SSETI Express SSETI Express Integration Logbook - Space.aau.dk

Project: SSETI Express 

Doc : SSETI Ex press Phase D – Integration Logbook 

Phase : Phase D 

Date : Started 5 th October 2004 

Ref : EXPRESS_D_ESA_Integration_Logbook.doc 

SSETI Express 

SSETI Express Integration Logbook 

Phase D 

Ref.: EXPRESS_D_ESA_Integration_Logbook.doc 

Issue: Updated daily 

Date: Started 5 th October 2004 

Prepared by: N Melville Date: 5 th October 2004 

Checked by: 

Date: 

Approved by: 

Date: 

SSETI, ESA-ESTEC, EXR-E 

P.O. Box 299, 2200 AG Noordwijk, The Netherlands 

Tel. (31) 71 565 6367 - Fax. (31) 71 565 5590 

E-mail: Neil.Melville@esa.int






Document change record 

Version Change By Date 

01 Drafting Neil Melville 05/06/2004 

X Daily updating Neil Melville N/A






Table of content 

1 Introduction......................................................................................................................................................9 

2 Defined procedures.......................................................................................................................................10 

Entering cleanroom..................................................................................................10 

Leaving cleanroom...................................................................................................10 

Taking equipment into cleanroom............................................................................10 

Cleaning aluminium boxes and parts .......................................................................11 

Cleaning honeycomb panels for insert potting.........................................................11 

Preparing glue and syringes .....................................................................................12 

Central insert potting................................................................................................14 

Side insert potting.....................................................................................................18 

Cleaning honeycomb panels after insert potting......................................................22 

Conformal coating....................................................................................................23 

Recommendations.................................................................................................24 

Tools and products................................................................................................24 

Procedure ..............................................................................................................24 

Lateral and solar panel integration...........................................................................36 

Integrated System Check procedure.........................................................................40 

3 General notes .................................................................................................................................................49 

2 nd August 2004........................................................................................................49 

9 th August 2004 ........................................................................................................49 

12 th August 2004 ......................................................................................................50 

16 th August 2004 ......................................................................................................50 

24th August 2004 .....................................................................................................50 

MILESTONE 1: EPS prototype and OBC EM successfully progress through 

operational modes, boot-up and watchdog sequences..........................................50 

30 th August 2004 ......................................................................................................51 

8 th September 2004...................................................................................................51 

13 th September 2004.................................................................................................51 

15 th September 2004.................................................................................................51 

16 th September 2004.................................................................................................52 

17 th September 2004.................................................................................................52 

19 th September 2004.................................................................................................53 

23 rd September 2004.................................................................................................53 

25 th September 2004.................................................................................................54 

27 th September 2004.................................................................................................54 

28 th September 2004.................................................................................................55 

MILESTONE 2: The first Cubesat arrives in ESTEC.........................................55 

30 th September 2004.................................................................................................55 

1 st October 2004.......................................................................................................55 

2 nd October 2004 ......................................................................................................56 

3 rd October 2004.......................................................................................................56 

4 th October 2004.......................................................................................................57






5 th October 2004.......................................................................................................57 

6 th October 2004.......................................................................................................57 

7 th October 2004.......................................................................................................67 

8 th October 2004.......................................................................................................70 

MILESTONE 3: Successful strip / mask tests on all DMM panels. ...................71 

9 th October 2004.......................................................................................................74 

MILESTONE 4: Completion of EM primary structure.......................................86 

10 th October 2004.....................................................................................................86 

11 th October 2004.....................................................................................................87 

12 th October 2004.....................................................................................................87 

13 th October 2004.....................................................................................................88 

14 th October 2004.....................................................................................................90 

15 th October 2004.....................................................................................................91 

16 th October 2004.....................................................................................................95 

17 th October 2004.....................................................................................................97 

18 th October 2004...................................................................................................100 

19 th October 2004...................................................................................................100 

20 th October 2004...................................................................................................101 

21 st October 2004...................................................................................................102 

22 nd October 2004 ..................................................................................................105 

MILESTONE 5: FM OBC booted up for the first time.....................................105 

23 rd October 2004...................................................................................................106 

24 th October 2004...................................................................................................108 

25 th October 2004...................................................................................................109 

MILESTONE 6: Stage 1 of the flight structure leaves ESTEC for Stuttgart ....113 

26 th October 2004...................................................................................................114 

27 th October 2004...................................................................................................116 

MILESTONE 7: The FM OBC downloads a picture from the FM CAM.........122 

28 th October 2004...................................................................................................122 

MILESTONE 8: EM structure completely assembled. .....................................123 

29 th October 2004...................................................................................................125 

30 th October 2004...................................................................................................126 

31 st October 2004...................................................................................................127 

MILESTONE 9: Successful functional integration of FM EPS and FM OBC. 128 

1 st November 2004 .................................................................................................132 

MILESTONE 10: This is the point of no return. ...............................................140 

2 nd November 2004 ................................................................................................141 

MILESTONE 11: The FM Primary Structure is completed..............................144 

3 rd November 2004.................................................................................................145 

4 th November 2004.................................................................................................145 

MILESTONE 12: The flight model Pressure Management System passes its 

vibration and leakage test. ..................................................................................145 

5 th November 2004.................................................................................................146 

6 th November 2004.................................................................................................148 

7 th November 2004.................................................................................................150






MILESTONE 13: Functional integration of OBC, EPS and MAGIC at the stage 

where the PROP payload is supported by the platform......................................153 

8 th November 2004.................................................................................................154 

MILESTONE 14: The SSETI Express safe mode and nominal mode beacons are 

received via a radio for the first time..................................................................156 

MILESTONE 15: Spacecraft hardware is commanded via an RF link for the first 

time. ....................................................................................................................157 

MILESTONE 16: SSETI Express transponds audio for the first time..............157 

9 th November 2004.................................................................................................157 

MILESTONE 17: The nominal mode beacon is received and decoded for the 

first time by the test ground station. ...................................................................160 

MILESTONE 18: SSETI Express receives, acknowledges and responds to its 

first RF telecommand. ........................................................................................161 

10 th November 2004...............................................................................................162 

11 th November 2004...............................................................................................166 

12 th November 2004...............................................................................................166 

13 th November 2004...............................................................................................167 

15 th November 2004...............................................................................................168 

16 th November 2004...............................................................................................171 

MILESTONE 19: SSETI Express passes its pressurised vibration tests...........174 

17 th November 2004...............................................................................................174 

18 th November 2004...............................................................................................175 

19 th November 2004...............................................................................................176 

21 st November 2004 ...............................................................................................178 

22 nd November 2004 ..............................................................................................179 

MILESTONE 20: 2-way link established between MCC, GND, UHF and OBC 

............................................................................................................................180 

23 rd November 2004...............................................................................................180 

24 th November 2004...............................................................................................183 

25 th November 2004...............................................................................................186 

26 th November 2004...............................................................................................189 

27 th November 2004...............................................................................................190 

MILESTONE 21: The propulsion thrusters are fired via OBC and MAGIC for 

the first time and work perfectly.........................................................................190 

28 th November 2004...............................................................................................190 

29 th November 2004...............................................................................................190 

30 th November 2004...............................................................................................191 

2 nd December 2004.................................................................................................191 

3 rd December 2004 .................................................................................................192 

6 th December 2004 .................................................................................................193 

MILESTONE 22: The FM PCU fires the T-Pods for the first time ..................195 

7 th December 2004 .................................................................................................195 

8 th December 2004 .................................................................................................198 

9 th December 2004 .................................................................................................200 

10 th December 2004 ...............................................................................................200






13 th December 2004 ...............................................................................................202 

14 th December 2004 ...............................................................................................204 

15 th December 2004 ...............................................................................................206 

16 th December 2004 ...............................................................................................207 

MILESTONE 23: Successful S-BAND data downlink for the first time..........207 

17 th December 2004 ...............................................................................................209 

MILESTONE 24: The satellite is powered-up using the battery box for the first 

time. ....................................................................................................................209 

12 th January 2005 ...................................................................................................210 

17 th January 2005 ...................................................................................................212 

19 th January 2005 ...................................................................................................213 

24 th January 2005 ...................................................................................................215 

25 th January 2005 ...................................................................................................216 

26 th January 2005 ...................................................................................................218 

27 th January 2005 ...................................................................................................219 

31 st January 2005....................................................................................................223 

1 st February 2005....................................................................................................224 

8 th February............................................................................................................225 

9 th February............................................................................................................225 

10 th February..........................................................................................................227 

11 th February..........................................................................................................229 

14 th February 2005 .................................................................................................230 

15 th February 2005 .................................................................................................231 

16 th February 2005 .................................................................................................235 

17 th February 2005 .................................................................................................239 

18 th February 2005 .................................................................................................240 

22 nd February 2005.................................................................................................240 

24 th February 2005 .................................................................................................241 

28 th February 2005 .................................................................................................244 

1 st March 2005........................................................................................................245 

2 nd March 2005.......................................................................................................246 

3 rd March 2005 .......................................................................................................250 

MILESTONE 25: The S-Band sub-system is declared flight-ready. ................252 

4 th March 2005........................................................................................................253 

5 th March 2005........................................................................................................255 

6 th March 2005 .......................................................................................................257 

7 th March 2005 .......................................................................................................257 

8 th March 2005 .......................................................................................................262 

9 th March 2005 .......................................................................................................266 

10 th March 2005 .....................................................................................................275 

13 th March 2005 .....................................................................................................278 

14 th March 2005 .....................................................................................................278 

15 th March 2005 .....................................................................................................279 

MILESTONE 26: The OBC and EPS software are declared FLIGHT READY 

............................................................................................................................286






16 th March 2005 .....................................................................................................286 

17 th March 2005 .....................................................................................................292 

18 th March 2005 .....................................................................................................298 

20 th March 2005 .....................................................................................................305 

21 st March 2005......................................................................................................312 

22 nd March 2005.....................................................................................................314 

23 rd March 2005 .....................................................................................................319 

24 th March 2005 .....................................................................................................326 

25 th March 2005 .....................................................................................................329 

26 th March 2005 .....................................................................................................331 

27 th March 2005 .....................................................................................................335 

28 th March 2005 .....................................................................................................342 

29 th March 2005 .....................................................................................................343 

30 th March 2005 .....................................................................................................346 

31 st March 2005......................................................................................................353 

1 st April 2005..........................................................................................................355 

2 nd April 2005.........................................................................................................358 

4 th April 2005 .........................................................................................................363 

5 th April 2005 .........................................................................................................368 

6 th April 2005 .........................................................................................................371 

7 th April 2005 .........................................................................................................374 

8 th April 2005 .........................................................................................................378 

9 th April 2005 .........................................................................................................379 

10 th April 2005 .......................................................................................................383 

MILESTONE 27: The integration is completed................................................391 

12 th April 2005 .......................................................................................................391 

MILESTONE 28: The spacecraft is ready for vibration testing. .......................394 

13 th April 2005 .......................................................................................................396 

14 th April 2005 .......................................................................................................400 

15 th April 2005 .......................................................................................................403 

MILESTONE 29: The vibration tests are completed. .......................................408 

16 th April 2005 .......................................................................................................408 

19 th April 2005 .......................................................................................................414 

20 th April 2005 .......................................................................................................415 

21 st April 2005........................................................................................................422 

22 nd April 2005.......................................................................................................424 

24 th April 2005 .......................................................................................................429 

25 th April 2005 .......................................................................................................430 

26 th April 2005 .......................................................................................................430 

27 th April 2005 .......................................................................................................434 

28 th April 2005 .......................................................................................................434 

29 th April 2005 .......................................................................................................437 

1 st May 2005...........................................................................................................438 

2 nd May 2005..........................................................................................................438 

3 rd May 2005 ..........................................................................................................438






4th May 2005 .........................................................................................................438 

May 5 th 2005 ..........................................................................................................446 

May 6 th 2005 ..........................................................................................................447 

7 th May 2005 ..........................................................................................................450 

MILESTONE 30: The fit-check is successfully completed ..............................451 

8 th May 2005 ..........................................................................................................452 

9 th May 2005 ..........................................................................................................452 

MILESTONE 31: The spacecraft is ready for thermal-vacuum testing............457 

10 th May 2005 ........................................................................................................457 

11 th May 2005 ........................................................................................................458 

12 th May 2005 ........................................................................................................460 

13 th May 2005 ........................................................................................................461 

16 th May 2005 ........................................................................................................465 

17 th May 2005 ........................................................................................................473 

18 th May 2005 ........................................................................................................473 

20 th May 2005 ........................................................................................................476 

21 st May 2005.........................................................................................................478 

24 th May 2005 ........................................................................................................480 

25 th May 2005 ........................................................................................................481 

26 th May 2005 ........................................................................................................482 

30 th May 2005 ........................................................................................................486 

1 st June 2005...........................................................................................................492 

2 nd June 2005..........................................................................................................498 

3 rd June 2005 ..........................................................................................................501 

5 th June 2005 ..........................................................................................................502 

6 th June 2005 ..........................................................................................................504 

7 th June 2005 ..........................................................................................................512 

8 th June 2005 ..........................................................................................................513 

9 th June 2005 ..........................................................................................................515 

10 th June 2005 ........................................................................................................517 

12 th June 2005 ........................................................................................................518 

13 th June 2005 ........................................................................................................518 

14 th June 2005 ........................................................................................................522 

15 th June 2005 ........................................................................................................522 

16 th June 2005 ........................................................................................................523 

17 th June 2005 ........................................................................................................536 

18 th June 2005 ........................................................................................................537 

20 th June 2005 ........................................................................................................539 

21 st June 2005.........................................................................................................540 

22 nd June 2005........................................................................................................541 

23 rd June 2005 ........................................................................................................548 

24 th June 2005 ........................................................................................................550 

27 th June 2005 ........................................................................................................551 

Solar panel preparations ......................................................................................................................................554 

4 Harness..........................................................................................................................................................556






5 Problems .......................................................................................................................................................558 

6 Modifications...............................................................................................................................................579 

7 Connector usage..........................................................................................................................................594 

8 Arrivals and Departures .............................................................................................................................596 

Glue batch usage...................................................................................................................................................600 

9 Milestones ....................................................................................................................................................602 

10 Lessons Learned.....................................................................................................................................604 

11 APPENDIX I – OBC report .................................................................................................................610 

12 APPENDIX II – Lessons learned from INFRA / UHF ...................................................................611 

UHF ........................................................................................................................611 

INFRA....................................................................................................................612






1 Introduction 

This document is intended to serve as an informal logbook detailing day-by-day the 

integration of the SSETI Express spacecraft as seen from the Managerial and System- 

Engineering levels here in ESTEC. Having such a detailed record will enable accurate and 

efficient troubleshooting and back-reference and also act as a valuable resource for future 

SSETI projects. 

The latter sections keep track of various issues such as connector usage and glue batches. 

A set of procedures for the various activities are defined and a set of lessons learned during 

the integration process are presented. 

This document is constantly updated. 

9






2 Defined procedures 

Entering cleanroom 

1. Enter airlock of room Em111 (ESA_Jason and ESA_Neil are in possession of access 

cards) 

2. Close door 

3. Walk over ‘dirty’ mat with both shoes 

4. Stay on the ‘dirty’ side of the bench 

5. Ensure that cleanroom lights are turned on (switch on the left wall) 

6. Take and wear hat from ‘guest’ cupboard 

7. Take and wear a cleanroom coat from ‘guest’ cupboard 

8. Take overshoes from ‘guest’ cupboard 

9. As you apply overshoes stop over the bench onto the ‘clean’ mat 

10. Enter door code (ask from ESA_Jason or ESA_Neil who will provide stagiers with the 

code under careful discretion) 

11. Enter cleanroom quickly 

12. Close the door behind you (there is an over-pressure) 

Leaving cleanroom 

1. Ensure that all equipment that you used is turned off 

2. If you are the last to leave then ensure that ALL equipment is turned off 

3. Enter into airlock and close door behind you to maintain over-pressure 

4. Step over bench 

5. Remove and discard overshoes 

6. Remove and replace cleanroom coat in cupboard 

7. Remove and discard hat 

8. If you are the last to leave ensure that the light is turned off (switch in the airlock on 

the right wall as you leave) 

9. Leave the airlock, closing the door behind you 

Taking equipment into cleanroom 

1. Inform ESA_Jason or ESA_Neil that you wish to take the equipment inside 

2. Wipe clean all exposed surfaces of the equipment 

3. Vacuum dust out of any ‘enclosures’ (laptop fans for example) 

4. There is MINIMAL PAPER allowed inside the cleanroom 

5. There is NO CARDBOARD allowed inside the cleanroom 

10






6. Only LINT FREE tissues allowed inside the cleanroom (provided) 

7. Carry equipment as you enter the cleanroom in the usual manner 

Cleaning aluminium boxes and parts 

1. Disassemble parts are far as possible 

2. Wear gloves 

3. Use wire wool (see ESA_Jason) to scrub surfaces, in ONE direction only, under 

running warm water (sink outside cleanroom) 

4. Dry parts thoroughly 

5. Clean thoroughly with IPA tissues 

6. Store on tissues in cleanroom 

Alternatively: Instead of steps 3,4,5 use the ultrasound bath, however, note that it is small, 

and dangerous to put hand in when it is in operation (and not too good when not in operation). 

Cleaning honeycomb panels for insert potting 

1. Wear gloves to avoid additional fingerprints 

2. Remove panels from packaging and place on soft, clean surface (like cardboard) 

3. Wipe all dust and debris off of panels with a dry cloth, taking care to maintain a clean 

working surface 

4. Run the nozzle of a working vacuum cleaner along the exposed core at the sides of 

each panel, taking care not to make contact. A good way to do this is to have a piece 

of cardboard beneath the panel and protruding slightly, you can then run the nozzle 

along the cardboard while lightly holding the panel and cardboard in place 

11






5. Stand the panel up and place the nozzle of a working vacuum cleaner within a 

centimetre or so of each of the central insert holes, again, taking care not to touch (a 

gloved finger in between is good) 

6. Lay the panel down and clean around each central insert hole with IPA impregnated 

tissues 

7. Any excessive grease should be removed with acetone, taking care to use appropriate 

hand protection and room ventilation 

Preparing glue and syringes 

1. Collect the following equipment: 

a) Weighing scales 

b) Tissues 

c) Gloves 

d) Syringes (around 30ml, with standard ‘screw’ top) 

e) Standard plastic “push on” syringe tips (these sill still screw in ok, and are 

easier to use than needles) 

f) Scissors / scalpel 

g) Tin of Araldite AY103-1 

h) Tin of Araldite Hardener HY991 

i) Clean mixing pot 

j) Stirring sticks (plastic) 

k) Small spatula 

l) IPA impregnated tissues 

m) Vacuum chamber 

2. Put the gloves on and have the tissues ready. Both components of the glue are 

harmful to the skin so care should be taken and any mess should be cleared up 

immediately. 

3. Ensure the area is well ventilated 

12






4. Weigh out the required amount of AY103 into the clean mixing pot (the ratio of 

AY103 to HY991 is 5 parts to 1, a standard insert uses 3-4 ml of the resulting 

mixture). This can simply be poured from the tin (being careful to clean up any mess 

using the tissues). 

5. Weigh out the required amount of HY991 into the same pot, it is best to use a spatula 

for this. 

6. Stir the mixture using a stirring stick until it is homogenous (a few minutes) 

7. Allow the mixture to settle for 10 minutes while you clean the area, the stirring stick 

and the spatula using IPA tissues. 

8. Place the pot in the vacuum chamber and reduce the pressure to around 30 Pa, or as 

low as it can go without the mixture “frothing” over the top of the pot (be warned that 

it increases dramatically in volume and is hard to clean) 

9. Allow the mixture to settle for around 10 minutes 

13






10. In the meantime take one of the tips for the syringe and cut the end off carefully so 

that the resulting end radius is as large as possible (for easy injection) whilst still being 

small enough to fit in the hole in the inserts (around 2mm diameter) 

11. Bleed air back into the chamber until ambient pressure is reached 

12. Remove the pot from the chamber 

13. Carefully and slowly suck glue up into the syringe (around 30ml is recommended). 

NOTE: This is not easy, it take a fair bit of strength and time and you should be very 

careful not to suck up air with the glue (try to tilt the pot if it helps). The glue made 

fade in colour while you ‘stretch’ it in the syringe, this is normal 

14. Screw the prepared tip onto the end of the syringe 

15. Hold the syringe vertically upwards, wait for any air to rise to the top, and then push it 

out 

16. Use the tissues to clean the syringe and tip exterior 

17. Inject glue into inserts as necessary until syringe empty 

18. Remove tip (using tissues) 

19. Expel all air and glue from the syringe 

20. Repeat steps 13 to 19 until job complete, up to a maximum of 90 minutes from the 

execution of step 6. 

21. When the pot is empty, or the glue is unusable (past 90 minute pot-life), clean the pot 

with tissues and then IPA tissues to remove all traces of glue 

22. It is recommended to retain the last few ml in the end of the syringe, label the batch 

number, time and date on the syringe, and leave it to harden as a specimen. 

Central insert potting 

1. Ensure panel is properly clean 

2. Clean the inserts by immersion and stirring in acetone for a few minutes, taking care 

to wear appropriate hand protection and ensure adequate ventilation for the fumes 

14






3. Dry the inserts with lint-free tissues 

4. Wear vinyl gloves as much as possible and avoid touching panel with bare skin so as 

not to leave additional fingerprints 

5. Inspect holes for any parts of core that might obstruct the insertion of an insert, 

carefully cut such offending parts away with a scalpel, or push them gently back into 

place 

6. Cut small lengths of kapton tape and cover each of one side of the central insert holes, 

ensuring that the tape is flat over the hole and securely attached all the way around the 

edges 

7. Carefully turn the panel over 

8. Refer to the integration manual and insert the correct inserts into the correct positions 

on the panel, placing them all up the same way 

9. Have someone else double-check that the correct types of insert are placed into the 

correct holes 

15






10. Cut small lengths of kapton tape and cover the other side of inserts and holes as 

before, ensuring that all inserts are within the skins of the honeycomb and al insert 

perimeters are surrounded by kapton. 

11. For those types of insert that are 10.3mm instead of 10.6mm press the kapton tape on 

the top and bottom of the insert from both sides of the panel simultaneously to ensure 

good adhesive contact and central location of the insert 

12. Using a sharp point make holes in the kapton tape to fit the holes manufactured in the 

inserts themselves (two per insert) 

13. Prepare some glue and syringes 

14. Equip yourself with many tissues 

15. Place the tip of the syringe in one hole of an insert and gently depress the plunger. 

NOTE: it will be very hard at first to tell if anything is happening! 

16. Inject the glue slowly into the insert until it starts to emerge from the other hole in the 

top, this should be done slowly enough to ensure that the glue flows into all the 

corners and gaps of the honeycomb. At least one to two minutes per insert is 

appropriate, but some may take longer depending upon the number of honeycomb 

cells the insert intersects 

16






17. Once the insert is full remove the syringe, leaving a blob of glue covering each of the 

holes, but taking care not to leave glue on the skin of the honeycomb (IPA tissues will 

clean any drops of glue from the skin of the honeycomb if necessary) 

18. As inserts are glued it is best to mark them using a soft pen on the kapton tape 

19. Repeat steps 15-17 for the other inserts in parallel with the following instructions 

20. The glue will slump and settle in the honeycomb, sucking material from the blobs left 

over the holes. As these decease, top them up as necessary by dripping glue onto the 

appropriate places 

21. After a few hours the glue should be beginning to harden and the inserts should no 

longer be ‘sucking’ material down inside. After a MAXIMUM of four hours the 

excess glue should be cleaned off of the kapton tape so that it doesn’t impair the 

surface of the panel. 

22. Using IPA tissues carefully clean around the outside of each insert and remove all glue 

off of the surface of the honeycomb and carefully from the top of the insert itself. 

23. Note which inserts have been glued with which batch of glue 

24. Ensure that the panel remains in its current orientation for at least 12 hours 

25. Do not remove the kapton tape for at least 48 hours 

17






26. Do not put any strain on the inserts until absolutely necessary. NOTE: full strength is 

only achieved after a week of curing time. 

Side insert potting 

1. Ensure panel is properly clean 

2. Clean the inserts by immersion and stirring in acetone for a few minutes, taking care 

to wear appropriate hand protection and ensure adequate ventilation for the fumes 

3. Wear gloves as much as possible and avoid touching panel and inserts with bare skin 

so as not to leave additional fingerprints 

4. Place or stick a strip of kapton tape sticky-side-up on the table and place side inserts 

and squares of paper onto it alternately. 

Mark II: The paper should be a little smaller than the width of the kapton tape. 

Mark III: It is even better to use more kapton tape “upside down” instead of paper. 

5. Cut down the centre of each piece of covered kapton to produce little “tie-fighters” 

18






6. With a scalpel cut carefully a ‘window’ around the bolt hole, removing a small circle 

of kapton covering the thread 

7. Identify the appropriate mounting strip (using the panel drawings) and attach the 

appropriate number of tie-fighters to one side in the correct orientation for the side 

pockets in question 

8. Place the strips on the panel ensuring that the reference end of the strip is flush with 

the correct side of the honeycomb skin. Tape the strip firmly in place with kapton 

tape. 

19






9. Tape the paper “handles” of the tie-fighters tightly onto the honeycomb surface so that 

the inserts are securely positioned 

10. Remove the strip first by unscrewing the bolts from the inserts, and then by removing 

the tape securing it to the panel (this order is important so that you do not disturb the 

inserts) 

20






11. Have someone else check that the inserts are corrected positioned and around the right 

way (they are not symmetrical) 

12. Now that the inserts are correctly positioned an extra layer of kapton should be applied 

to protect the thread from any stray glue 

13. Apply kapton tape to ensure that all top-layer honeycomb cells intersecting with the 

side pocket are covered 

14. Prepare some glue and syringes 

15. Using a needle make small holes in the kapton tape above the air holes of the insert, 

and then also make small holes either side of the insert, one of them large enough to 

accommodate the end of the tip of the syringe 

16. Equip yourself with many tissues 

17. Place the tip of the syringe into the larger hole one side of the insert and gently depress 

the plunger. NOTE: it will be very hard at first to tell if anything is happening! 

18. Inject the glue slowly into the pocket until it starts to emerge from the other hole in the 

top and from the holes in the top of the insert, this should be done slowly enough to 

ensure that the glue flows into all the corners and gaps of the honeycomb. At least one 

to two minutes per insert is appropriate, but some may take longer depending upon the 

number of honeycomb cells the pocket intersects 

19. Once the pocket is full remove the syringe, leaving a blob of glue covering each of the 

four holes, but taking care not to leave glue on the skin of the honeycomb (IPA tissues 

will clean any drops of glue from the skin of the honeycomb if necessary) 

20. Repeat steps 16-18 for the other inserts in parallel with the following instructions 

21. The glue will slump and settle in the honeycomb, sucking material from the blobs left 

over the holes. As these decease, top them up as necessary by dripping glue onto the 

appropriate places 

22. After a few hours the glue should be beginning to harden and the pockets should no 

longer be ‘sucking’ material down inside. After a MAXIMUM of four hours the 

excess glue should be cleaned off of the kapton tape so that it doesn’t impair the 

surface of the panel. 

23. Using IPA tissues carefully clean around the outside of each insert and remove all glue 

off of the surface of the honeycomb and carefully from the top of the insert itself. 

21






24. Note which inserts have been glued with which batch of glue 

25. Ensure that the panel remains in its current orientation for at least 12 hours 

26. Do not remove the kapton tape for at least 48 hours 

27. Do not put any strain on the inserts until absolutely necessary. NOTE: full strength is 

only achieved after a week of curing time. 

28. In a future gluing session top-up glue in a side insert if the existing glue does not come 

at least halfway up the insert on both sides 

Cleaning honeycomb panels after insert potting 

1. Wear gloves so as to avoid adding fingerprints to the panels 

2. Remove all kapton tape, carefully using a scalpel if necessary to cut kapton away from 

any exposed glue. It is important to remove the kapton, since this particular type uses 

adhesive which will outgas in space 

3. Ensure that the area is adequately ventilated 

4. Equip yourself with many lint free tissues, a small bowl, a bin and a bottle of acetone. 

5. Wear appropriate hand protection 

6. Pour some acetone into the bowl 

7. Dip a clean area of tissue into the acetone and then wipe it from one end of the panel 

to the other in one fairly slow and continuous motion. The panel should become 

visibly cleaner where you have wiped, and the tissue will pick up grease. NOTE: Care 

should be taken to AVOID wiping directly over an insert, as the acetone can damage 

the properties of the glue. 

8. Repeat step 7 until the panel will get no cleaner, always wiping in the same direction, 

taking care not to contaminate the acetone with dirty tissue and topping up the acetone 

as required 

9. After cleaning both sides of the panel dispose of the tissues in a safe bin and deposit 

the remaining acetone in a disposal bottle 

10. Take the panel into the cleanroom 

22






Conformal coating 

1. Prepare a well ventilated area. NuSil CV 1152 conformal coating is harmful if 

inhaled. 

2. Clean the appropriate PCB by placing a lint-free tissue over it and then brushing 

isopropanol through it onto the board using a stiff wire brush, taking care not to 

damage the components. 

3. All the isopropanol to dry. 

4. Ensure that the PCB is level and stable. 

5. Wear a mask to reduce inhalation of fumes. 

6. Check with the PCB manufacturer which areas, if any, should be avoided. 

7. Open the CV 1152 and pour some into a clean container (like a plastic cup). 

8. Close the bottle of CV 1152. 

9. Using a small flexible brush apply the conformal coating to the PCB. It is better to 

dab it on rather than using broad brush strokes. 

10. While you work allow the coating to flow into all the gaps between component legs 

etc, it will do so naturally and does not need to be forced. 

11. Ensure that a layer of around 1mm thickness is covering the entire board, including 

the tops, sides and legs of all components. 

12. Place the PCB in a vacuum oven and reduce to the lowest pressure possible. You 

will probably see the coating froth and expand slightly as the air is forced out. 

13. Once it is flat again, or after about 10 minutes (whichever comes first), slowly, 

return the pressure to normal and remove the PCB. 

14. Leave the PCB to dry in a well ventilated area for at least 12 hours (preferably 24) 

before manipulating it. 

23






Solar panel laydown 

Recommendations 

All operations must be accomplished with gloves, hat, coat and overshoes in a cleanroom. 

The primer bottle must be closed immediately after use because it is hydrophile. 

The silicone in the bonding material is very difficult to clean and easily contaminates every 

surface it comes in contact with. Avoid having smears of silicone on the gloves since it can be 

easily deposited on the cells or any other object. 

Check the gloves for any trace of RTV before manipulating the panel or the cells. 

Tools and products 

Gloves 

White sheets 

Alcohol-impregnated tissue 

Clean tissue 

Alcohol 

Another cleaning product 

Bonding material : 

RTV655A (Silicon) 

RTV655B (hardener) 

Primer: 

DC1200 

Silicon – primer mixing plate 

Low-tack tape (LTT), thickness 0.07mm, width 2 inches (50mm). 

Kapton tape, thickness 0.05mm, width 1 and 2 inches. 

A spatule, approximately one centimeter wider than the solar cell. 

Procedure 

Workspace preparation: 

- setup separate table for all chemical products 

- clean the table with alcohol-impregnated tissue 

- lay down white sheets on the table and tape them to the table 

- clean the aluminum panel with truc qui arrache 

24






Prepare the string 

- write down the string number, cell class and panel number 

- Check that the solders have been verified with a binocular and visually check the 

solders again 

- position it so that the glass faces the table 

- clean the back of the string with an alcohol tissue 

Lay the kapton down on the aluminium panel 

- tape one strip of LTT on the table 

- align the panel next to it 

- tape the start of the kapton tape to the LTT 

- while the first operator holds the kapton tape aligned over the panel, the second 

operator tapes the kapton to the panel inch by inch using a dry tissue and checking for 

the presence of air bubbles 

- if there is no air bubble at all between the kapton and the panel, the procedure is 

reiterated for the next kapton strip. Two adjacent strips must overlap by approximately 

3mm. The panel must be completely covered with kapton tape 

- cut out the surplus of kapton with a scalpel 

25






26






Kapton laydown 

Attach the aluminium panel to the table using LTT 

- attach the aluminium panel with two parallel strips, one on each side 

- if the strip is large enough, it can be used both to attach the panel and to position the 

bonding material area 

27






Clean the kapton 

- use alcohol on a clean sheet 

Lay down the positioning tape 

- in order to obtain the correct bonding material thickness, three layers of tape are 

required on each side of the string.The first two layers are LTT and the third is kapton. 

Total height is therefore 0.019mm, except on areas where the kapton strips are 

overlapping: there thickness is reduced to 0.014mm 

- the tape strips must be placed on the panel so that the bonding area has a 10mm 

clearance to any hole. The left and right strips must be parallel and separated by the 

width of the string plus a 1mm margin on each side(in this case, 67mm) 

- top and bottom positioning strips are single-layer LTTs. They are separated by the 

length of the string excluding both end pads plus a 2mm margin. There must be no 

bonding material beneath both end pads 

28






Clean the back of the string 

- use alcohol on a clean sheet 

Apply the primer on the kapton and on the back of the string 

- Write down the batch number of the primer bottle 

- fold a clean tissue several times to obtain a spatule shape 1inch wide 

- put some primer in a cup and dip the spatule-shaped tissue in it. If the primer is 

already white, do not use it and take new one 

29






- tap the tissue against another clean tissue to remove all surplus of primer. It is 

important that the primer evaporates quickly in order to prevent it to flow on the other 

side of the cell 

- apply the primer along the width of the cell 

- apply the primer along the width of the bonding area 

- write down the time of application 

- wait 15 minutes before applying the bonding material onto any primer-covered area. 

The primer has to become white 

Attach LTT at both ends of the string 

- the string must be hold at each end with a special LTT strip 

- see figures for the shape of the each LTT strip 

- only the surface in direct contact with the glass surface of the cell must be sticky. All 

the rest of the strip’s surface must be covered with another strip 

30






Prepare the bonding material 

- write down the batch number of the RTV bottle 

- open a bottle of RTV (red paste) 

- put between 25 and 100g of RTV in a plate 

- add a drop of hardener 

- mix the RTV and the hardener with the spatule 

- put a sample of the bonding preparation in a cup. Identify it with the string number 

31






Apply the bonding material on the kapton with the spatula 

- put some bonding preparation on the entire span of the spatula 

- start on the top LTT strip and move down the entire length of the bonding area while 

keeping both ends of the spatula on the side strips 

- repeat the operation until all the bonding area is uniformly covered with the bonding 

material 

32






Remove all the positioning tape 

- take care not to lift the aluminum panel or to touch the bonding material while 

removing the positioning tape 

Lay down the string 

- this part requires three persons 

- the string should be oriented with its two original pads (+) towards the bottom of 

the panel 

- two operators hold each one side of the string while the third operator holds a 

tissue beneath the string during its manipulation to prevent the string to hit the 

table in case one of the end tapes detaches from the string 

33






- the first operator lowers his end of the string onto the panel while the second 

operator keeps his end at approximately 30cm above the panel 

- the second operator then lowers his end while the first operator gently presses the 

cells individually against the glue 

Adjust string position 

- once laid down, the string’s position can be adjusted cell by cell, very carefully 

- press very gently with a finger from the center to the border of each cell in order to 

evacuate any remaining air bubble 

- wait at least 30 minutes before proceeding to the next laydown 

34






LTT 

outline 

Kapton 

tape 

RTV 

35






Lateral and solar panel integration 

1) Integrate the +x patch antenna to the +x lateral panel and the +x+y corner 

profile. Do not tighten the bolts fully. 

2) Integrate solar panels 4 and 5 to the +x lateral panel and the +x+y corner 

profile. Use only those bolt holes that do not also interface with the primary 

structure. Do not tighten the bolts fully. 

3) Integrate the +x lateral panel to the primary structure. 

4) Tighten all the bolts on the +x lateral panel, apart from those on the +x+y 

corner profile. 

5) Integrate the –x patch antenna to the –x lateral panel and the –x+y corner 


6) Integrate solar panel 9 to the –x lateral panel. Use only those bolt holes that do 

not also interface with the primary structure. Do not tighten the bolts fully. 

7) Integrate the –x lateral panel to the primary structure. Do not tighten the bolts 

fully. 

8) Integrate solar panel 10 to the –x lateral panel and primary structure. 

9) Tighten all of the bolts on the –x lateral panel, apart from those on the –x+y 


10) Integrate solar panel 6 to the +y lateral panel. Use only those bolt holes that do 


11) Integrate the +y sun sensor to the +y lateral panel and solar panel 6. 

i. Pass the trailing sun sensor harness through the 4mm hole adjacent to 

the fixation holes. 

ii. Remove the protective cover from the sun sensor and take care at all 

times not to touch the delicate soldering near the window on the front. 

iii. Tighten the two M2 securing bolts until the head of each is touching 

the PCB 

iv. Tighten the M2 nuts on the reverse side until they are touching the 

PCB. 

v. Place two M2.5 washers on top of each of the fixation holes on solar 

panel 6. 

vi. Locate the sun sensor so that the bolts line up with the holes and the 

harness passes between the bolts underneath and into the harness hole. 

vii. Gently push the bolts through the washers, solar panel 6, and the +y 

lateral panel. 

viii. Holding the sun sensor in place attach M2 nuts to the bolts on the rear 

side of the +y lateral panel. 

ix. Tighten the nuts with a spanner. 

36






x. Replace the sun sensor protective cover. 



13) Integrate the +y lateral panel to the primary structure and to both the +y corner 

profiles. Do not tighten the bolts fully. 

14) Integrate solar panel 7 to the +y lateral panel and the primary structure. 

15) Tighten all the bolts on the +y lateral panel. 

16) Tighten all the bolts on the +x and –x sides of the +y corner profiles. 

17) Locate the coil driver correctly on the inside of the –y lateral panel, place M3 

bolts through the fixation holes and tape the heads in place so that they will not 

fall when the panel is turned. 

18) Turn the –y lateral panel over and carefully place solar panel 2 over the coil 

driver fixation bolts. Add nuts to these bolts to hold them in place. The 

kapton on the reverse side can then be removed. 

19) Integrate solar panel 2 to the –y lateral panel. Use only those bolt holes that do 


20) Integrate solar panel 3 to the –y lateral panel and the –y+x corner profile. Use 

only those bolt holes that do not also interface with the primary structure. Do 

not tighten the bolts fully. 

21) Integrate the –y sun sensor to the –y lateral panel. For this, repeat the 

instructions given in step 11 above (replacing +y with –y and solar panel 6 

with solar panel 3 throughout). 

37






22) Integrate solar panel 1 to the –y lateral panel and the –x-x corner profile. Use 



23) Integrate the –y lateral panel to the primary structure, taking care to slide the 

+x-y corner profile under the +x magnetorquer coil. 

24) Tighten all the bolts on the –y lateral panel. 

25) Tighten all the bolts on the +x and –x sides of the –y corner profiles. 

38






NOTE: During integration and disintegration of the lateral panels it will often be necessary 

to store a lateral panel to one side for a while. This should not be done by leaning them up 

against a wall or a table leg, as they will easily fall. Instead the side protector, with the nuts 

done up tightly, makes an ideal stand – whichever way you need to store / work on the panel. 

39






Integrated System Check procedure 

1) Integrated System Check information form 

Check Number: 

Date: 

Start time: 

Test carried out by: 

Circumstances: 

Powering: 

Internal battery 

External power source 

Internal battery and 

solar panel simulation 

External power source 

and solar panel simulation 

End time: 

Log-file path and filename on FTP: 

This report path and filename on FTP: 

40






2) Test setup 

For external powering 

1) Ready to supply 28 volts to pins 13 and 14 of the external power supply connecter, 

with the ground running through pin 5 of the OBC debug connector (or to the ASAP 

ring if it is accessible). 

2) Current limit set to maximum, and run the positive line through a digital ammeter. 

3) The EPS ARM plug should NOT be connected. 

4) The S-Band ARM plug should be present. 

5) When the spacecraft should be powered up, simply turn the power supply on. 

For internal powering 

1) The external power connector should be left empty. Ready to connect the EPS ARM 

plug to turn the spacecraft on. 

2) The S-Band ARM plug should be present. 

3) When the spacecraft should be powered up then apply the ABF connector, wait for 

five seconds (to charge the timer skip capacitor), disconnect it, and then reapply. The 

spacecraft should boot-up. (If the last boot-up was within the last 30 minutes then 

only one application is necessary as the capacitor should still be charged. 

Groundstation 

1) Should already be set up internally – don’t change anything. 

2) Ensure that the ICOM is connected to COM port 1 

3) Ensure that the TNC is connected to COM port 4 

4) Turn the power supply on and check all the lights come on 

5) Run PCR-100 controller, should see a non-zero signal (noise). If communication is 

not possible then a dialogue box will appear to inform 

6) Run term.exe and ensure it is set to 56k7 

7) Connect and send a test TC, should see the PTT lights come on on the UHF side of the 

TNC and the UHF radio 

8) Run DTMF.exe and make sure that the audio spectrum is present (noise) 

41






3) Platform checkout 

1 

2 

3 

4 

5 

6 

7 

8 

Action Test Response 

Measure battery voltage from pin 2 Voltage should be between 22.3V 

of EPS SAFE/ARM and pin 5 of (safe mode entry) and 24.67V (full). If 

OBC debug 

it is lower than 24.3V then it must be 

charge before continuing 

Power the spacecraft by turning on 

external power supply, or by 

applying the safe/arm connector 

Listen, with an open squelch, using 

a UHF handheld on 437.250MHz 

Watch the current consumption if 

powering externally 

Watch for data reception on the 

groundstation and note the boot 

attempt number 

When data is received, send 

telecommand to set TX delay: 

OBC 32 010F 

Wait 5 seconds, then send it again 

Send two time-synch commands 

with a gap of ten seconds in 

between 

Command S-Band carrier up with 

DTMF tones “B0” and measure 

the frequency with the counter 

If on external power then current 

consumption should read around 

220mA for the first 30 seconds 

At least one safe mode beacon should 

be heard, noise should be replaced by 

bursts of silence over 1.4 seconds 

every 30 seconds (only 0.2s if on 

external power) 

Should read 300mA in recovery mode, 

then 357mA in nominal mode (once 

ACDS has settled) 

If the battery voltage in step 1 is over 

24.3V then data should be received, 

otherwise the spacecraft will stay in 

safe mode and needs charging. If on 

internal power then beacon data 

should reflect voltage from step 1 

(unless powering externally) 

Should see an acknowledge from OBC 

(MID = F8, Data = 00 32 0F 01) 

For each see an acknowledge and a 

time difference. The second time 

difference should be small, and the 

next beacon should have an accurate 

time stamp 

Should see signal level raise and 

current consumption reach 600- 

750mA. Frequency should be close to 

2401.8 MHz 

V 

OK 

mA 

OK 

OK 

mA 

OK 

V 

Boot # 

OK 

OK 

Delta 

Time 

OK 

mA 

MHz 

OK 

42






9 

Command S-Band carrier down 

with DTMF tones “B0” 

Should see signal level drop back to 

noise level, and current back to 

357mA 

mA 

OK 

10 

Turn live stream on down UHF 

with telecommand: 

OBC 18 FF04 

Should see an acknowledge from 

OBC, and then telemetry as it arrives 

in the stack 

OK 

11 

Check ACDS data in the nominal 

mode beacon 

Field strength should be static, and 

field rate should be zero (may take a 

few minutes to settle down) 

Strength 

OK 

12 

Note TCS temperatures 

Should be close to ambient 

Deg C 

Deg C 

Deg C 

OK 

13 

14 

Request the entire alarm stack on 

UHF with the telecommand: 

OBC 02 FF04 

Command ACDS telemetry rate 

increase with telecommand: 

ACDS 01 010A 

Should get alarm from ACDS with 

ASCII translation “ACDS Normal 

Operation” and a timestamp from 

boot-up 

Should see telemetry rate of ACDS 

increase from every 30 seconds to 

every 6 seconds 

OK 

OK 

15 

16 

17 

Command S-Band TNC on with 

DTMF tones “B5” and make sure 

that the ICOM is using a wide 

filter 

Request 100 picture packets via S- 

Band with telecommand: 

OBC 03 000C 

Command S-Band TNC off with 

DTMF tones “B5” 

Should see the current consumption 

rise by 27mA 

Should see the acknowledge, and then 

100 packets arrive while the signal is 

up on the ICOM 


drop back to where it was before step 

15 

mA 

OK 

Packets 

OK 

mA 

OK 

The above procedure confirms that: 

- The PCU is alive 

43






- The UHF radio and TNC are alive 

- The OBC is alive 

- The S-Band radio and TNC are alive 

- The magnetometer is alive 

- The ACDS driver is alive 

4) Payload checkout 

18 


Re-iterate the OBC live streaming 

command: 

OBC 18 FF04 

Should receive acknowledge and see 

live streaming HK and AL data 

OK 

19 

20 

21 

22 

23 

Turn on MAGIC low power with 

telecommand: 

EPS 70 8F8F 

Turn on MAGIC high power with 

telecommand: 

EPS 70 AFAF 

Request pressure transducer 

readings from MAGIC with 

telecommand: 

MAGIC 81 XXXX 

Request temperature readings from 

MAGIC with telecommand: 

MAGIC 91 XXXX 

Turn off MAGIC high power with 

telecommand: 

EPS 70 BFBF 

Should see MAGIC TM received (two 

packets initially) and current 

consumption rise to 405mA 


rise to 700mA and then drop off back 

to 405mA over around 5 seconds 

Should receive TM from MAGIC with 

MID=1, length=6, then two bytes for 

each of the three transducers. Pressure 

should read close to expected internal 

pressure 

Should receive TM from MAGIC with 

MID=11, length=4, then one byte for 

each of the four thermistors. 

Temperatures should be close to 

expected internal temperatures 

The current consumption will fluctuate 

for a few seconds 

mA 

OK 

mA 

OK 

Bar 

Bar 

Bar 

OK 

Deg C 

Deg C 

Deg C 

Deg C 

OK 

OK 

23 

Turn off MAGIC low power with 

telecommand: 

EPS 70 9F9F 

Should see current consumption drop 

back to where it was before step 19 

mA 

OK 

44






24 

Turn on CAM with telecommand: 

EPS 70 4F4F 

Should see current consumption rise 

by 13mA 

mA 

OK 

25 

Send ping to CAM with 

telecommand: 

CAM 50 XXXX 

Should see pong in TM as a packet 

from CAM with the ASCII translation 

“CAM” 

OK 

26 

Turn CAM back off with 

telecommand: 

EPS 70 5F5F 

Should see current drop back to where 

it was before step 24 

mA 

OK 

27 

28 

Command S-Band into 

telemetry mode with 

DTMF tones “B*”, and 

ensure that the DTMF 

software is running with 

the audio line connected. 

The ICOM should be using 

a narrow filter. 

Command S-Band into 

transponder mode with DTMF 

tones “BA” 

Wait for DTMF tone 

telemetry, once every two 

minutes. Power should be 

around 1.6 watts, and 

temperature should be 

ambient to start with but 

rise as S-band carrier is 

used. Record all info from 

three bursts 

Carrier will jump up if control tones 

are on 

1 st 2 nd 3 rd Deg C 

W 

RSSI 

OK 

OK 

29 

Using control tones on from a 

handheld UHF transceiver, 

transmit audio 

S-band carrier should come up and 

repeat audio via groundstation 

OK 

30 

31 

32 

33 

Command S-band carrier up using 

DTMF tones “B0” 

If possible measure S-band 

frequency using frequency counter 

and patch antenna 

Command S-Band carrier down 

using DTMF tones “B0” 

Request a small thumbnail via 

UHF using telecommand: 

OBC 0D FF04 

Signal should rise and current 

consumption reach around 730mA 

Frequency should be very close to 

2401.84 MHz 

Signal should drop and current should 

return to 357mA 

Current consumption should rise for a 

minute or two to approximately 

780mA and 320 packets should be 

recieved 

mA 

OK 

MHz 

OK 

mA 

OK 

mA 

packets 

OK 

45






34 

During transmission from step 33, 

measure UHF frequency using 

frequency counter and patch 

antenna 

Frequency should be very close to 

437.250 MHz 

MHz 

OK 

5) Power down 

35 


Command the OBC to shut down 

using telecommand: 

OBC 15 XXXX 

Should see the acknowledge 

OK 

36 

At least 5 seconds, but not more 

than 20 seconds, after step 35, 

disconnect power source 

Current consumption should go to zero 

and all telemetry ceases 

OK 

37 

38 

Save the log of the term.exe 

session to a file called “YYYY- 

MM-DD_HH-MM_ISC.txt” and 

this filled in report to “YYYY- 

MM-DD_HH-MM_ISC.doc” and 

upload to the FTP in the AIV 

folder 

Close all relevant applications and 

turn off the groundstation 

6) Charging the spacecraft batteries 

There are two possible ways of charging the spacecraft batteries: 

Via the battery charge stud 

The spacecraft must be turned OFF. 

Set up the power supply and external battery charge regulator (EBCR) such that 28V is ready 

to be supplied to the EBCR, ensuring the polarity is the right way around. Connect the 

46






positive battery side of the EBCR to the battery charge stud, running it through an ammeter 

first, and the negative to the structure (ground). 

Connect a voltmeter across the battery side of the EBCR and note the voltage. If it is lower 

than 24.3V then it should be charged. 

Turn the power supply on. The voltage will raise slightly and the current consumption should 

be less than 600mA. 

When the battery consumption is less than 50mA turn off the power and check the voltage. If 

it is higher than 24.4V then the charging can be stopped. 

Trickle charging can continue indefinitely. The battery can be considered “full” when the 

consumption drops to 1 or 2mA. 

Via a solar panel simulation 

If the ABF is not already configured for solar panel simulation then power down the 

spacecraft 

The ABF should be configured such that: 

- The loops from pins 17 to 24 (from pins 4 to 11 respectively) are disconnected 

- The positive plug of the current source should be connected to pin 4 of the 

ABF, and the negative to ground. 

The spacecraft must be turned ON. 

The large Keithly current source should be set to 700mA and 28 volts. 

Turn the current source on. This should charge the battery while the spacecraft is running. 

Once the battery voltage is at the desired level the solar panel simulation can be turned off. 

The voltage can be measured across pin 2 of the EPS ABF and ground. 

47






Possible anomalies 

OBC / EPS unstable 

The presence of un-requested acknowledges with data like “02 70 6F 6F” signify unwanted 

OBC reboots, as do increments of the OBC boot counter. 

This is almost certainly due to lingering software bugs and is probably not due to current test 

circumstances. Just restart live streaming (OBC 18 FF04) if you were using it, send a time 

synch, and carry on – noting the time and boot attempt number. 

Battery charge under 90% 

This will cause the satellite to remain in safe-mode upon power-up. If the battery is not being 

charged then after approximately 12 minutes it will be assumed as broken and the spacecraft 

will proceed to recovery mode and ignore the battery voltage for the rest of the run (no safe 

mode entry). 

In this mode the OBC / EPS combination is known to be a little unstable. 

48






3 General notes 

Started on 05 th October 2004 

(previous work written from memory and with the power of hindsight) 

2 nd August 2004 

ARRIVAL 0: OBC team arrive with bags of components, empty PCBs and a car-full of 

equipment. Started soldering and testing EM in offices. 

9 th August 2004 

OBC complete the EM 

ARRIVAL: Solar panels get taken of the Nuna-II solar car and arrive in office 

Initiation to cleanroom, introduction to Jason, transfer of soldering work of OBC to 

cleanroom with tutoring from Jason. 

49







ARRIVAL 1: Arrival of EPS_Fulvio with software prototype board, start to iterate software 

and test with OBC EM 


ARRIVAL 2: Arrival of test UHF modems, loaned to OBC team for testing. OBC start 

soldering FM. 

24th August 2004 

MILESTONE 1: EPS prototype and OBC EM successfully progress through 

operational modes, boot-up and watchdog sequences. 

50







EPS software largely complete. OBC flight model 60% finished. 

DEPARTURE 1: EPS_Fulvio goes home, taking prototype with him. 

DEPARTURE 2: OBC team go home with OBC EM and test modems. 

8 th September 2004 

ARRIVAL 3: EPS_Tommasso arrives in ESTEC with EPS PSU components and empty 

PCBs. Introduction to cleanroom, starts soldering with tutoring from Jason. 


OBC_Karl arrives in ESTEC to continue work on OBC FM in cleanroom. 


ARRIVAL 4: PIN arrives in ESTEC. Visual inspection ok, minor required modifications 

identified. 

51







ESA_Neil buzzes out the PIN. Pin-outs all correct and joints positive. 


OBC_Mike arrives in ESTEC for “emergency” visit and starts software debugging with 

OBC_Karl. Software at “3 out of 500”. 

ARRIVAL 5: CAM arrives in ESTEC. Visual inspection good, minor alterations required. 

ESA_Neil takes the box apart and cleans it (see procedures). 

ARRIVAL 6: ACDS MAGNETOMETER arrives in ESTEC. Visual inspection ok (COTS), 

connector requires replacing and paint probably needs coating. 

ARRIVAL 7: ACDS MAGNETS arrives in ESTEC. Visual inspection ok. 

ARRIVAL 8: Magic EM arrives in ESTEC. Visual inspection fie but testing impossible due 

to OBC not ready 

CAM flight connectors undergo one cycle total (insert of savers in Denmark, removal of 

savers in ESTEC). 

All EPS flight hardware apart from BCR arrives in ESTEC in big metal drum. 

52







OBC_Mike leaves ESTEC. Software at “70 out of 500” 

23 rd September 2004 

ESA_Neil, ESA_Jason and EPS_Tommasso make decision that the EPS boards are not fit to 

fly. Tommasso begins to redo schematics and layouts while ESA_Jason finds supplier and 

ESA_Neil chases authorisation. 

LESSON LEARNED 1: Hold courses on flight PCBs before the start of hardware 

manufacture. 

ARRIVAL 9: Both titanium rings, DMM base-plate, DMM top-plate and all small DMM 

shear panels arrive in ESTEC. Visual inspection good. ESA_Neil and ESA_Marie clean with 

vacuum cleaner and IPA. 

ARRIVAL 10: Inserts arrive from STRU_Melro. 

53







OBC_ Karl leaves ESTEC. OBC FM 99% complete (missing thermistors and prom). 

ESA_Neil arranges for side-pockets to be cut into base-plate in mechanical workshop. 


ARRIVAL 11: New EPS boards arrive in ESTEC. EPS_Tommasso starts to populate them. 

ARRIVAL 12: Aluminium brackets, screws, nuts and bolts arrive from WestendBV. 

ESA_Neil cleans them all with IPA / ultrasound. 

SYS_Joerg comes to ESTEC. 

54







MILESTONE 2: The first Cubesat arrives in ESTEC. 

ARRIVAL 13: XI-V_Yuya arrives with the engineering model of Xi-V Cubesat. The 

engineering model is labelled Xi-III. Lengthy interface discussions with ESA_Neil. 


ARRIVAL 14: DMM shear panels +y and –y arrive in ESTEC. ESA_Neil and ESA_Marie 

clean them for insert potting. 

ARRIVAL 15: ESA_Neil and ESA_Marie go and pick up Araldite AY103 and Hardener 

HY991 from Viba. 

Morning, ESA_Neil and ESA_Marie define procedure and commence central insert potting 

on all small shear panels apart from –x0y (S1) 

Afternoon, ESA_Neil and ESA_Marie continue central insert potting on –x0y and Base-plate 

(S2) 

1 st October 2004 

ESA_Neil and ESA_Marie continue central insert potting with plates –y. +y and top-plate (S3 

and S4). 

ESA_Neil gets workshop to manufacture side insert “masks” to aid with side insert potting. 

55






2 nd October 2004 

EPS_Tommasso leaves ESTEC. EPS PSU and TIMERS 95% complete. PDU incomplete. 

DEPART 3: EPS_Tommasso takes EM hardware with him to Naples 

ESA_Neil and SYS_Joerg cut all other side pockets in DMM panels. 

3 rd October 2004 

ESA_Neil and SYS_Joerg define side-insert potting procedure and commence side insert 

potting on all small shear panels and the +x sides of the –y, +y and top plates (S5). 

PROBLEM 1: In panel –x+y a mistake was made in marking and cutting of side pockets, 

four of the pockets are 7mm misplaced. 

MODIFICATION 1: The offending pockets are ground larger 

56






PROBLEM 2: In panel +x+y the strip-mounted side inserts do not fit properly. 

MODIFICATION 2: To fix the above the panel is re-ground until the strip sits flush with 

the reference point. 

4 th October 2004 

OBC_Mike reports via IRC that OBC software at “400 out of 500” 

ARRIVAL 16: ACDS COILS arrive in ESTEC. Visual inspection ok, requires some kind of 

coating. 

ESA_Neil performs insert potting on +x side of base plate and tops up all S5 potting. 

SYS_Joerg works on thermal simulation. 


ESA_Neil performs insert potting on –y sides of base-plate and top-plate and on –x sides of 

+y and –y panels (S7). 

SYS_Joerg works on thermal simulation 

MODIFICATION 3: ESA_Jason touches up the PIN box (few solder joints, cable ties, etc.) 

ESA_Neil meets with Børre Pedersen of K-SAT. Usage of Svalsat for SSETI Express still 

looks possible. 

LOG DONE IN REAL TIME FROM HERE 


ESA_Neil and SYS_Joerg work on DMM honeycomb panels: 

Panel –x-y 

- Removed all kapton 

- PROBLEM 3: Paper stuck to honeycomb around side pockets! 

57






- MODIFICAITON 4: Use scalpel to carefully remove excess glue, results in 

some scraping of aluminium skins and is not recommend for flight model. 

Damage is superficial but looks awful. 

58






- PROCEDURE: Change design of “tie-fighters” so that kapton covers paper 

completely and gap between insert and kapton is larger 

- PROCEDURE: After any gluing session inspect all relevant panel skins for 

excess glue 

- PROBLEM 4: Some holes not filled completely with glue 

- PROCEDURE: Top-up glue in a side insert if the existing glue does not come at 

least halfway up the insert on both sides 

- Tested with mounting strip and bolts: POSTIVE 

59






- Two inserts require filling 

Panel +x+y 


- Panel better than –x-y 

- Some inserts with air bubbles at surface (not critical) 

- See procedure point above for standard 


- Panel passes 

- It is cleaned and put into cleanroom 

Panel -x+y 



- PROBLEM 5: One large drip of glue was missed and had run all the way down 

the side and stuck to the cardboard beneath (arg!) 

60






- MODIFICATION 4b: Glue taken off (in order to leave a smooth surface) using 

a heat gun and a scalpel. Result is ugly, but damage is superficial and panel is 

smooth again. 

- Panel passes 

- It is cleaned and put into cleanroom 

Panel +x-y 



- Some paper removed with scalpel as other panels 

- One insert requires topping-up 

SYS_Joerg manufactures new “mark 2 tie-fighters” for all remaining side inserts: 

61






ESA_Neil mixes batch 8 of glue 

ESA_Neil tests extended vacuum exposure by leaving mixture at approx. 25 Pa for 15 

minutes. 

62






Result is indistinguishable from shorter exposures, and many air bubbles still present. 

However, result can hardly be worse than original, so we proceed to use it. 

ESA_Neil and SYS_Joerg proceed to use batch 8 to glue side inserts: 

- Side inserts in the +y side of the base plate 

- Side inserts in the +y side of the top plate 

- Top-up offending inserts in –x+y plate 

- Top-up offending inserts in –x-y plate 

- Top-up offending inserts in -y plate 

- Top-up offending inserts in +y plate 

ARRIVAL 17: Flight structure (except plate 16) arrives in ESTEC 

63






64






SYS_Joerg and ESA_Neil plan the coming days and discuss design of mass / CoG 

“mushroom” with ESA_Marcel 

ESA_Jason finishes PIN and looks at CAM. 

ESA_Neil and SYS_Joerg clean the FM baseplate, the FM +y plate, the FM –x+y plate and 

the FM +x+y plate. 

65






SYS_Joerg marks the side pockets on the FM baseplate and FM +y plate. 

ESA_Neil and SYS_Joerg cut the side pockets out with a scalpel in the FM baseplate and FM 

+y plate. 

ESA_Neil grinds the side pockets smooth and performs fit tests with mounting strips. 

66






ESA_Neil cleans the panels after grinding. 

SYS_Joerg places main central inserts into baseplate and use kapton tape to secure them. 

ESA_Neil tapes in the activation switch plate insert to the FM baseplate. 

ESA_Neil and SYS_Joerg give up for the night since it is 2:30am. 


ESA_Neil mixes batch 9 of the glue using 133g of AY103 and 53.3g of HY991, as per 

procedure. 

ARRIVAL 18: Lifting frame and panel protectors from Vienna. Initial inspection 

everything seems very good. Requires cleaning. 

67






SYS_Joerg prepares “tie-fighters” and strips for the DMM baseplate –x, the FM baseplate –x 

and the DMM top-plate –x sides. SYS_Joerg cleans and positions centre inserts for the FM 

+y panel. 

ESA_Neil glues side inserts into DMM baseplate –x, the FM baseplate –x and the DMM topplate 

–x sides. 

SYS_Joerg and ESA_Neil glue the majority of the centre inserts for the FM +y panel. 

PANEL WEIGHTS BEFORE GLUING (a bit late, but we subtract additions): 

FM Baseplate 

FM +y 

FM +x+y 

FM -x+y 

956g - ~30g glue – 4 x 4.64g side insert – 12 x 2.77g asap insert 

= 874g 

788g - ~30g glue – 29 x 1.76 central insert 

= 707g 

233g – 8 x 1.76 central insert 

= 219g 

224g – 3 x 1.76 central insert 

= 219g 

PROBLEM 6: Accidentally left glue too long with ‘excess blobs’ on the top, they hardened 

a lot. Also, the remains of the glue in the mixing pot hardened a lot. So all of S9 inserts have 

a problem. 

ATTEMPT: Clean pot with hot water, knife (we broke it), then IPA, then acetone. FAIL 

ATTEMPT: Cut off the excess glue from the top of an insert with a scalpel, tested on one of 

the –x side inserts on the DMM baseplate. FAIL – too hard, rubbery and sticky 

68






ATTEMPT: Heat gently with heat gun whilst cutting away with a scalpel, tested on one of 

the –x side inserts on the DMM baseplate. FAIL – doesn’t help much 

ATTEMPT: Try to remove kapton tape from inserts before the excess becomes too rigid to 

do so, tested on one of the –x side inserts of the DMM baseplate. FAIL – pulls too much on 

the glue around the insert, not recommended for flight model 

ATTEMPT: Cut away the excess glue just around the injection and air holes, tested on one 

of the –x side inserts of the DMM baseplate. POTENTIAL SUCCESS – hopefully will 

enable the removal of the kapton tomorrow when the glue is hard enough not to disturb the 

insert – requires a further test 

ATTEMPT: Remove kapton from centre inserts before the excess glue is too rigid to do so, 

tests on one of the centre inserts of FM +y panel. SUCCESS 

MODIFICATION 5: Cut away the excess glue just around the injection and air holes on all 

the –x side inserts of the DMM baseplate, the –x side inserts of the DMM top-plate and the –x 

side inserts of the FM baseplate. Hopefully will be able to remove kapton with the excess 

glue tomorrow. 

69






MODIFICATION 6: Remove kapton from all potted centre inserts of FM +y panel, rekapton 

and top-up as necessary 

PROCEDURE: Define maximum of four hours after mixing before the excess glue is wiped 

away from the inserts. 

ESA_Neil mixes batch 10 of the glue using 100g of AY103 and 40g of HY991. Since the 

mixing pot is ruined two plastic disposable coffee cups are used. They are cleaned with IPA 

directly before, and half of the glue is tipped from one to the other. During vacuuming the 

froth decreases dramatically, and the glue is much better than usual. 

ESA_Neil and SYS_Joerg use batch 10 for: 

- The centre inserts of the FM +x+y panel 

- The centre inserts of the FM -x+y panel 

- The remaining centre inserts of the FM +y panel 

- Topping up the S9 inserts 

NOTE: Glue much better – it does not slump at all on the majority of inserts. 

POTENTIAL PROCEDURE: Always prepare glue in this manner! 

ESA_Neil and SYS_Joerg give up and go home as it is 01:30am. 


ATTEMPT: As defined yesterday attempt to remove (early) kapton and excess glue from –x 

side of the DMM baseplate. Attempt successful. 

70






MODIFICATION 7: Remove all kapton and excess glue from –x sides of DMM baseplate, 

DMM top-plate and FM top-plate 

ESA_Neil cleans all DMM honeycomb panels and labels them. 

MILESTONE 3: Successful strip / mask tests on all DMM panels. 

71






PROCEDURE: When cleaning panels noticed that some paper had still stuck with glue 

(minor correction with scalpel), therefore consider instead using “upside down” kapton to 

make the ‘wings’ of the tie-fighters. (Mark III fighters.) 

SYS_Joerg cleans inserts and prepares mark III tie-fighters and strips for next batch of gluing. 

ARRIVAL 19: The EM e-box arrives in ESTEC from Canada, visual inspection looks very 

good, will test with EPS as soon as possible. User manual on FTP. 

72






ARRIVAL 20: Two potential transport containers arrived in ESTEC, they are about the right 

size but need cleaning and have NO mounting points (so should look for Proba box still). 

ARRIVAL 21: Two integration pillars arrive in ESTEC. Extremely heavy and hard to 

move, requires a lot of cleaning, requires mounting adapter. 

At 2100 ESA_Neil mixes batches 11 and 12 of glue using 100g of AY103 and 40g of HY991 

for each. The process for batch 10 is mimicked with small test variations – none of them 

work as well as batch 10 did. 

ESA_Neil and SYS_Joerg apply batches 11 and 12 to 

- side inserts on +x side of the FM baseplate 

- side inserts on +x side of FM +y plate (15) 

- side inserts on FM +x+y 

- side inserts on FM –x+y 

73






SYS_Joerg and ESA_Neil clean FM panels +x0y, -x0y, +x-y, -x-y and top-plate. 

SYS_Joerg marks side pockets in FM panels +x0y, -x0y, +x-y, -x-y and top-plate. 

SYS_Joerg and ESA_Neil cut the side pockets out, with a scalpel, from FM panels +x0y, - 

x0y, +x-y, -x-y and top-plate. 

ESA_Neil grinds the side pockets smooth. 

PANEL WEIGHTS BEFORE GLUING: 

FM Top-plate 

FM +x0y 

FM –x0y 

FM +x-y 

FM –x-y 

512g 

234g 

232g 

217g 

218g 

SYS_Joerg and ESA_Neil give up and go home as it is 00:30am. 


10:00 ESA_Neil mixes batch 13 of the glue using 100g of AY103 and 40g of HY991. It is 

used to glue: 

- side inserts on +y side of the FM baseplate 

- side inserts on -x side of FM +y plate (15) 

- side inserts on FM +x-y 

- side inserts on FM –x-y 

PROBLEM 7: Just after gluing ESA_Neil notices that the inserts in the –x side of the +y 

shear panel are the wrong way around. SYS_Joerg inspects and agrees. 

MODIFICATION 8: Inserts are carefully removed from the –x side of the +_y shear panel 

while the glue is still liquid. Inserts are cleaned by SYS_Joerg. ESA_Neil uses small strips 

of tissue to remove some glue from the side pockets by wicking. SYS_Joerg prepares tiefighters 

and the mounting strip (8) to re-pot the inserts. ESA_Neil applies kapton to protect 

the panel from excess overflowing glue. SYS_Joerg re-applies mounting strip and ESA_Neil 

secures strip and inserts in place and then removes mounting strip. Gluing is redone and 

appears unproblematic. 

74






PROCEDURE: Double check BEFORE you glue! (Obvious really…) 

LESSON LEARNED 2: All inserts should be symmetrical in terms of mounting points 

75






LESSON LEARNED 3: All structural items should be symmetrical as far as possible. In 

the case where they are not symmetrical they should be obviously or explicitly not 

symmetrical. 

21:00, ESA_Neil mixes glue batch #14, using 100g of AY103 and 40g of HY991. 

SYS_Joerg prepares mounting strips. 

SYS_Joerg and ESA_Neil use batch #14 to glue side inserts into the –y side of the FM 

baseplate and the –x side of the FM top-plate. 

MATERIALS: Low on kapton tape, low on IPA tissues 

PROBLEM 8: SYS_Joerg counted side inserts and there are not enough, we need to order 4 

more at least (or make in workshop here) 

PROBLEM 9: Rest of centre insert potting cannot proceed until ordered inserts have been 

delivered (before Wednesday we hope) 

SYS_Joerg and ESA_Neil enter cleanroom to perform integration of engineering model of 

primary structure. 

1) We start by connecting the base brackets to the titanium ring using M4x16mm bolts 

PROBLEM 10: The integration manual specifies M4x16mm bolts, but the holes are M5. 

MODIFICATION 9: The manual is wrong; there was a typo in the CATIA model. 

Subsequently the wrong types of bolts were ordered, we must order some more. Luckily 

we already have M5x16mm bolts for other purposes in the spacecraft, so we use them. 

76






2) Since there is no engineering model of the separation ring available we simply proceed 

using four M6x45mm bolts, with large washers, to hold the baseplate to the basebrackets 

and titanium ring 

3) The assembly so far rests upon the shaker adapter plate so as not to damage the table 

(could not do flight model like this, but the primary structure alone is light enough to 

not require the integration table.) It corresponds nicely to the integration manual. 

4) We try to insert a shear panel (+y) to the base brackets. The fit is VERY tight and 

scratches the surface of the panels. However, the marks are superficial and with some 

persistent (but careful) coercion the panel does slide into place. We find this with all 

the shear panels. 

77






5) We insert the other two +y small shear panels. The manual does not specify which 

way around to put the bolts, we define the convention to always have the bolt heads in 

the outer corner on the base-brackets. 

6) We mount the two applicable mid-height brackets into the 0x+y compartment 

78






7) We notice a gap of a few millimetres between the panels. This is probably just a result 

of tolerance combinations and should not be a problem since the brackets are holding 

it very firmly. (Later loosening the bolts and slightly adjusting the position equalised 

gaps like this.) 

8) We mount the two top-brackets (including the special T-Pod one) into the –x+y and 

+x+y compartments 

79






9) We mount the +x0y panel and it’s brackets onto the existing structure 

80






10) We mount the –y shear panel to the existing structure with no problem, fastening it 

securely to the base brackets and the mid-height bracket on the +x0y panel 

11) We attempt to mount the +x-y shear panel but have a lot of difficulty getting it into 

place since it is a very tight fit. To start with we cannot align the hole with the base 

bracket. The same problem is encountered with the –x-y shear panel. 

12) By loosening the other bolts on the connecting panels and adjusting their positioning 

slightly it becomes possible to put the –y small shear panels into place, but it takes a 

considerable amount of force, we have to be very careful. 

81






PROCEDURE: Do NOT tighten the bolts as you go, just leave them ‘hanging’ in place 

but not tight at all, don’t even make them “finger tight” (as we were doing). This will 

allow the structure to be ‘flexible’ enough to avoid problems of combined tolerances 

making integration difficult. Once the whole structure is in place the bolts can then be 

tightened to secure it. 

13) We mount the mid-height brackets to support these panels 

14) We slide the –x0y panel into its place and attempt to mount the brackets that hold it 

there. 

PROBLEM 11: The integration manual specifies three M5x16 bolts only for the midheight 

bracket #12, however, the insert in the –y panel (and the drawings that led us to put 

the insert there) is an M4 insert. 

82






MODIFICATION 10: We therefore use an M4 bolt instead of an M5. This needs to be 

checked by STRU to see whether or not it is acceptable. 

PROBLEM 12: The bracket that shares a mounting point with the –x+y top-bracket is 

mis-aligned with its bolt hole in the –x0y shear panel by a few millimetres. Somehow 

both the panel and the bracket appear to be consistent with their drawings. 

MODIFICATION 11: We carry on without fixing this bracket to the –x0y panel. We 

ask STRU to check where this error comes from. A possible solution would be to modify 

the bracket slightly by drilling a hole in the appropriate place. We ask STRU is this will 

be acceptable (the bolt would then be able to slide sideways a little in the –y direction). 

15) We add the final top-brackets 

83






16) We add the top-plate using M6x30mm bolts 

17) In our joy and elation we take a moment to verify STRU_Melro’s claim that the 

structure could be assembled by a pair of blind monkeys 

84






18) We add the lifting frame to the top of the satellite with no problem 

19) We mount the Perspex side protectors, making sure that the one with the access hole 

lines up with the location of the flight-preparation-panel. They seem sturdy and 

strong. 

PROBLEM 13: There is a slight mis-alignment of the ‘lower-right’ screw on each of the 

protectors, such that the bolts could be fixed, but it puts unnecessary torque on the 

mounting points (side inserts). 

REQUIRED MODIFICATION: We need to slightly move the position of one of the 

holes (lower-right) in three (+x,+y,-x) of the side protectors. This should be easy though 

20) We do a quick fit-test with the camera, it looks fine 

85






21) We congratulate each other, tidy up the cleanroom and go home (it is 3am) 

MILESTONE 4: Completion of EM primary structure 


16:00 ESA_Neil mixes batch 15 of the glue using 200g of AY103 and 80g of HY991. 

ESA_Neil and SYS_Joerg use it to glue the –y side of the FM top-plate and the centre inserts 

(not thrusters) of the FM baseplate. 

PROBLEM 14: Some glue in the central holes of the ASAP inserts 

86






REQUIRED MODIFICATION: Will cut kapton and clean through with IPA when ‘mop 

up’ the excess glue. Otherwise can grind it out later (tested grinder on sample and it works 

fine.) 


19:30 ESA_Neil mixes batch #16 of the glue in ratio of 45:18g. 

SYS_Joerg prepares tie fighters for the side inserts of the +x side of the FM top-plate. 

ESA_Neil and SYS_Joerg glue in the side inserts of the +x side of the FM top-plate. 

ESA_Neil and SYS_Joerg inspect the base-plate and decide that grinding the glue out later is 

safer than removing the kapton on semi-set inserts, and therefore leave the plate as it is. 

SYS_Joerg decides that we need to cover the inside of the satellite in black thermal paint. 

ESA_Neil tracks down black thermal paint and finds some local friendly Royal Marines to 

spray it for us. 


ARRIVAL 22: The engineering models of the separation rings (two) arrive from SSTL. 

ESA_Neil discusses and arranges the manufacture of the missing side-inserts and the adapter 

plate (between the integration table and the satellite) with ESA_Marcel. 

87






ARRIVAL 23: The spare transport boxes arrive in the office. They are only just big enough 

and have no mounting points. They may be ok for fit-check dummy, but no good for flight 

model. 

ESA_Neil negotiates usage of PROBA container. 


ARRIVAL 24: Corner profiles from Westend 

ARRIVAL 25: Lateral panels from Westend 

ARRIVAL 26: All remaining inserts and the passive magnet housing from Westend 

88






ESA_Neil mixes batch 17 of the glue using a ratio of 100:40g. 

SYS_Joerg cleans appropriate inserts. 

ESA_Marie tapes inserts into appropriate panels. 

ESA_Neil, SYS_Joerg and ESA_Marie apply the glue to the bottom layer of the 10 inserts for 

the thruster mounting in the FM baseplate, and all the central inserts of the FM top-plate. 

PROBLEM 15: The thruster mounting plates have one bolt hole too large for the associated 

bolt (the one over the thruster insert should be M4 instead of M5. 

SOLUTION: ESA_Neil and ESA_Marcel quickly manufacture two new plates with the 

correct mounting pattern. 

ESA_Neil mixes batch 18 of the glue using 120:48g. 

ESA_Marie and ESA_Neil apply the glue to all the centre inserts of the FM –x0y, +x0y, +x-y, 

+x+y panels. 

TEST: ESA_Neil performs a simple vacuum test on the ‘dregs’ (full of air bubbles, from the 

froth at the top of the batch) of batches 4 and 5. Sample is taken down to 25mb over five 

minutes, sustained for five minutes, and then returned to atmospheric pressure. No 

discernible damage. Test positive. 

89







ARRIVAL 27: PROBA box. It is huge and heavy (140kg) and it has the wrong mounting 

pattern in it. 

ESA_Dave, ESA_Neil, ESA_Marie remove PROBA SQM from the box and store it in the 

high-bay of Erasmus. 

ESA_Marie arranges cleaning of the box. 

ESA_Neil arranges manufacture of adapter plate with ESA_Marcel. 

90






ARRIVAL 28: Final panels from ISF, including FM –y panel, spare large shear panel, spare 

baseplate and spare top-plate. Visual inspection ok apart from minor clamp damage on +x 

side of +y surface at approximately one third of the +z extent. 

ESA_Neil and ESA_Marie mark, cut and grind the side pockets into the –y panel. 

PROBLEM 16: Uppermost (+z) side pocket on the –x side of the –y panel is originally cut 

7mm too high by mistake. 

MODIFICATION 12: As with the precedent set on the engineering model correction is 

made by enlargement of the pocket in the right direction. 


ESA_Neil prepares mounting masks for the five inserts of each of the two thruster clusters. 

The masks are entirely covered with a protective layer of kapton on the contact side in order 

to prevent accidental gluing of the mask to the baseplate. 

91






Small bolt holes are cut into the kapton to prevent getting any into the threads, and then the 

inserts are bolted to the mask lightly, position in the base-plate, and then secured firmly in 

place on the mask. 

Simple paper ‘depth testers’ are prepared so that we can tell if the glue in the holes is deep 

enough 

ESA_Neil mixes batch 19 of the glue using a ratio of 200:80. 

Glue is then applied to the holes until it covers the bottom millimetre or two of the depth 

testers (which are only temporarily applied). The point is to have enough glue that the inserts 

are held in place by their bases, but not so much glue that it flows onto the surface of the 

baseplate. We intend to top-up the holes later once the inserts are secured in place. 

The reason for not using the “tie-fighter method” (like the side inserts) is because the ‘wings’ 

of the tie-fighters would have to be impractically long to extend beyond the mounting mask – 

therefore they would not provide a secured fixing of the insert positions. 

The mounting masks and inserts can then be placed into position and left for the glue to dry. 

92






The rest of glue batch 19 is also used to pot the centre inserts into the FM –y panel. 

ESA_Neil and ESA_Marie remove all kapton tape from FM panels +y, -x-y, +x,-y, -x0y, 

+x0y, -x+y, +x+y and the top-plate. 

PROBLEM 17: For most of the FM we have been using thinner (width) kapton tape and two 

strips were needed to cover each insert. Glue has run down the join in between each pair of 

kapton strips, on TOP of the panels. 

SOLUTION: It is decided that such line of glue on the surface will almost certainly be 

removed when we scrub the surface abrasively to prepare it for the application of thermal 

paint. In the event that this scrubbing is not sufficient to remove the glue then the precedent 

from the engineering model of using a scalpel (possibly in conjunction with a heat gun) to 

remove the glue will be employed. 

LESSON LEARNED 4: When insert potting no joins between different strips of kapton tape 

should be accessible to the glue being used. (I.e. use tape that is wide enough for the job.) 

93






ARRIVAL 29: The remaining nuts, bolts and washers from Westend BV. 

ARRIVAL 30: The remaining side inserts from ESA_Marcel. 

MODIFICATION 13: For ease of manufacture it was decided not to have injection holes in 

these new side inserts, since they are not really used in the potting procedure anyway. This 

will give a slightly reduced mechanical strength, which, although negligible, should be 

reflected by placing these inserts in positions where they bear the smallest loads. 

ARRIVAL 31 : The magnetorquer coil clamps from Westend BV. Visual inspection looks 

fine. 

ARRIVAL 32: The adapter plate from ESA_Marcel. This plate has multiple footprints to 

interface with the lower side of the ASAP ring, the top of the integration table, the mounting 

frame in the PROBA transport container, and the activation pillar. 

94







ESA_Neil mixes batch #20 of the glue, using 50g of AY103 and 20g of HY991. 

ESA_Marie and RANDOM_Fred prepare tie-fighters using a mixture of old and new style 

side-inserts. (New ones in central positions.) 

ESA_Marie, ESA_Neil and RANDOM_Fred pot the side inserts into the +x side of the FM –y 

panel. Then we clean the integration table, the adapter plate and the second engineering 

dassault ring and take them into cleanroom, then clean all relevant bolts. 

ESA_Neil, ESA_Marie and RANDOM_Fred secure the adapter plate to the integration table, 

and the engineering ring to the adapter plate. Then we remove the temporary bolts from the 

bottom of the EM structure, and mount it to the engineering ring. 

95






PROBLEM 18: Several of the bolts are extremely tight and difficult to get into place. 

Several iterations of loosening and tightening all bolts are required. This is NOT 

recommended for flight model, as it puts unnecessary strain on various threaded items (such 

as the base-brackets). 

LESSON LEARNED 5: All flight bolts should be cleaned VERY thoroughly, using dry 

tissues, then IPA, then an ultrasound bath, then acetone. 

ESA_Neil, ESA_Marie and RANDOM_Fred apply the second engineering ring to the large 

shaker adapter plate, ready for integration of the FM structure. 

96






ESA_Marie prepares tie-fighters for +x side FM –y panel. 

The excess glue is cleaned off of the side inserts before we all go home. 


ESA_Neil mixes batch #21 of the glue using 100g of AY103 and 40g of HY991. 

ESA_Neil mounts tie-fighters to the +x side of the FM –y panel. 

ESA_Neil attempts to remove the mounting plates from the thruster inserts on the FM 

baseplate. 

PROBLEM 19: On the –x thruster insert cluster some glue has run onto the top of the panel 

from one of the inserts. 

SOLUTION: Careful use of a scalpel, as per engineering model precedent, is employed to 

clean the glue off of the surface. 

PROBLEM 20: The mounting plate on the +x thruster insert cluster has become stuck to the 

base-plate and cannot be removed with bare hands. 

97






SOLUTION: Did not want to lever the mounting plate against the skin, as it could easily 

damage the ply. It was therefore decided that the best way to remove it would be a rotation 

about the z-axis and in the x/y plane, as this poses the lowest potential damage to the 

baseplate. (It flexes in this direction slightly anyway.) The mounting plate was carefully 

secured in a small vice (without touching the baseplate), and then a force was applied to rotate 

the vice and the baseplate in opposite directions. The additional strength and leverage 

provided by the vice allowed a relatively simple removal of the plate with no apparent 

damage to the baseplate – most of the force is actually absorbed by the tearing of the kapton 

tape adhesive from both surfaces. 

98






PROBLEM 21: Some glue is left on the surface of the baseplate. 

SOLUTION: As per precedent this glue is removed carefully with a scalpel. In order to then 

investigate the final surface properties of the panel a very light sandpaper is used to remove 

the oxide layer in the area. The results look fine and seem to take the surface down to the 

same ‘level’ as the use of the scapel. The final result will therefore not have suffered any ill 

effects from the use of the scalpel. 

LESSON LEARNED 6: A better procedure should be developed for this process to ensure 

that the initial glue depth is insufficient to result in flow of glue on the surface of the skin. 

ESA_Neil covers the inserts in kapton tape again and applies the glue from batch 21 to the 

thruster inserts and to the side inserts in the +x side of the FM –y panel. 

99






PROBLEM 22: The bolts used to secure the separation ring to the EM structure (same as 

will be used for FM) are actually 5mm longer than planned. This could potentially interfere 

with the propulsion tubing. Must check this with PROP! 


ESA_Neil Removed kapton from centre (ASAP) inserts of FM baseplate. 

MODIFICATION 14: ESA_Neil ground the excess glue out of the holes in the ASAP 

inserts on the FM baseplate. 

ESA_Neil cleans workshop and tidies area. 


ARRIVAL 33: The FM Magic box arrives in ESTEC. Visual inspection fine, several small 

modifications by Jason required. 

100






PROBLEM 23: Westend BV discovered that they used alloy 6082 in the manufacture of all 

recent inserts! (Instead of 7075.) ESA_Neil asks STRU_Antonio for advice on this issue. 

ESA_Neil, ESA_Marie and SYS_Joerg meet with Andy Currie of SSTL to discuss schedule, 

fit-check and launch campaign. 

MODIFICATION REQUIRED: Need to make shock absorbers in PROBA box, as advisd 

(strongly) by Andy Currie of SSTL. 


PROBLEM 24: STRU_Melro advises that the inserts made with alu 6082 are not suitable 

for the flight model as the strength is almost 50% less than that of 7075. This renders “FM” – 

y, -x0y, +x0y, -x-y and +x-y panels as unusable! (Later decided that –x-y and +x-y are ok as 

they carry little load.) 

MODIFICAION 15: After extensive discussion ESA_Neil, SYS_Joerg, ESA_Marie and 

STRU_Melro decide to upgrade the –y, -x0y, and +x0y panels of the Engineering Model to 

flight hardware, and downgrade the –y, -x0y, and +x0y panels of the Flight Model to 

engineering hardware. 

ESA_Neil starts to negotiate with Westend BV to replace panels ruined by incorrect inserts. 

ARRIVAL 34: UWE-1 mass dummy. Visual inspection: fine. 

101







ESA_Neil and ESA_Marie clean a test panel ready to be sprayed. This involves removing 

kapton, scrubbing with wire wool, de-greasing with acetone, then cleaning with IPA tissues, 

before storing in a bag to prevent dust. Gloves worn at all times. 

The panel selected was the “FM” (pre-downgrade) –x0y, since it would no longer be used for 

the flight structure. 

The clean panel was then taken to Valkenburg airport, where cooperation with the Royal 

Netherlands Marine Korps procured a promise for them to spray the thermal coating onto our 

panels in their paint shop. 

ARRIVAL 35: New paint arrives from Map (base, hardener and thinner). Visual inspection 

looks fine, confirm is the same as “old” paint. 

102






ESA_Neil and ESA_Marie disintegrate the relevant parts of the engineering model and 

replace them with the downgraded “FM” panels. Specifically: 

Panel DMM –y becomes FM –y 

Panel DMM –x0y becomes FM –x0y 

Panel DMM +x0y becomes FM +x0y 

Panel FM –y becomes EM –y 

Panel FM –x0y becomes EM –x0y (actual panel is still at airport) 

Panel FM +x0y becomes EM +x0y 

ESA_Neil, ESA_Marie and ESA_Jason manufacture custom-sized polythene bags to store 

panels in. 

OBC_Karl mounts components to the FM ACDS magnetorquer driver PCB. 

ESA_Neil, ESA_Marie, OBC_Karl, SYS_Joerg and ESA_Iñaki perform the following tasks 

on all FM panels, including laterals and corner profiles: 

1) Remove kapton and excess glue 

2) Scrub the oxide layer off with “scotch bright” (3M) 

3) Dry-wipe dust from panel 

4) De-grease with acetone (clean tissue surface for each dip-and-wipe) 

5) Clean with IPA tissues (clean tissue surface for each wipe) 

103






6) Mask off all areas that should not be sprayed (this includes all threaded inserts, 

all side inserts, the propulsion tube mounting points, the titanium ring, all 

bracket-panel contacts, underneath the PDU, underneath the UHF, underneath 

Magi and underneath the S-band unit. 

104






7) Place in polythene bags 

ESA_Neil, ESA_Marie, OBC_Karl, SYS_Joerg and ESA_Iñaki go home to bed as it is 1am. 

22 nd October 2004 

08:00. ESA_Neil and ESA_Marie pick up panels and paint and transport them to Valkenburg 

military airport. 

ESA_Neil and ESA_Marie completely mask off all not-to-be-painted areas, and negotiate 

spray-painting to be done on the same day. 

ARRIVAL 36: Otto switches arrive from Westend BV. Visual inspection: they look fine – 

just like it says on the tin. 

MODIFICATION 16: ESA_Neil arranges with ESA_Marcel to have the erroneous holes 

slotted in the number #13 aluminium brackets, and to have four of the M6x60 bolts cut-down 

to 55mm in length so as not to interfere with PROP tubing. 

MILESTONE 5: FM OBC booted up for the first time 

OBC_Karl boots up the FM OBC and it works perfectly. 

Royal Dutch Marine (Charles) informs ESA_Neil that spray job is complete and went well. 

Panels to be picked up on Monday. 

PROBLEM 25: TCS thermistors do not have a power supply defined! 

105






MODIFICATION 17: OBC to provide one 3.3V power supply, one ground, and six signal 

wires (two redundant per thermistor), in the TCS connector. Power splitting is to be handled 

in special TCS harness. Pin-out: 

1 Thermistor 1 signal 



4 Empty 

5 Ground 

6 Thermistor 1 signal (redundant) 



9 Power supply 3.3V 

OBC_Karl adds the connectors to the computer. 

23 rd October 2004 

OBC_Karl puts the FM computer “in the box” and proceeds to test. 

PROBLEM 26: No external communications on the computer are working and the utility 

processor simply oscillates instead of exchanging proper CAN messages. 

TESTS: OBC_Karl: 

- Removes boards from box and verifies similar problem on three processors 

- Replaces the utility processor with another version (no difference) 

- MODIFICATION 18: OBC_Karl disconnects power supply between boards (by 

de-soldering a hook-up wire) and powers the computer directly in order to check 

that the power supply is stable (no difference) 

SOLUTION: The OBC has several internal CAN devices on the bus, and then connects 

externally to the Magic box (the only other CAN device on board). However, it is the Magic 

box that houses the termination of the CAN bus (128 ohm resistor), so, since the magic box is 

not plugged in, the devices on the un-terminated bus just talk gibberish to each other. (The 

EM had termination built in, since it was designed to work in a stand-alone context.) 

DECISION: Since the OBC cannot work at all without a CAN bus termination it is decided 

by ESA_Neil that the termination should be present in OBC and not rely on a connection to 

an external device. 

MODIFICATION 19: OBC_Karl adds an appropriate resistor to terminate the CAN bus and 

re-solders the hook-up wire between the two boards. The system then functions normally. 

106






REQUIRED MODIFICATION: We need to remove the CAN bus termination in the Magic 

box, as it can’t be terminated twice. 

OBC_Karl adds a temporary reset wire to pin 7 of the power connector (flight PROM will 

remove the need for this) and reassembles computer in box. 

ESA_Neil integrates the top s-band patch antenna mounting plate to the top plate as a simple 

fit check. No problems encountered. 

ESA_Neil integrates the PIN to the EM structure as a simple fit check. The mounting holes 

on the PIN are too large for the M4 bolts intended to hold it in place. Suggest either to have 

small stand-offs to fill the holes, or do bolts up tighter than usual (system is not fragile). No 

other problems encountered. 

107






ESA_Neil cleans all bolts, nuts, washers, rivet nuts and coil clamps, storing each type in 

individual labelled bags. This is to get ready for FM structural integration tomorrow morning. 

OBC_Karl glues wires and heavy components down on the ACDS magnetorquer coil driver 

and the FM OBC. 


ARRIVAL 37: EPS_Fulvio arrives to complete the EPS subsystem with FM PDU boards 

and components. 

PROBLEM 27: EPS_Fulvio does not bring the Battery Charge Regulator, as it is still not 

finished. Must chase this up tomorrow. 

OBC_Karl tests all communications ports on the FM OBC. 

PROBLEM 28: The holes on the ACDS magnetorquer driver do not match up with the holes 

on the mounting brackets. 

REQUIRED MODIFICATION: New holes must be drilled into the mounting brackets. 

PROBLEM 29: The ‘reset’ wire that OBC_Karl attached to the FM OBC yesterday 

accidentally touched a positive power connection. This fried the ARM processor and the 

flash chips on board. 

108






MODIFICATION 20: OBC_Karl needs to make the computer again, this will take 

approximately a day. The components will arrive on Monday night with ACDS_Lars and 

CAM_Morten. 

LESSON LEARNED 7: Don’t do anything to flight model that you didn’t do to the EM. 

PROBLEM 30: EPS_Fulvio has lost all of his components and can’t start soldering his 

board. We have to order them again. 


ARRIVAL 38: ESA_Neil and ESA_Marcel go to Valkenburg airport to pick up the painted 

FM panels from the Royal Dutch Marines. They look great… 

109






ESA_Neil and ESA_Marie attempt to remove the kapton tape from all three FM –y panels 

and the FM baseplate. 

PROBLEM 31: The glue from the kapton is often left behind on the surface. 

ATTEMPTS: Tried to clean it with acetone, IPA, scrubbing, but little effect. 

MODIFICATION 21: Excess kapton glue is removed by “scrubbing” with a vinyl glove (to 

which it adheres better than to the panels), or, in extreme cases, with a scalpel. 

LESSON LEARNED 8: Do not use “flight” kapton tape if you are going to want to take it 

off again for any reason, as the glue is often too adhesive to so easily. 

ESA_Neil and ESA_Marie successfully remove all kapton tape from the FM –y panels and 

the FM baseplate. 

110






PROBLEM 32: The FM titanium ring is slightly to large for the exposed area of panel, and 

instead the edges rest on the layer of thermal paint at the outside. 

MODIFICATION 22: ESA_Neil and ESA_Marie carefully cut extra thermal paint away 

from the panel with a scalpel, until the area is large enough to encompass the titanium ring. 

The result is not too pretty, but the damage is only aesthetic. 

111






OBC_Karl teaches some soldering techniques to EPS_Fulvio. 

ESA_Neil and ESA_Marie attempt to start integration of the FM structure. 

PROBLEM 33: The M6 bolts between the separation system and the satellite do not run 

smoothly through the base brackets – some are extremely tight. 

SOLUTION: ESA_Neil re-taps the threads. 

ESA_Neil and ESA_Marie assemble the first stage of the Flight Structure, up to and including 

page 24 of Express_D_STRU+CONF_040926_02_Integration_Manual.pdf. 

NOTE: All bolts are done up just over “finger tight”, so that there is still enough play in the 

structure to integrate the rest of the panels unproblematically (as per lessons learned from EM 

integration). They will need to be “torqued” and glued for final integration. 

ESA_Neil removes the adapter plate from under the EM structure for mounting the FM 

structure in the transport box. 

112






ESA_Neil and OBC_Karl wrap the partially complete FM structure in plastic coating in an 

attempt to retain some level of cleanliness. 

ESA_Neil, ESA_Marie and ESA_Andre mount the FM structure into the transport container 

and close the lid. The shaker adapter and all the brackets are also taken for shipment to 

Stuttgart for the PROP integration. 

DEPARTURE 4: The Flight Structure in the PROBA box leaves for Stuttgart. 

MILESTONE 6: Stage 1 of the flight structure leaves ESTEC for Stuttgart 

113






ESA_Neil cleans and integrates the single painted engineering panel. The modified bracket 

#13 now fits fine, some glue could be added in the slot for the FM. 

ARRIVAL 39: Second lifting frame and a dummy antenna arrive from INFRA_Lars 


ARRIVAL 40: OBC FM replacement components. 

ARRIVAL 41: CAM spare components, along with CAM_Morten. 

ARRIVAL 42: Remaining ACDS components, along with ACDS_Lars 

OBC_Karl solders the new FM OBC. 

EPS_Fulvio and ESA_Marie hunt for component sources. Remaining components for the 

PDU are identified and ordered, will be delivered on Thursday. (BCR still to be ordered.) 

114






ESA_Jason discusses issues with CAM and ACDS teams. 

ESA_Neil cleans kapton and glue of off two more small shear panels. 

PROBLEM 34: ACDS power connector is inverted. 

SOLUTION: Fix it in the harness. 

REQUIRED MODIFICATION: ACDS power harness needs to be flipped. 

ESA_Neil attempts to integrate the dummy antenna to the EM structure. 

PROBLEM 35: The dummy antenna is too wide beneath the mounting plate and does not fit 

into the insert. Maybe the flight one is the same size! 

REQUIRED MODIFICATION: Should cut the bottom off of the dummy. INFRA_Lars 

says that the flight one is slightly smaller – we hope this is true. 

OBC_Karl completes the soldering of the FM OBC, but it does not work upon testing. 

MODIFICATION 23: OBC_Karl and CAM_Morten cut some hidden stray tracks on the 

FM OBC circuit board. Afterwards it seems to works fine, but more testing is needed. 

EPS_Fulvio practices his soldering. 

ACDS_Lars solders flight connectors onto the FM ACDS driver board. 

115







EPS_Fulvio practices soldering. 

ACDS_Lars attempts to define the harness and mounting for ACDS. 

PROBLEM 36: The mounting points on the –y lateral panel for ACDS coil driver do not 

match up to the size of the board. These will have to be re-positioned. 

REQUIRED MODIFICATION: The holes on the –y lateral panel for the lower of the 

ACDS coil driver brackets have to be re-positioned. Waiting on ACDS_Lars for the details. 

PROBLEM 37: The ACDS ‘internal’ connectors are not flight-worthy, but hard-soldering is 

not feasible since we don’t want to have the thread fragile items like sun-sensors through 

harness holes. 

116






REQUIRED MODIFCATION: A 25-pin d-sub connector will have to be added to the 

ACDS system. There is no mounting point for this, so an L-profile will have to be 

manufactured and then glued to the panel beside the board, there is not much room, but it will 

just fit above the coil driver. The pins for the harness connector will have to be inserted into 

the harness connector during integration of the side panels. 

PROBLEM 38: The ACDS magnetometer has a cadmium connector that is not suitable to 

fly in space as it will off-gas severely. 

REQUIRED MODIFICATION: The ACDS magnetometer connector must be replaced. 

ARRIVAL 43: ESA_Neil picks up the pyro connectors donated by ESA_Neil_Cable. 

OBC_Karl and CAM_Morten test the FM OBC with the FM CAMERA. 

117






PROBLEM 39: Cannot establish proper communication between OBC and CAM. 

ATTEMPT: OBC_Karl and CAM_Morten attempt to isolate the problem by testing 

communications with another CAN device (namely, the EM MAGIC). 

PROBLEM 40: Communication cannot be established with the EM MAGIC, because both 

ends have a CAN termination. We phone MAGIC_Renato to ask if OBC_Karl can de-solder 

the CAN termination on the EM MAGIC. MAGIC_Renato says yes. 

MODIFICATION 24: OBC_Karl de-solders the CAN termination on the EM MAGIC. 

(Two pairs of parallel 120-ohm resistors and a capacitor.) 

ATTEMPT: OBC_Karl and CAM_Morten attempt to isolate the problem by testing 

communications with another CAN device (namely, the EM MAGIC). The attempt fails. 

SOLUTION: After extensive testing and discussion with team-mates in Aalborg, OBC_Karl 

and CAM_Morten discover that the utility processor can only recognise CAN IDs above 127. 

This is why ACDS testing works fine, but nothing else does. New software will be sent from 

Aalborg tomorrow, and implemented on Friday. 

118






Using a “hack” OBC_Karl and CAM_Morten establish communication between the FM OBC 

and FM CAM. 

ESA_Neil disintegrates the corner profiles (all) from the FM lateral panels. 

PROBLEM 41: The rivet nuts don’t fit into the holes on the corner profiles, they are too 

large. 

REQUIRED MODIFICATION: The holes in the corner profiles must be made slightly 

larger to accommodate the rivet nuts. 

ARRIVAL 44: The FM PDU components arrive for EPS_Fulvio. 

ESA_Neil integrates the EM lateral panels (without corner profiles) to the EM structure as a 

simple fit check. 

119






PROBLEM 42: The M4x8mm bolts to fix the lateral panels to the primary structure are 

slightly too long for the inserts and therefore protrude a couple of mm past the lateral panels, 

allowing them to move away from the satellite body. 

120






REQUIRED MODIFCATION: Some kind of thick washers will be required to go between 

the lateral panels and bolt heads. 

PROBLEM 43: Many of the bolts are a very tight fit because of tolerance problems potting 

the side inserts. However, on the whole EM structure only two bolts do not fit at all, which is 

probably quite a good result. These two however are the centre two on one side of the topplate, 

which is a little worrying as it means that the top-plate is flexing down at the corners. 

SOLUTION: When integrating the lateral panels the centre bolts should be placed first and 

very very loosely, therefore allowing a lot of movement during integration. Once all bolts are 

in place they can all be tightened up. Hopefully the positioning will be more accurate in the 

flight model. Also, the top-brackets should be mounted as low as possible (within the bolt 

tolerances) in order to make sure that the top-plate can sit directly onto the top of the shear 

plates and line up to the holes in the top of each lateral panel. 

EPS_Fulvio begins to solder the FM EPS PDU. 

121






MILESTONE 7: The FM OBC downloads a picture from the FM CAM 


EPS_Fulvio continues to solder his flight PDU. 

ACDS_Lars works on the ACDS harness, completing wire soldering to the sun-sensors and 

the magnetorquer driver. 

OBC_Karl completed functional testing successfully with the ACDS FM Magnetometer. 

OBC_Karl and ACDS_Lars re-configured the pinouts from OBC to ACDS. 

OBC_Karl calibrated the analogue inputs on the FM OBC. 

122






PROBLEM 44: CAM_Morten and OBC_Karl killed the FM CAM computer board. Testing 

suggests that the FPGA failed. 

SOLUTION: CAM_Morten will take the whole thing back to Aalborg and attempt to resolder 

a new computer for the camera. 

MODIFICATION 25: ESA_Neil enlarged the holes on the corner profiles to accommodate 

the rivet nuts. 

ESA_Neil integrates the rivet nuts to the corner profiles. 

ESA_Neil integrates the corner profiles to the EM structure. 

MILESTONE 8: EM structure completely assembled. 

PROBLEM 45: The bolts are not really long enough to extend far enough into the rivet nuts 

after they have gone through the corner profiles. 

123






REQUIRED MODIFICATION: New, longer bolts are required. 

ESA_Neil and ESA_Marie clean all remaining katon tape and glue off of flight primary 

structure panels. 

ESA_Neil and ESA_Jason prepare all the torque wrenches and attachments that will be 

needed during the Stuttgart PROP / STRU integration. 

PROBLEM 46: There are not enough M6 nuts for the rest of the primary structure 

integration – we are short by one. 

MODIFICATION 26: ESA_Neil removes an M6 nut from under the EM lifting frame, 

taking care not to let the top bold fall into the satellite. This bolt to solve problem 46. 

REQUIRED MODIFICATION: Need to add an M6 nut to the top of the EM structure. 

124







EPS_Fulvio continues soldering the FM PDU. 

CAM_Morten packs up all camera parts to take them back to Aalborg. 

MODIFICATION 27: ACDS_Lars and OBC_Karl re-configure the pinouts between OBC 

and ACDS 

DEPARTURE 5: CAM_Morten takes all the CAM hardware (apart from the box) back to 

Aalborg, where he will attempt to build a new computer. 

ARRIVAL 45: Remaining OBC components arrive. 

OBC_Karl prepares and mounts thermistors and inserts “flight” software PROMS. (Arranges 

on a blank board first) 

ESA_Neil completes removal of the kapton tape from the flight shear panels. 

125






EPS_Fulvio glues heavy components down to the FM PDU. 


OBC_Karl tests all serial ports on the FM OBC using simple RX-TX loop-back – successful! 

OBC_Karl glues thermistors harness into place on FM OBC (needs thermal glue for unit 

itself). 

126






EPS_Fulvio continues to prepare the FM PCU. 

ESA_Neil prepares and packs all materials and tools required for the rest of the PROP and 

STRU (primary) integration in Stuttgart. 

PROP team and SYS_Joerg erect a “cleantent”. 


MODIFICATION 28: OBC_Karl decides that he does not need all three of the analogue 

inputs on the FM OBC for thermistors as originally planned. After discussion with ESA_Neil 

the remaining two inputs are then to be used as extra TCS thermistors. (Note: they will still 

register as OBC telemetry.) 

ESA_Neil collates the hardware to take to Stuttgart for the PROP / STRU integration. 

127






ESA_Neil drives from ESTEC to Stuttgart with the remaining flight panels, the lifting frame, 

the bolts and a set of torque wrenches. 

EPS_Fulvio finishes soldering the PDU and connects it to the other PCDU boards. 

OBC_Karl and EPS_Fulvio commence functional testing with FM EPS and FM OBC. 

- Boot-up sequence demonstrated 

- Serial communication established 

- Watchdog system stable 

- Telemetry received 

- Telecommands executed properly (turn system X ON or OFF) 

- EPS shutdown of OBC successful 

MILESTONE 9: Successful functional integration of FM EPS and FM OBC. 

Meanwhile, in Stuttgart, ESA_Neil and PROP_Nils do a simple, and successful, fit check 

with the pyro connectors supplied by ESA_Neil_Cable. 

128






PROP_Nils and PROP_Sascha integrate thrusters clusters to the low-pressure tubing, 

including simple leak tests on all new connections (10 bar). 

PROBLEM 47: The mounting bolts for the tensioning of the tank are not 25mm as specified 

by PROP, but only 13mm as specified by STRU_Antonio (so as not to interfere with the +z 

T-Pod e-box). 

MODIFICATION 29: PROP_Hanno, PROP_Nils and PROP_Matthias develop method of 

tensioning of tank clamps when using shorter bolts and tank under pressure. (Tank taken to 

~210bar using direct connection with manual valve, and compressed air and time to cool the 

tank.) 

129






PROP_Nils, SYS_Joerg and PROP_Hanno prepare and clean all relevant equipment to enter 

the cleantent. 

PROP_Sascha and PROP_Matthias enter cleantent to mount the ACS thrusters and the lowpressure 

tubing, including gluing the base of the mounting cones, and applying glue around 

the edges of all ‘insecure’ connections (all those without safety wires). 

PROBLEM 48: Thruster insert has an M4 thread instead of an M5 thread. 

MODIFICATION 30: An M4 bolt is used to mount the thrusters clusters to the thrusters 

inserts instead of an M5. The other bolts nearby are M5, so the system as a whole should be 

strong enough. 

130






PROBLEM 49: The conical low-pressure tubing mounts do not sit right down onto the 

baseplate. This is because the tubing was bent by hand and is not accurate to the millimetre. 

MODIFICATION 31: The addition of washers to space the mounts, and then a large weight 

(7kg metal plate) placed on top of the tubing to hold them down while they dry overnight. 

PROBLEM 50: One of the low-pressure tubing mounts is not entirely touching the baseplate 

– this one will be slightly weaker than the others, but should be within acceptable limits. 

In order to be able to tell when the glue is sufficiently dry, two bolt heads are also glued to a 

test plate. 

131






1 st November 2004 

ARRIVAL 46: UHF EM arrives in ESTEC. Visual inspection, somewhat worrying. 

PROP_Hanno cleans the tank and the high-pressure tubing. 

PROP_Hanno and PROP_Matthias enter the cleantent to mount the tank and the high pressure 

tubing. 

The test from yesterday is not quite dry, so before the weight is removed the clamp mountings 

are disconnected from the low-pressure tubing to relieve any strain on them. 

PROP_Hanno and PROP_Matthias integrate the tank and the high pressure tubing. 

132






PROBLEM 51: The PMS and of the high pressure tubing is slightly misplaced. 

ATTEMPT: Correct by slight rotation of the high pressure unit (tank and tubing). Better, 

but still not good enough. 

ATTMPET: Mount the PMS box to the tubing (high and low) to assess the alignment. 

Results, initial misalignment, but tubing is flexible enough to manoeuvre into position 

temporarily (a little bit of strain). 

SOLUTION: Only the inner-most mounting point is glued initially. The rest will wait until 

the primary structure and the PMS box are integrated. In this way we can be sure that they 

are in the correct position, as the panel will hold it there. 

133






PROP_Hanno secures the wires temporarily with plastic cable ties. 

PROP_Hanno and PROP_Matthias lower the tank slightly to provide some pressure on the 

mounting that is being glued. 

Meanwhile, in ESTEC: OBC_Karl removes EM PCBs from UHF box, so that ESA_Marie 

can send the FM box and FM DC-DC converter back to UHF_Holger for integration of the 

FM. 

PROP perform a shake test on the PMS box using random vibrations of the qualification loads 

on the z axis and the x axis, for four minutes each. 

134






Visual inspection suggests that the PMS box survived intact, no problems were found. 

However, a full leak-test is the only way to tell, which we will only perform tomorrow when 

the GSE equipment (fill drain valve) returns to the workshop. 

PROP_Hanno slightly raises the tank and then tightens the mounting straps into flight 

configuration. 

PROP_Hanno connects the low-pressure tubing clamps to the mounting points. 

PROP_Hanno glues the tank thermistor harness to the tank, temporarily securing the wire 

with kapton tape. 

ESA_Neil and SYS_Joerg loosely integrate all the flight panels to the flight structure around 

the propulsion subsystem. The process was markedly easier than the integration of the EM 

primary structure and no significant problems were encountered. (Towards the end the 

kapton tape securing the thermistor harness was carefully removed, leaving the wire in place.) 

135






There is just time for one more final check of the tank and the tubing before it disappears 

from view forever. 

136






137






There was a minor problem with the masking areas on a couple of the brackets, but 

TCS_Joerg was happy with the expected level of conduction, so no modifications were 

required. 

PROP_Hanno and PROP_Sascha loosely mount the FM PMS box to the –y panel and connect 

it firmly to the high- and low-pressure tubing. 

138






ESA_Neil and SYS_Joerg tighten all the structure bolts, making sure that all the panels are 

aligned properly, all bolts turn smoothly, and the top-plate sits as low as possible – as per 

experience gained during integration of EM secondary structure. Each bolt is torqued, and 

then glued (apart from the top-plate). 

The top-plate is removed and ESA_Neil and SYS_Joerg glue in the final brackets, securing 

them in place with aluminium tape until tomorrow. 

139






MODIFICATION 32: Due to the lack of one aluminium “patch”, a spare of the smallest 

mid-height bracket is used in the –x-y+z corner of the central compartment. This will add a 

little mass, but provide much better strength. 

MILESTONE 10: This is the point of no return. 

NOTE: The structure cannot be disintegrated once these brackets are glued in. From 

this moment forward we are on a “one shot” strategy. 

PROBLEM 52: PROP_Hanno is not happy with the stress on the high-pressure tubing. 

MODIFICATION 33: ESA_Neil and PROP_Hanno temporarily remove the PMS box and 

bend the tubing into place. The PMS box is then replaced. During this operation the topplate 

is temporarily replaced, in order to help make sure that the other panels do not move. 

PROP_Hanno glues the high-pressure tubing mounts to the baseplate, adding washers to the 

mounting points where necessary to space them correctly onto the baseplate. 

140






Since it is 3am, everyone goes to bed to sleep while the various bits of glue harden. 

2 nd November 2004 

PROP_Hanno and PROP_Nils glue a thermistor to the tank in the cleantent. The tank 

thermistor requires aluminium tape to hold it against the tank in the correct position. This 

tape is equipped with a nut and a piece of string hanging down to the bottom of the satellite, 

so that even after the top plate is on the string, nut and tape can be removed. 

PROP_Hanno and PROP_Nils disintegrate the PMS box and remove it from the cleantent. 

PROP_Hanno glues thermistors onto the thrusters clusters. 

141






PROP_Nils glues thermistors onto the PMS box. 

SYS_Joerg and ESA_Neil shop for supplies, finding some potential rubber parts to use as 

shock absorbers. 

ESA_Neil prepares the shock absorbers. 

142






They are perhaps too rigid, but there are no alternatives at this point. 

PROP_Hanno, PROP_Sascha and PROP_Nils solder test connectors to the PMS box. 

ESA_Neil and SYS_Joerg remove the aluminium tape from the glued brackets, some of the 

glue remains on the brackets and panels, but is cleaned off with an IPA (ish) tissue. 

143






ESA_Neil and SYS_Joerg integrate the top-plate, and torque and glue the nuts, bolts and 

washers. The upper ASAP bolts are also torqued, but not glued since these need to be 

exchanged at the launch site. 

MILESTONE 11: The FM Primary Structure is completed. 

SYS_Joerg makes a protective plastic bag to cover the satellite. 

No other work can progress until we have the ground half of the fill-drain valve. This should 

arrive tomorrow morning, so we can get some proper sleep tonight. 

144






3 rd November 2004 

OBC_Karl attempts to establish communication between the FM OBC and the EM UHF 

modem. Attempts are unsuccessful. 

EPS_Fulvio continues to work on the FM PDU. 

The PROP ground-half of the fill-and-drain valve arrives, but is still broken. PROP arrange 

to have a replacement sent, but cannot work without it. 

ESA_Neil drives back to ESTEC. 

4 th November 2004 

PROP receive a replacement ground-half of the fill-and-drain valve and perform leakage tests 

on the FM PMS. 

MILESTONE 12: The flight model Pressure Management System passes its vibration 

and leakage test. 

OBC_Karl, MAGIC_Renato and MAGIC_Luigi establish two-way communication between 

the FM OBC and the FM MAGIC. Telecommands and telemetry are exchanged with no 

significant problems. 

PROBLEM 53: The magic box uses flash memory, which may not fare too well in space. 

OBC_Karl and MAGIC_Renato investigate ways to replace it. 

145







MAGIC_Luigi and EPS_Fulvio glue capacitors to their boards. 

OBC_Karl fixed the flash bug of the day. 

EPS_Fulvio soldered wires onto his board. He has many wires. 

PROBLEM 54: One-time-programmable chip not available as replacement for Texas 

Instruments DSI in the Magic box. Solution: MAGIC_Renato to investigate boot-loader 

options. 

PROBLEM 55: The modem in the EM UHF box also appears to have flash on-board. 

SOLUTION: Advice given to S-Band team was that this is not so dangerous as long as it is 

not updated in space. Also, we will put tantalum over the chips to protect them. 

SYS_Joerg and the PROP team integrate the PMS box into the structure, perform a pressure 

test on the connections, torque the bolts, glue the bolts, cover the satellite in the bag, and 

install it in the transport container. 

146






147






The satellite is now ready to go to IABG for the shaker test next week. 


We learn that the UHF and S-Band units also use flash chips. Advice from AMSAT and 

SSTL is that flash is ok in low-earth orbit if it is not updated. Advice from ESA and Aalborg 

is that flash is not ok in space, at all. 

MODIFICATION 34: We decide to leave the Magic box hardware as it is but make the 

software more robust by having a minimalist (and therefore small target) random number 

generator, which then directs to any of a large number (however many fit on the chip) of 

148






copies of the main code. Any particular copy of the main code will then report it’s own 

checksum to the OBC, which will toggle the unit if it doesn’t get the right response. 

MODIFICATION 35: We decide to cover all flash chips on-board with small protective 

layers of tantalum to err on the side of caution. 

LESSON LEARNED 9: Take the time at the beginning of the design phase define which 

components and standards are acceptable for the project. 

EPS_Fulvio puts together the FM PCU box. 

MAGIC_Renato readies the MAGIC boards and connectors. 

OBC_Karl installs and configures the linux box to use on the OBC debugging port. 

149







ESA_Neil removes the acrylic jig from Xi-III (Xi-V EM) in order to send it and its box back 

to XI-V_Yuya. 

ESA_Neil removes the –y lateral panel from the EM structure and removes the FM PIN. 

OBC_Karl programs “ground station” software. 

MAGIC_Renato cleans, and glues wires into, the FM MAGIC. 

ESP_Fulvio wires up connectors to the FM PCU. 

PROBLEM 56: EPS_Fulvio and ESA_Neil discover that when proper redundancy is applied 

to the connector leaving the PCU for the RBF connector, there are not enough pins 

MODIFICATION 36: Decide to split harness from EPS RBF connector such that the 

battery lines come direct from the battery box and back, while the others come from the PCU. 

ESA_Neil prepares harness for inter-subsystem testing between EPS, OBC, PIN, MAGIC, 

UHF and S-BAND. 

150






OBC_Karl prepares the FM OBC for inter-subsystem testing and prepares his laptop as a 

dummy uplink mission control computer. 

EPS_Fulvio finishes wiring to his connectors and closes the FM PDU. 

MAGIC_Renato tidies and cleans the FM MAGIC boards and cables and wires up the 

connectors. 

ARRIVAL 47: OBC_Karl gets the linux box (OBC debugger) up and running. 

ESA_Neil makes test harness for interconnections between all present subsystems. 

ESA_Neil, OBC_Karl, MAGIC_Renato and EPS_Fulvio perform functional testing with the 

FM MAGIC, FM PCU, FM OBC, FM PIN and a dummy pyro-valve (simple resistance). 

151






PROBLEM 57: The OBC flight proms have not been added, so it is required to load 

software into to memory upon every boot-up. This upload takes about 50 seconds, which is 

longer than the EPS-OBC watchdog period. Therefore the OBC currently cannot be powered 

up directly from the PCU. 

SOLUTION: The OBC is powered up directly from a normal power supply for the software 

to be uploaded in advance, and the boot-up synchronised by hand with the power cycling 

from EPS (which is measured using a dummy load and a voltmeter). 

PROBLEM 58: When progressing through the SECOND (first is ok) cycle of recovery 

mode to nominal mode, EPS sends a shutdown command to OBC even after OBC responds to 

the watchdog ping. Current workaround: cycle power to the PCU. 

All operational modes ok apart from above. 

Nominal mode pings and telemetry between EPS and OBC are good. 

MAGIC_Renato replaces MAGIC connectors with crimp instead of solder. 

152






CAN link between OBC and MAGIC is found to be fine and stable. Telemetry is received 

and telecommands are processed (can hear the pyro relay click). 

A dummy load is attached to the pyro valve ports and the firing command sequence is given 

by to the OBC. A scope across the load measures the result, which is a very nice discharge 

curve. 

MILESTONE 13: Functional integration of OBC, EPS and MAGIC at the stage where 

the PROP payload is supported by the platform. 

The safe mode beacon is picked up on a scope connected to the Push-To-Talk line. 

PROBLEM 59: The power levels seem dodgy. For example, when MAGIC is off it still gets 

1.4 volts through the PIN. 

EPS_Fulvio reprograms the PIC multiple times to correct the software. (This includes cycling 

the chip in the socket.) 

153







ARRIVAL 48: The FM S-BAND unit and a large amount of communications equipment 

arrives with the AMSAT UK team. 

ARRIVAL 49: The FM UHF box and antenna arrives with UHF_Holger. 

ARRIVAL 50: The Low Gain Antennas, FS S-BAND enclosure, Microwave Cables and 

Antenna Caps arrive with COMMPL. 

ESA_Neil, OBC_Karl, EPS_Fulvio and AMSAT-UK functional testing EPS, OBC, PIN, 

SBAND, UHF 

PROBLEM 60: The S-band unit fails to power up using the EPS. This is found to be due to 

a large inrush current on the DC-DC converter that EPS is treating like a latch-up. 

REQUIRED MODIFICATION: Some kind of choke must be fitted to the S-band unit 

power input so spread the surge out and keep it below the critical current stipulated by EPS 

(about 1 amp). Temporarily the S-Band unit is powered directly from a normal power supply. 

PROBLEM 61: The S-band units needs ground on the RS232 line. 

MODIFICATION 37: A line is added connecting pin 5 of the RS232 line to the ground 

arriving from the PCU. 

PROBLEM 62: The S-band unit requires a ground on the audio line with UHF. 

MODIFICATION 38: A ground is taken from after the DC-DC converter (to keep galvanic 

isolation) to the audio link. 

154






PROBLEM 63: The OBC experiences a line driver failure on the EPS and UHF ports. This 

is probably due to an unconnected cable from the port acting like an antenna and running lots 

of powerful RF into the computer. 

MODIFICATION 39: OBC_Karl replaces the line driver. 

LESSON LEARNED 10: Anything can act as an antenna. Keep all cables connected at 

both ends and keep all at low level and at a reasonable distance from sensitive electronics. 

SOLUTION: EPS_Fulvio fixes the power levels to the load (the PDU was not being fed 

power correctly.) 

ESA_Neil and ESA_Eric glue the test solar panel to a dummy aluminium plate (with thermal 

paint on the other side). 

The UHF box powers-up ok. 

155






MILESTONE 14: The SSETI Express safe mode and nominal mode beacons are 

received via a radio for the first time. 

This verifies the push-to-talk functionality of the PCU, and the UHF radio. 

PROBLEM 64: AMS_Sam and AMS_David discover that the third harmonic from the S- 

BAND unit is only 30dB below the carrier. 

MODIFICATION 40: A low pass filter needs to be added to remove this problem. 

The nominal mode beacon is received on a handheld radio (audio). This verifies the nominal 

mode on OBC and the RS232 ports on OBC and UHF. 

The audio link from UHF to S-BAND is ok, this verifies the UHF output. 

The S-BAND box modes are successfully controlled by the AMSAT-UK team. 

156






MILESTONE 15: Spacecraft hardware is commanded via an RF link for the first time. 

When in the correct modes the telemetry sent from S-BAND is received and decoded in the 

“groundstation” next door. 

MILESTONE 16: SSETI Express transponds audio for the first time. 

This verifies the UHF uplink, the audio connection to S-BAND, and the S-BAND downlink. 

Various bizarre and complex tests are performed by the AMSAT-UK team. 

PROBLEM 65: Packets downlinked from UHF or S-BAND are valid frames but gibberish. 

This is because the protocol is not implemented correctly on the OBC. 

Meanwhile, in Munich, the PROP team have been preparing the spacecraft for the pressurised 

shake. Pressurisation takes a lot longer than expected because of the lack of convection inside 

the centre compartment. 


ESA_Neil bolts three low-gain antennas, and the UHF antenna, onto the engineering 

structure. 

PROBLEM 66: The holes in the LGAs do not quite match the mounting points on the 

mounting bracket on the top plate. 

157






REQUIRED MOFICATION: The +z LGA should have the holes slightly widened. 

PROBLEM 67: The structure of LGA looks far to sensitive to mechanical loads and 

vibrations, we expect that they will fail during the vibration testing. 

One of the LGAs is hooked up to the S-BAND unit and much testing of the radiation pattern 

and signal ensues. The AMSAT-UK team are happy with the results. 

158






Test solar panel is completed and delivered to test centre. 

PROBLEM 68: A significant amount (about 10%) of the RF from the antennas is 

transmitted backwards into the satellite. This will cause problems for the on-board 

subsystems, and the compartments will act as waveguides. 

REQUIRED MODIFICAION: Addition of back-shields for the antennas to provide 

mechanical strength and to reflect ‘back’ RF out into space. 

AMS_Graham, AMS_David and AMS_Sam check the polarity of the circular polarisation on 

the antennas and find it to be right hand (IEEE). 

159






AMSAT-UK team perform various complex tests on the antennas… 

MODIFICAION 41: OBC_Karl and AMS_Jason fix the AX25 protocol implementation on 

the OBC. 

MILESTONE 17: The nominal mode beacon is received and decoded for the first time 

by the test ground station. 

UHF_Holger gets the EM UHF up and running again. 

PROBLEM 69: The nominal mode beacon has wrong call sign (AMS_Jason’s), and 

contains incorrect information about the website. 

REQUIRED MODIFICATION: The nominal mode beacon call-sign must be changed 

before the flight (once we know what it is), and the web address should be given as 

www.sseti.net. 

160






Data received at ground from UHF TX packets, valid but undecodable due to non-binary 

nature of groundstation. 

Data received at ground from S-BAND TX packets valid but undecodable due to non-binary 

nature of groundstation. 

Data received at space segment, UHF RX packets arriving but gibberish (can't send in binary 

from groundstation). 

MILESTONE 18: SSETI Express receives, acknowledges and responds to its first RF 

telecommand. 

OBC_Karl finally goes home. 

AMS_Howard prepares simplest and minimalist groundstation imaginable. 

We all go for dinner. 

MODIFICATION 42: An inductance, diode, load resistor are tested as a choke on the S- 

BAND power input and work fine to reduce the current inrush. S-BAND can now be 

powered by EPS directly, although the flight implementation of the modification remains 

pending. 

The PROP team finally fill their tank to 290 bar, watch it for two hours and the pressure 

seems stable. 

161







PROBLEM 70: The PROP team get back to IABG and find that the tank is now at only 33 

bar. It is leaking from below the tank, where an o-ring has popped its housing. 

LESSON LEARNED 11: Before reaching any “point of no return” make ABSOLUTELY 

sure that everything “before” the point is complete, and demonstrate this by testing IN 

CONTEXT. For example: even though the tank had passed a pressure test previously, the 

same tests should have been done again after integration to the structure, before the point of 

no return. 

REQUIRED MODIFICATIONS: ESA_Neil and SYS_Joerg define the following list of 

modifications during an emergency teleconf: 

1) Goal is to take the top plate off, therefore the following steps are performed: 

1.1) remove the washers and nuts of the top plate. As the washer and the nut can 

stay together, only the glue between top-plate and washer has to be removed. 

For that, the glue can be heated carefully to a maximum of 75°C. 

WARNING! A temperature above 80°C *WILL* damage the top-plate honeycomb 

and destroy it! 

The heat is required to remove the shear strength of the glue, so you can remove 

the washer off the plate by turning the nut (and the washer). 

1.2) After removing all nuts and washer, and taking out the bolts carefully, you 

can remove the top-plate and store it clean and safe. 

2) Goal is to cut the Thermistor wire without removing the thermistor from the tank: 

The thermistor wire is glued to the tank, to the straps, winded around the tubing 

and glued to the tubing. So, the wire has to be cut! 

162






Find a way to cut it, so the thermistor can be re-connected later. 

Suggestion: 

2.1) Cut at bottom and at top and find a way to cut it between the straps. 

2.2) don't remove the thermistor! 

3) Goal is to remove the tank: 

3.1) Release tension on the tank mounting straps carefully. You know best how to do 

that, remember that nothing should fall in the S/C and that you have to fix it 

later again without being able to access from the sides! 

3.2) If possible, hold the HP tubing somehow for not putting stress on it 

3.3) Turn the tank and remove it. Good luck. 

4) Put in the washer to fix the sealing. 

5) Put in the tank again - same procedure reversed. 

5.1) Fix it 

5.2) Torque it 

5.3) Torque the straps 

5.4) find a way how to re-connect the thermistor 

6) Fill the tank 

7) Put the top plate on 

7.1) use new washers and nuts, please 

7.2) Torque them with 4 Nm (four Nm) 

8) ready for shake? 

You will have terrible logistic problems, how to access the S/C from top and bottom. 

Please remember that you have to lift it by the lifting frame and probably have to 

fix it to the adapter plate when putting torques on it. 

Please, think about that before for not ending in a 'dead-lock' situation. 

Good luck! :-) 

In ESTEC the AMSAT-UK team take apart the S-BAND box to identify and address any 

issues internally. 

163






ESA_Jason casts his wary eye over the interior of the S-BAND box, identifying several solder 

joints and strain relief of wires that need to be fixed. Upon testing it turns out the “PTFE” 

wire was not so at all, and needs to be replaced. 

ARRIVAL 51: AMS_Howard completes setup of a simple test groundstation that can be 

used for functional testing. 

DEPARTURE 6: AMS_Sam, AMS_David and AMS_Jason leave ESTEC taking with them 

the FM S-BAND unit, the FS S-BAND enclosure, ready-crimped lengths of 24 gauge PTFE 

wire, flight solder and rosin flux. 

AMS_Howard gives ESA_Neil crash course in setting up and operating the “loaned” test 

groundstation. 

ESA_Neil sets up and operates the test groundstation. 

164






MODIFICATION 43: The PROP team progress through to step 5.1 inclusive from the list 

given as a required modification above. 

PROBLEM 71: The PROP team cannot torque the tank properly (step 5.2 from required 

modification listed above) because of bad access through the structure. 

EPS_Fulvio implements a boot-loader on the PCU and installs the PIC. This removes the 

need to keep cycling the socket every time software is uploaded. 

AMS_Howard ESA_Neil performs functional testing the FM OBC, FM UHF and test 

groundstation before AMS_Howard leaves for the UK 

Uplink of packets works, usually with a “TC too long”, “TC too short” or “Rubbish in buffer” 

ending up in the alarm stack (as expected since the groundstation can currently only send 

ASCII data). Sometimes the right length of packet is achieved for it to be accepted by the 

OBC as a telecommand (although a bogus one), acknowledged via an undecodable UHF 

downlink packet, and stored in the flight plan. 

PROBLEM 72: The exact length of telecommand uplink that gets recognised by OBC is 

indeterminate and non-repeatable. OBC_Karl is made aware of this by ESA_Neil. 

PROBLEM 73: Randomly, twice, a particular ASCII string is sent which causes the OBC to 

kill the nominal mode beacon (otherwise the OBC seems fine). This should not be possible. 

OBC_Karl is made aware of this problem by ESA_Neil. 

PROBLEM 74: On the rare occasions where an attempt at uploading a telecommand is 

actually successful in placing the command into the flight plan, it overwrites any command 

that is already there – then resulting in a maximum of one TC in the flight plan at any given 

time – which is clearly not enough for the mission. OBC_Karl is made aware of this problem 

by ESA_Neil. 

165






ESA_Neil and EPS_Fulvio perform functional testing between the FM EPS and the FM OBC. 

The problems with the ping responses are still present but the timing is a lot better. 


MODIFICATION 44: PROP team temporarily glue some metal blocks to the top of the tank 

to assist with applying the correct torque. 

The PROP team place the satellite in a protected room and head back to Stuttgart to prepare 

and procure the necessary tools to torque the tank on Monday. 

EPS_Fulvio solders the FM BCR. 

ESA_Neil reviews EPS_Fulvio’s software for the PIC on the EPS PCU. 

ESA_Ceaser begins thermal vacuum testing on the test solar panel. 

ESA_Neil and EPS_Fulvio perform functional testing on the FM OBC and FM PCU. 

PROBLEM 75: EPS_Fulvio accidentally powers a 5V board from a 28V supply. This 

appears to damage either the MAX232 line driver or the PIC. After further testing it is 

determined that the PIC is partially damaged (still partly functional). 

LESSON LEARNED 12: DOUBLE check ALL the connections to the power supply 

EVERY TIME before you turn it on. 

LESSON LEARNED 13: At the end of any particular session remove all the test cables to 

the external power supplies, forcing them to be set up again the next time instead of relying 

on it being set up correctly already (maybe you were doing something slightly different in the 

last session and you forgot the details). 

MODIFICAITON 45: EPS_Fulvio replaces the PIC with the only remaining spare. 

ESA_Neil and EPS_Fulvio perform functional testing on the FM OBC and FM PCU. The 

timings are now at the stage where the OBC can be powered from the PIN directly, with a 

quick software upload on the OBC in time to come up before the first ping arrives. Problem 

number 58 is still unresolved. 


EPS_Fulvio hunts for his software bugs. 

166






EPS_Fulvio solders on the BCR. 

ESA_Neil and ESA_Marie set up the groundstation and test some TC uplink, unsuccessfully 

due to ASCII / HEX software limitations. 

EPS_Fulvio reprograms and retests his PCU software. 


EPS_Fulvio and ESA_Neil perform extensive functional testing on the PCU and OBC. 

Eventually it is concluded that the problems with the PCU must be due to an intermittent 

failure on the PIC, which could have been damaged during one of the two occasions when the 

incorrect voltage has fed the PDU. No other explanation seems to make sense: the code is 

carefully checked and re-written many times, the link is tested both sides by manual 

emulation, the timings are tested, the beacon is turned off to prevent beat frequencies on 

timings (one / two minute cycles on ping and beacon respectively in recovery mode), and yet 

still the problems are intermittent, unrepeatable and random. 

ESA_Neil orders new PIC chips. 

EPS_Fulvio instructs ESA_Neil on the programming of PICs and their insertion into the 

PDU. 

ESA_Neil performs functional testing on the uplink to UHF and OBC via the test 

groundstation. Overall the following issues are identified in the software of the OBC 

(extracted from email to OBC_Karl): 

167






PROBLEM 76: Occasionally get looping "de-kissisifcation error" down the debugger upon 

startup, and once we do it doesn't get better. This seems to be resolved once the linux box is 

rebooted. 

PROBLEM 77: When using TCINS to send commands to EPS while looking down the EPS 

port with a laptop the parameters get swapped around. Is this another least-sig / most-sig 

issue, or is this intentional? (It is probably general to all tcins, but this was the case I 

discovered.) 

PROBLEM 78: Upon startup of OBC the bytes "0f 6f 6f" get sent to EPS. Why is this? (It 

even appears in the flight planner briefly, I caught it once.) This is why EPS was shutting you 

down sometimes... now it just counts errors... 

PROBLEM 79: I have just about found a way to uplink HEX-based TCs to the OBC via the 

UHF, but it is still a little shaky (is only half tested, the RF link is not strong enough to be 

THAT reliable, and it has to end in a carriage return). I can get stuff into the flight planner 

quite well, and the right numbers seem to be appearing in the right places. Issue is: if I send a 

TC starting with "00" (i.e.: subsystem ID is OBC) then it does not acknowledge, does not go 

into the light plan, the nominal mode beacon stops, and it ignores anything incoming from 

UHF completely from then onwards. (Although it all looks fine on the debugger.) 

PROBLEM 80: TCs from UHF are overwritten, not stacked. This would make for a tedious 

mission... 

REQUIRED MODIFICATION: It would be really good to have either everything ("time", 

time on "tmdata", time on "plist" etc) in EITHER linux time, OR real time. I don't actually 

mind too much which, but it is confusing that it changes a lot, as comparisons are then tedious 

and time consuming. Obviously real time is better for the general layman, but consistency is 

the most important thing. 

PROBLEM 81: After adding commands on TCINS to be executed immediately (which they 

do), I have found them still listed in the flight planner later on, which is a tad odd. Doesn't a 

TC get taken out the planner when it's executed? This has not been a repeatable error... 

unfortunately. 

ESA_Neil and EPS_Fulvio pack up everything and tidy the cleanroom. 

DEPARTURE 7: EPS_Fulvio finally gets to go home, taking with him the FM BCR and the 

EM PDU. He will finish them in Naples and send them back. 


ESA_Neil flies to Munich to join PROP_Sascha and PROP_Hanno. 

168






The new tank adapter is manufactured and fitted to the tank already with a maximum torque 

applied. No additional tubing is required. 

PROP_Hanno inserts tank and connects to the high-pressure tubing. The system is 

pressurised using the Cryosat pressure intensifier and checked for leakage. There is none. 

The satellite is lifted and the connection to the high pressure tubing is torqued and locked off 

with safety wire. The satellite is lowered and the thermistor is glued to the tank, with the 

cable running through the extra insert above the PCU box. 

169






PROP_Sascha and ESA_Neil replace the shock absorbers in the PROBA box with “softer” 

ones. 

PROP_Hanno and PROP_Sascha top up the pressure as necessary as we wait for thermal 

equilibrium to be established. 

PROP_Sascha, PROP_Hanno and ESA_Neil move the satellite up into the shaker test room. 

PROBLEM 82: We notice some structural damage to the exposed core on the +y side of the 

baseplate. We have no idea when this occurred, but it looks like an impact by something like a 

screwdriver handle. 

170






The top-plate is replaced, torqued and glued. The satellite is left pressurised to 301.5 bar 

overnight for the decay test. 


The decay test was successful, and the pressure in the morning read at 299 bar. (Expect slight 

drop as cooler overnight.) 

PROP_Sascha, PROP_Hanno and ESA_Neil prepare the satellite for the shake, including 

securing lose cables, bolting the test adapter to the table and the satellite to the test adapter. 

A 0.5g resonance search on the x-axis is performed. No problems experienced. 

171






The random vibrations on the x-axis are performed, then the satellite is inspected. The 

pressure is still at 302 bar. 

PROBLEM 83: Upon visual inspection it is found the high pressure tubing mount closest to 

the tank has been torn off of the baseplate. 

PROBLEM 84: Upon visual inspection it is found that the inserts holding the tank 

mountings have been pulled into the centre compartment by a couple of millimetres, therefore 

damaging the honeycomb panels slightly by pulling the core away from the skin. This is 

because the washers used on these bolts were too small, and within the circumference of the 

inserts, therefore not transmitting loads onto the honeycomb in the proper way. This is highly 

concerning structural damage. 

PROBLEM 85: ESA_Neil notices that the PMS box mountings also have washers that are 

too small for the inserts. 

MODIFICATION 46: The high-pressure tubing mount is glued back down with “UHU” 

two-part resin (fast drying). 

MODIFICATION 47: It is decided that washers “as big as possible” (quote STRU_Melro) 

should be added to the bolts on the tank mounting. Therefore two small stainless steel plates 

are made, each of which act as a washer for one pair of the tank mounting bolts. While the 

tank mountings are tightened again the honeycomb can be heard being pulled back into place. 

This seems like a very strong solution and is judged adequate to continue the testing. 

MODIFICATION 48: Larger M5 washers are added beneath the M4 ones on the PMS 

mounting points so as to distribute the loads properly. These bolts are torqued and glued 

again (UHU). 

172






NOTE: We were extremely lucky that we randomly chose to shake on the x-axis first, and 

that we had a 15 degree offset. This allowed us to detect and fix the problems above that 

would have destroyed the structure on the y-axis shake. 

LESSON LEARNED 14: Before a major system level test EVERY SINGLE 

COMPONENT should be checked and rechecked on ALL criteria. 

The x-axis shocks are applied. During the high frequency shock simulation there is a very 

strong shock as the shaker over-runs its displacement limit and hits a mechanical stop. It is 

decided to finish these shock tests where they stand. 

The pre- and post-vibration resonance searches do not agree very well. It doesn’t look as it 

something is broken, but it looks as if something is stiffer on the PMS. This can be explained 

by the larger washers employed in the modification, which transmit the loads into the 

honeycomb directly, instead of through the more flexible glue. 

The mid- and low-pressure sections are taken to 17 and 2.4 bar respectively. The mid 

pressure appears to leak intermittently. They are left at this pressure for the next tests. 

The satellite is turned by 90 degrees for the y-axis tests. 

All the vibration and shock tests are written up and documented in full in the PROP test 

reports. 

The y-axis random shakes are applied, and then the shocks. 

PROBLEM 86: The mid-pressure system has depressurised. This is bad, but not critical, 

since the SSTL requirements are on the high-pressure system. The leak is intermittent, and 

only seems to happen when the filling hose is disconnected, but no leak can be found in that 

area. We decide to proceed with the shake anyway. 

173






PROBLEM 87: The clamp between the high- and low-pressure tubing has come loose. It is 

simple to torque it back up again, but this is concerning because it was torqued previously and 

uses a “self securing” nut, like most of the rest of the PROP system. 

The shaker is turned and the satellite is mounted for the z-axis shake. 

The z-axis vibrations and shocks are performed. 

Visual inspection of the spacecraft seems fine, and the high-pressure pressure transducer still 

reads at 302 bar. 

Overlays of pre- and post-vibration resonance searches match quite well in the y and z axes, 

there is just a small amount of “settling”. 

MILESTONE 19: SSETI Express passes its pressurised vibration tests. 

ESA_Neil gets on a plane to go home. 

PROP_Hanno, PROP_Sascha and SYS_Joerg dismount the spacecraft from the table, put it in 

the PROBA box and drive back to Stuttgart with it. 


PROBLEM 88: ESA_Cesar reports damage to the solar cell test panel and ESA_Neil goes 

to investigate. It appears that the cover glasses cracked upon depressurisation, although –40 

174






to +80 degree temperature cycles were also carried out. The string still responds well to 

impinging light, but a decision must be made as to whether or not to continue with the tests. 

ESA_Neil decides and advises ESA_Cesar to continue with the testing. This is reasonable 

because we are short on time and options, and we have several spare strings that we can 

depressurise again later once we have identified the failure point and attempted to implement 

a solution. 

PROP_Hanno (in Stuttgart) attempts to isolate and fix the mid-pressure leak in the PMS box. 


ARRIVAL 52: The flight structure and propulsion system arrive back in ESTEC. 

ESA_Neil and ESA_Marie move the flight structure into the cleanroom and mount it to the 

integration table. 

175






ESA_Neil cleans the flight structure and mounts the protective side panels to it. 

ESA_Neil sets up the groundstation and evaluates OBC_Karl’s test groundstation software. 

After a few teething problems it seems to work fine and telecommands can be sent 

successfully to the OBC. 

PROBLEM 88: When requesting X units of the ALARM stack, if there are less than X units 

then it simply repeats until X units are transmitted. 

PROBLEM 89: The timestamp on downlinked telemetry seems to fluctuate from a sensible 

time, to sometime in 2033. 


ESA_Jason glues components, wires and thermistors down in the MAGIC and UHF flight 

models. 

176






ESA_Neil continues functional testing with OBC and test ground station. 

SOLUTION (to problem 88): There is lots of stuff in the stacks in the flash RAM… 

PROBLEM 90: When requesting camera data the debugger receives “camera data sent” 

repeatedly and then the OBC crashes. 

PROBLEM 91: Using the “tcins” command (from the debugger) to flush the alarm stack 

causes the OBC to crash. 

(These are to be added to the current list of OBC issues: 

1) Occasionally get looping "de-kissisifcation error" down the debugger upon startup, and once we do 

it doesn't get better. This seems to be resolved once the linux box is rebooted. Could it possibly be 

from your side? (Probably is crap linux box though.) 

2) When using TCINS to send commands to EPS while looking down the EPS port with a laptop the 

parameters get swapped around. Is this another least-sig / most-sig issue, or is this intentional? (It is 

probably general to all tcins, but this was the case I discovered.) 

3) Upon startup of OBC the bytes "0f 6f 6f" get sent to EPS. Why is this? (It even appears in the 

flight planner briefly, I caught it once.) This is why EPS was shutting you down sometimes... now it 

just counts errors... SOLUTION: this is OBC telling EPS to turn ACDS on, it is not necessary, but not 

harmful either. 

4) I have just about found a way to uplink TCs to the OBC via the UHF, but it is still a little shaky (is 

only half tested, the RF link is not strong enough to be THAT reliable, and it has to end in a carriage 

return). I can get stuff into the flight planner quite well, and the right numbers seem to be appearing in 

the right places. Issue is: if I send a TC starting with "00" (i.e.: subsystem ID is OBC) then it does not 

acknowledge, does not go into the light plan, the nominal mode beacon stops, and it ignores anything 

incoming from UHF completely from then onwards. (Although it all looks fine on the debugger.) 

ALTHOUGH NEW SOFTWARE MUCH MORE RELIABLE FOR SENDING TCs, THE FACT THAT 

THIS IS POSSIBLE AT ALL IS WORRYING. 

5) TCs from UHF are overwritten, not stacked. This would make for a tedious mission.... ;-) 

177






6) It would be really good to have either everything ("time", time on "tmdata", time on "plist" etc) in 

EITHER linux time, OR real time. I don't actually mind too much which, but it is confusing that it 

changes a lot, as comparisons are then tedious and time consuming. Obviously real time is better for 

the general layman, but consistency is the most important thing. 

7) After adding commands on TCINS to be executed immediately (which they do), I have found them 

still listed in the flight planner later on, which is a tad odd. Doesn't a TC get taken out the planner 

when it's executed? This has not been a repeatable error... unfortunately. 

8) When requesting X units of the ALARM stack, if there are less than X units then it simply 

repeats until X units are transmitted. POTENTIALLY SOLVED (it’s coming from flash). 

9) The timestamp on downlinked telemetry seems to fluctuate from a sensible time, to 

sometime in 2033. NEEDS TO BE REPEATED AND DOCUMENTED PROPERLY. 

10) When requesting camera data the debugger receives “camera data sent” repeatedly and 

then the OBC crashes. 

11) Using the “tcins” command (from the debugger) to flush the alarm stack (and probably 

the HK stack too) causes the OBC to crash.) 

Commands that do work fine via UHF uplink: 

Non-valid (makes it into flight plan) 

GET_HK 

GET_AL 

SHUTDOWN 

Other testing is limited by the lack of EPS being fully functional. 

21 st November 2004 

UHF_Lars and ESA_Neil perform some quick functional testing using the test groundstation, 

FM OBC, FM UHF, FM PIN and FM EPS. 

ESA_Neil and UHF_Lars mount the lateral panels to the EM structure in order to ensure an 

accurate simulation. 

178






UHF_Lars “plays” with the network analyser and tests the response of the UHF antenna. 

22 nd November 2004 

ARRIVAL 53: The FM CAM returns, rebuilt, to ESTEC with CAM_Morten. 

ARRIVAL 54: The current MCC and GND laptops and software arrive. 

OBC_Karl uploads new software to the OBC, fixing problems 74, 77 78, 80, 81 and 91 

PROBLEM 92: The GND does not successfully converse with OBC. 

MODIFICATION 49: ACDS_Lars modifies EM lateral panels and coil-driver brackets to fit 

the coil driver PCB properly and to accommodate the sun-sensors and associated harness, and 

ACDS_Lars replaces connector on magnetometer. 

PROBLEM 93: ACDS magnetometer does not fit properly. 

179






LESSON LEARNED 15: Get every team to double check all mechanical interfaces 

explicitly with the relevant teams. 

PROBLEM 94: UHF_Lars needs a network analyser. 

LESSON LEARNED 16: Make sure that all required equipment is specified and provided 

before arrival and work of the teams, otherwise time will be wasted looking for it. 

FM CAM takes a picture and transfers to OBC. 

ESA_Neil, MCC, GND and OBC_Karl do functional testing on GND, MCC, UHF and OBC. 

MILESTONE 20: 2-way link established between MCC, GND, UHF and OBC 

PROBLEM 95: Still loosing packets and lots of erroneous data 

23 rd November 2004 

ARRIVAL 55: S-BAND comes back to ESTEC with AMS_Graham. Visual inspection 

reveals some gluing issues, but nothing too serious. 

PROP_Sascha tries to find leak with evacuation and helium sniffer, but fails 

180






PROBLEM 96: The branch valve is not responding properly, making leak testing impossible. 

It seems that the branch valve one-way only (back pressure leaks through). 

ACDS_Lars wraps the ACDS magnetorquer coils in Kapton tape. 

MODIFICATION 50: ACDS_Lars drills new holes into the magnetometer mounting plate. 

ACDS_Lars mounts the passive magnet into the EM structure and it fits perfectly. 

181






PROBLEM 97: A second line driver fails in the OBC. 

MODIFICATION 51: OBC_Karl once again replaces the dead line-driver. It is the same 

one as last time, therefore resulting in the third soldering of that part of the PCB. This is not 

really acceptable, but we have little choice. 

SOLUTION: Suddenly we realise that the galvanically isolated UHF and S-BAND boxes 

have two, different, totally floating, local grounds – both of which are used for comparison to 

the RS232 with OBC. This could cause all sorts of interesting problems, including the S- 

BAND power-up issue. 

ESA_Neil, OBC_Karl and AMS_Graham add grounding wires between the boxes and buzz 

them through to make sure that all the boxes are coupled together, as if the spacecraft is there, 

and that the subsystems that are meant to have galvanically isolated components do. 

SBAND now powers up from PCU. (Solves problem 60.) 

AMS_Graham and ESA_Neil perform functional testing with S-Band, which all goes rather 

well and no significant issues are identified. 

LESSON LEARNED 17: Make sure you get the grounding scheme implemented properly 

during the testing! On SSETI Express we forgot that the plastic table doesn’t work the same 

way as the metal spacecraft and lost a lot of time and effort because of it. 

PROBLEM 98: After extensive investigation OBC_Karl and CAM_Morten conclude that the 

byte errors could be coming from an overflow in the utility processor on the OBC. A new 

chip therefore needs to be programmed, burnt, and replaced. 

PROP_Sascha crimps connectors and sorts harness 

182







This is the first day of the third SSETI Express workshop. 

CAM functionally integrates with OBC 

PROBLEM 99: Some bytes dropped during transfer from CAM to OBC. 

SOLUTION: Partially fixed through software changes on OBC. Further needs to be done 

with new utility processor. 

PROBLEM 100: The FM CAM PCB grounding plane connects to the bolts that hold it to the 

box, therefore forming a ground loop through the EPS harness and then back through the 

structure. This would be very problematic, especially with the RF antennas so close. 

SOLTUION: The addition of Kapton tape above and below the PCB, and around the top of 

the bolt shafts, ensures isolation between the CAM components and the box. 

ESA_Neil and AMS_Graham test S-Band power-up repeatedly. It works fine and looks 

MUCH better than before on a scope. 

183






ACDS_Lars integrates the various parts of his system to the EM structure. 

PROBLEM 101: The ACDS coils are too small for the mounting pattern prepared on the 

lateral panels. 

SOLUTION: As long as the coils fit between the vertical mounting points then it is ok to 

bend them as they don’t get in the way of anything else. They will need to be glued to the 

panels anyway. This actually INCREASES efficiency of the coils, probably, since they will 

have a larger internal area. 

MODIFICATION 52: ACDS_Lars drills holes in EM lateral panel (+x) for the coil-holdingtie-wraps. 

SYS_Joerg and UHF_Lars move the EM back into the cleanroom. The cut antenna seems 

fine, although the return loss is not as good as expected. 

184






ACDS integrates magnetometer with OBC, goes well 

GND and AMS try to get RF link working between the cleanroom and Fr413. ESA_Neil 

plays with antennas to facilitate (lengthens cable and tapes UHF antenna to the window.) 

PROBLEM 102: Signal is just too weak for reliable reception at GND. (Although works 

fine the other way.) 

SOLUTION: Move groundstation back beneath the cleanroom. 

PROP_Sascha pressurises the mid- and low-pressure systems, still can’t find the leak. There 

is no sound and no bubbles, just a randomly sudden rapid pressure drop. 

PROP_Sascha sorts harness. 

MAGIC_Renato tries to reprogram box 

PROBLEM 103: MAGIC_Renato locks the FM MAGIC processor. 

EPS_Stefano programs a new PIC and inserts in the PCU. ESA_Neil and EPS_Stefano 

upload new software and attempt functional testing. Snooping on the port looks ok 

PROBLEM 104: No pings sent, OBC not powered up 

ATTEMPT: Re-program EPS with old code, fail 

ATTEMPT: Replace PIC with old one, (back to pre-EPS_Fulvio departure s/w) 

EPS_Stefano uploads new software. 

Functional tests done by ESA_Neil and EPS_Stefano seem ok. This implies that the original 

PIC was not broken (but the new one is). 

LESSON LEARNED 18: Handle microcontrollers carefully. One in the PCU broke within 

about 5 minutes of getting it out the wrapper. 

AMS_Graham demonstrates reception of AO51 out in the ESTEC car park using a simple 

handheld and a patch antenna. Signal is not great due to low pass, but voice of Danish Radio 

Amateur clearly heard. Everyone is happy and impressed. 

GND receives the nominal mode beacon from OBC and UHF. 

PROBLEM 105: MCC interprets the nominal mode beacon and it looks like rubbish. 

185






SOLUTION: MCC didn’t know beacon format was different from TM 

They fix it and it works fine. 


ACDS_Lars takes all ACDS off of EM STRU 

ESA_Neil, UHF_Lars, SYS_Joerg reassemble EM and make top plate realistic. 

SYS_Joerg and UHF_Lars transport EM outside. 

UHF_Lars, AMS_Graham and AMS_Howard test response of the antenna and gradually 

shorten it to tune it. 

ACDS_Lars integrates the coil driver with OBC. It seems to work fine but is not in full 

context without the other subsubsystems. 

186






EPS new test software (6.2) uploaded. This includes increased error-tolerance and debugging 

in the nominal mode beacon. 

PROBLEM 106: The new EPS software doesn’t work. At all. 

PROBLEM 107: S-Band carrier-up command suddenly doesn’t work 

SOLTUION: Revert to the previous software. 

PROBLEM 108: OBC won’t power up properly. 

SOLUTION: Yes it does, jus the laptop buffer is full and won’t display it 

SOLUTION: Wait until laptop fixed and nominal mode stable. Try s-band again – no 

problem. 

AMS_Graham tests output using EM STRU and EM LGA antenna. All seems good and third 

harmonic is now 50dB under main carrier. 

AMS_Graham performs further functional testing, including walking away to weaken UHF 

reception during transponding – no real difference, which is good. 

MAGIC_Renato looks for chip, finds one, hopefully comes tomorrow. 

MODIFICATION 53: ESA_Jason takes off the old magic chip. 

PROP_Sascha continues to search for leakage. Now that connectors are on the branch valve 

can be closed in combination with the fill and drain valve – proving that the leak is between 

fill and drain and branch (mid pressure). 

PROBLEM 109: Branch valve not responding. 

SOLUTION: A new connector fixes it 

CAM separates local and structural ground, and functionally integrates the camera to the 

OBC. Picture transfer seems ok. 

EPS_Stefano tries to fix software and works on definition issues with SYS_Joerg. 

ACDS coil driver is powered up via PCU and PIN. 

PROBLEM 110: ACDS power-up works but there is a slight bleed current when turned OFF 

(as noticed before), which slowly charges up the capacitors and bleeds into the system. 

187






CAM is powered up by the PIN. 

PROBLEM 111: CAM power-up works but there is a slight bleed current when turned OFF 

(as noticed before), which slowly charges up the capacitors and bleeds into the system, 

causing some oscillations. 

PROBLEM 112: OBC_Karl decided that he doesn’t like this (problems 110 and 111) very 

much either. 

REQUIRED MODIFICATION: Add pull-down resistors to the PIN box. OBC_Karl, 

ACDS_Lars and CAM_Morten discuss and advise 10 kilo ohm. 

PROBLEM 113: The CAM doesn’t fit in to the hole on the top of the spacecraft. 

REQUIRED MODIFICATION: The sides of the box should be milled down by about 1mm 

each so that the camera will fit into the hole. 

SYS_Joerg and PROP_Sascha pressurise mid-pressure part, close the fill-drain valve very 

tightly and remove the (leaking) ground half. Pressure seems stable – leave overnight. 

OBC_Karl and CAM_Morten prepare PCU power bypass hack. 

188






OBC_Karl fixes the call sign and webpage in the nominal mode beacon. 

OBC_Karl adds live streaming capability to downlink and tests briefly with GND and MCC. 

ACDS_Lars prepares for rest of integration and packs up his stuff. 

CAM takes picture and transfers to OBC. GND asks for it and downloads it. 

PROBLEM 114: GND miss packets – maybe this is still a utility processor buffer problem. 

GND download it again, and miss some different data. (Files don’t match.) 

We tidy up and go home, it is 2am. 


PROBLEM 115: GND and MCC interpret picture – there is so much missing it is not 

recognisable. 

OBC_Karl and ESA_Jason look for prom burner but fail. 

EPS_Stefano, ESA_Neil and ESA_Jason discuss work to be performed on the battery box. 

There is much still to be done, and a test battery is really needed. 

MAGIC_Renato and PROP_Sascha go and pick up the new processor. 

MODIFICATION 54: OBC_Karl solders the new MAGIC processor to the board. 

SYS_Joerg and PROP_Sascha potentially find leak: perhaps the fill-drain valve was not 

closed properly. Now it seems to hold pressure ok. 

189






LESSON LEARNED 19: Close all valves properly before leak testing. 

MODIFICATION 55: UHF_ON and UHF_OFF commands removed and replaced with 

UHF_CYCLE (much more sensible). This command was then added to the “satellite has not 

heard form the ground for 24 hours” routine ensuring that the UHF will be power cycled if 

necessary. 

OBC_Karl and MAGIC_Renato test FM MAGIC and FM OBC. It seems fine. 

Everyone tidies up and goes out to celebrate the end of the workshop. 


PROP_Sascha finalises harnessing of PROP. 

OBC_Karl works down his to-do list adding and adapting outstanding issues from workshop. 

EPS_Stefano and OBC_Karl work on ping problem, snooping down the line in both 

directions between EPS and OBC. OBC receives and responds to all pings fine, but EPS still 

sends shutdown commands. 

OBC_Karl, MAGIC_Renato, PROP_Sascha and ESA_Marie test out OBC, MAGIC and 

PROP. 

MILESTONE 21: The propulsion thrusters are fired via OBC and MAGIC for the first 

time and work perfectly. 


PROBLEM 116: OBC_Karl notices the occasional missing byte in EPS telemetry, 

OBC_Karl and EPS_Stefano make software changes to make communications more robust 

from the EPS and OBC side. This seems to resolve the ping issues and EPS keeps OBC alive 

for two 1.5-hour stints. 

OBC_Karl sets up debugging laptop in cleanroom for remote operation of OBC. 


ESA_Marie and ESA_Neil dismount the flight antenna and move the EM structure out to 

Einstein for a presentation. Afterwards they move it into the office for temporary storage. 

190







PROBLEM 117: The test solar panel emerges from the thermal vacuum chamber 

significantly damaged. 

ESA_Neil spends the rest of the day searching for an alternative source of solar panels. Best 

bet: buy TECSTAR cells from EMCORE, borrow solder and lay-down jigs from SSTL and 

get soldering gurus ESA_Jason and ???? to manufacture them. (Get enough spares to get it 

wrong a few times.) 

2 nd December 2004 

ESA_Jason and ESA_Neil define and list pending work to be done for each box. 

ESA_Jason starts work on the FM UHF. 

MODIFICATION 56: ESA_Neil drills / mills new holes into the side protectors to fit the 

lower right hand corners properly. (And the lower left hand corner of the –y protector.) 

191






ESA_Neil cleans protectors. 

3 rd December 2004 

ESA_Jason adds RS232 and Audio redundancy, replaces connectors and re-works the routing 

in the FM UHF box. 

ESA_Neil tests FM UHF with EPS, OBC and S-BAND, it all seems to work fine except for 

the PTT. 

ESA_Jason adds CAN redundancy and replaces and solders the thermistor wires in the FM 

MAGIC box. 

192






PROBLEM 118: Push-to-talk wired up like an RS232 in the UHF box. 

REQUIRED MODIFICATION: Re-do pin-out of PTT in FM UHF. 

6 th December 2004 

ACDS_Lars does all markings for holes that should be drilled in the FM. Also makes some 

test harness. 

ARRIVAL 56: The T-Pods arrive with CANX_Fred. 

ARRIVAL 57: The new Utility processor arrives with OBC_Karl. 

OBC_Karl replaces the utility processor. 

OBC_Karl improves groundstation software and downloads a picture via direct link (not 

involving UHF). Some bytes still lost, but no-where near as many. 

PROBLEM 119: 150ms turnaround on half-duplex 

POSSIBLE SOLUTION: Switch modem to higher TXD (200ms) 

ACDS_Lars integrates the passive magnet and the magnetometer to the flight model. 

193






CANX_Fred and ESA_Neil activate the test E-Box manually. 

PROBLEM 120: E-Box does not hold itself on like it should. 

SOLUTION: Reset pin was permanently depressed. 

ESA_Neil tries to test PCU timers. 

CANX_Fred inserts mass dummies into two T-Pods. 

194






CANX_Fred, ESA_Neil and ACDS_Lars activate the three T-Pods from the PCU, they fire 

satisfactorily. 

MILESTONE 22: The FM PCU fires the T-Pods for the first time 


ESA_Neil tests PCU timers. 

PROBLEM 121: No combination of specified wires produces the correct result. 23 and 24 

power all loads as normal, but also power T-Pods for 10 seconds initially. 23 and 25 eat 32 

milliamps permanently and do nothing else. 24 and 25 power T-Pods for ten seconds and 

then eat 32 milliamps permanently and does nothing else. 

195






MODIFICATION 57: ACDS_Lars and ESA_Neil drill holes in FM secondary structure to 

support ACDS integration. 

ACDS_Lars integrates all items to the FM lateral panels (nothing glued yet). 

ACDS_Lars and OBC_Karl debug OBC with regards to analogue I/O for ACDS. 

OBC_Karl downloaded “complete” picture from the spacecraft to the groundstation. 

PROBLEM 122: Black holes in picture (data lost during transmission). 

CANX_Fred and ESA_Neil ground the FM structure to the FM EPS. 

CANX_Fred and ESA_Neil integrate the T-Pods to the FM. Larger washers are used to 

secure them. All goes smoothly. 

196






PROBLEM 123: The upper part of the +y lateral panel interferes with the release 

mechanism of the +z T-Pod. 

REQUIRED MODIFICATION: Need to cut a bit out of the +y lateral panel to 

accommodate the release mechanism of the +z T-pod. 

PROBLEM 124: The +x and –x side protectors no long fit, as the T-Pods protrude more 

than planned. 

197






REQUIRED MODIFICATION: Need to cut sections out of the +x and –x side protectors 

to fit around the T-Pods. 

PROBLEM 125: Power-pin on analogue multiplexer not connected on ACDS coil-driver. 

MODIFICATION 58: OBC_Karl adds modification to ACDS coil-driver to correct 

analogue multiplexer problem. 

ACDS_Lars and OBC_Karl test the coil-driver, magnetometers and magnetorquers with the 

OBC. 


ACDS_Lars and OBC_Karl test and debug the sun-sensors with the coil-driver and computer. 

198






CANX_Fred trains ESA_Neil on loading procedure of T-Pods. 

199






ESA_Neil and CANX_Fred load mass dummies into two T-Pods. 

ESA_Neil mounts two side protectors to the flight model. (The only two that still fit.) 

OBC_Karl makes a help screen for his groundstation software. 

ACDS_Lars and OBC_Karl perform full functional testing of the entire ACDS system, in 

context, via the UHF groundstation. Everything goes smoothly. 


OBC_Karl and MCC_Krage get direct link from groundstation in ESTEC to MCC in Aalborg 

working. 

OBC_Karl adds features to his software. 


MODIFICATION 59: ESA_Jason mills sides of FM CAM down by 1mm each. 

MODIFICAITON 60: ESA_Jason mills hole for crystal out of OFM OBC box, and re-taps 

thread in OBC box. 

MODIFICATION 61: OBC_Karl replaces two resistors on main board and IF board. 

ESA_Neil bolts CAM back together and performs fit-check. Works fine. 

200






ESA_Neil performs functional testing on S-Band. 

OBC_Karl and ESA_Neil perform functional testing on all OBC telecommands via the 

groundstation. The following actions were identified. 

201







ARRIVAL 58: LGA back-shields and reinforced LGAs arrive with COMM_Damian. 

EPS_Fulvio and EPS_Tommy arrive and troubleshoot the PCU. 

SOLUTION: The ramp-up on the external power supply was not accurate compared to that 

of the BDR, and consequently was not starting the timers properly. 

ESA_Neil and COMM_Damian attempt to integrate LGAs and back-shields to the EM 

structure. 

PROBLEM 126: Back-shields do not fit through the holes in the +z LGA mounting plate 

and the top-plate, due to minimum radius in the corners of these holes when machining. 

MODIFICATION 62: ESA_Neil uses a Dremmel to “square” the corners of the holes in the 

EM and then the FM top-plates 

202






MODIFICATION 63: COMM_Damian uses a hand file to “square” the corners of the +z 

LGA mounting plate. 

PROBLEM 127: LGA holes don’t line up very well with the +z LGA mounting plate. 

MODIFICATION 64: ESA_Neil modifies an EM LGA with a Dremmel to fit the +z LGA 

mounting plate. Results are ok but COMM_Damian is not happy with the process. 

SOLUTION: By careful selection of LGA and mounting plate the differing tolerances allow 

a fit to be found. 

PROBLEM 128: The +z LGA mounting plate cannot be mounted to the top-plate as the 

washer on the +x+y lifting bolt is too wide. 

MODIFICATION 65: ESA_Neil uses a Dremmel to make a small cut-away in the FM +z 

LGA mounting plate so that it fits around the washer on the top-plate. 

COMM_Damian integrates RBF cap to the +z LGA. 

203






ESA_Neil integrates the +z LGA and RBF cap to the FM. 


ESA_Neil and SYS_Joerg re-iterate the state machine documentation. 

EPS_Fulvio assembles the battery box. 

EPS_Tommy tests the timers, steadily adjusting the resistance to get the time down to 74 

minutes. 

ESA_Jason, EPS_Fulvio and EPS_Tommy take the plastic sheathing off of the battery to 

inspect the inner workings 

PROBLEM 129: The battery uses one or two layers of thin plastic sheathing as insulation 

between cells. If this off-gasses and ‘dissolves’ then the battery will fail. Quite a lot of 

204






plastic insulation is also used on the wires connecting the small circuit boards on the top of 

each cell. 

MODIFICATION 66: EPS_Fulvio and EPS_Tommy take the battery apart and then reintegrate 

it, replacing plastic wiring with PTFE and plastic sheathing with kapton tape. They 

take many pictures throughout the process and refer to these to ensure that the cells are 

correctly connected. 

OBC_Karl makes the following suggestion for modification of the timers such that they only 

activate the first time the spacecraft activation switch is released. They can be reset manually 

under this scheme for the purposes of testing. 

205






OBC_Karl iterates the OBC software on the linux box in the cleanroom via the direct internet 

link with Aalborg. 

OBC_Karl iterates the test groundstation software and provides a copy to ESA_Neil. 


EPS_Fulvio and EPS_Tommy continue to re-integrate the battery. 

206






EPS_Tommy continues to test the timers, 5 attempts with the same resistance are within 90 

seconds of the target 74 minutes. 

ESA_Neil installs the new test groundstation software onto the test groundstation. 

ESA_Neil boots up the OBC to test the new software. 

EPS_Fulvio and EPS_Tommy test the reintegrated battery. It seems to work fine. 

PROBLEM 130: The OBC is not working at all, there is no TM, no flight plan execution, 

and a “tcins” causes it to crash. 

SOLUTION: Didn’t have the correct version of the software (without hardware flow 

control, since the utility processor here is different to the one in Aalborg until final prom 

insertion). 


ESA_Neil tests the new OBC software and test groundstation software. 

SOLUTION (to problem 130): The OBC is fine now. 

MILESTONE 23: Successful S-BAND data downlink for the first time 

Based on downloading picture chunks or large amounts of telemetry, S-band is about 50% 

faster than UHF. This is not as much as expected, but the utility processor limits it. Further 

tuning may be possible. 

NOTE: The test groundstation software does not run fast enough on the g/s laptop to keep up 

with the reception. To time the transmissions it is best to watch the DCD indicators on the 

TNC. 

207






Every second beacon is sent on S-Band. 

Commanding S-Band between voice and data modes whilst transmitting works fine (shuts off 

and comes back as expected.) 

File opening works ok. 

File deletion works ok. 

File protection works ok. 

PROBLEM 131: The OPEN_FILE command uses a target comms system parameter, which 

then means that only that comms system can be used with the GET_FILE command. This is 

too restrictive and should be the other way around. 

EPS_Fulvio and EPS_Tommy test their subsystem and develop a current source to test it 

further. 

208






SOLUTION: OBC_Karl swaps the target comms system from the OPEN command to the 

GET command. 

EPS_Fulvio and EPS_Tommy glue the battery cells together, and the lose wiring on the top 

down to the cells. 


EPS_Fulvio and EPS_Tommy make the EPS harness between FPP, ACT, PCU and BAT. It 

is a five-headed orange monster. 

EPS_Fulvio and EPS_Tommy test the harness they have manufactured, it seems to work fine. 

MILESTONE 24: The satellite is powered-up using the battery box for the first time. 

209






The timers activate the test ebox (and two buzzers) too early, after just 61 minutes. This 

could be because they have not discharged properly since the testing. 

All testing seems nominal. 

PROBLEM 132: When S-BAND attempts to bring the carrier up and transmit right away 

the power spike is too high and EPS cuts it off. This happens when the carrier is down and 

the system in is voice config and attempts to send telemetry, and it happens when the carrier 

is down and the system is in data config and attempts to send telemetry. 

PROBLEM 133: The battery box is at 14V instead of 0V. This is because the wall of one of 

the cells is touching the sides of the box through the kapton somewhere. 

MODIFICATION 67: EPS_Fulvio and EPS_Tommy line the battery box with kapton and 

recharge the batteries, and touch up some of the soldering on the BCR. This probably fixes 

the problem (132), but only testing in full context (using the structure) will tell for sure. 

12 th January 2005 

ARRIVAL 59: UWE-1 Engineering model. Visual checkout: all seems fine. Stored in lowhumidity 

cupboard in cleanroom. 

210






ARRIVAL 60: Solar cells from Emcore. Visual checkout impossible due to packaging – it 

is more risky to open the boxes now than to simply send them to Cannes for the lay-down and 

have them check them out there. One discrepancy: the quantities marked on the outside of the 

boxes add up to 162 cells, not the 155 expected. 

ARRIVAL 61: Box of stuff from SSTL, presumably lay-down jigs and fit-check and launch 

hardware. Will open and itemise tomorrow. 

ARRIVAL 62: NCube-II mass dummy arrives. 

PROBLEM 134: Visual inspection reveals apparent damage during transportation. Situation 

reported to the NCube-II team and am awaiting appropriate response. 

211







ESA_Neil integrates the lateral panels to the EM structure, the new 6mm M4 bolts fit much 

better to the side inserts, and the 12mm M4 bolts fit much better to the rivet nuts. The lateral 

panels are then much tighter fitting than before. (Although the same two bolts in the top-plate 

do not fit, hopefully this will be better in the FM.) 

ESA_Neil integrates two of the EM patch antennas to the EM structure. 

PROBLEM 135: The EM UHF antenna doesn’t fit in the mounting hole and is too long. 

MODIFICAITON 68: ESA_Neil modifies the EM UHF Antenna by cutting off both ends 

with a hacksaw, grinding the edges smooth and grinding the corners rounded to fit into the 

insert on the top-plate. 

ESA_Jason glues two harness clamps to the S-Band box. 

ESA_Neil mounts the EM UHF Antenna to the EM structure. 

ESA_Neil performs a simple fit-check with the activation switch plate. It fits perfectly. 

212







DEPARTURE 8: The damaged NCube-2 mass dummy is sent back to Norway for 

evaluation. 

ESA_Neil assembles the activation switch plate from SSTL. There are no instructions, but 

there are not many sensible options and the result appears to concur with drawings. (Note: 

two spare “otto” switches provided by SSTL are used, they are not four-pin like ours.) 

ESA_Marcel completes modification of side protectors. 

213






ESA_Neil requests FPP manufacture from ESA_Marcel. 

ESA_Neil installs the modified side protectors, they fit fine. 

ESA_Neil removes –y side panel and +x-y corner profile of EM. 

ESA_Neil and ESA_Jason install the activation switch plate into the EM. It fits perfectly, 

with the two switches and the battery charge stud protruding from the base as expected. 

214






ESA_Neil replaces the +x-y corner profile and the –y side panel of EM. 


ESA_Neil prepares RBF and ABF tags. 

ESA_Neil prepares the remaining foam “dummy” items and attaches them to the EM using 

double-sided tape. 

ESA_Ferry completes the EM FPP. 

ESA_Neil prepares the EM FPP – everything fits perfectly. NOTE: screw holders on the 

connectors are used to prevent lots of small object being present during manipulation of RBFs 

and ABFs. 

215







ESA_Ferry finishes the FM FPP 

ESA_Neil attaches RBF tags and ABF tags to all items apart from the antenna caps (in Poland 

for modification) and the side protectors). 

216






ESA_Neil glues the EM FPP to the EM. 

217







ESA_Jason removes the TNC from the lid of the S-Band box. 

ESA_Jason declares gluing complete for CAM. 

MODIFICATION 69: ESA_Bas redoes pin-out of PTT in FM UHF. 

ESA_Jason glues the last items in the FM UHF and S-Band into place securely. 

ESA_Neil removes support from the EM FPP and confirms that it is glued securely. 

218







ESA_Neil and ESA_Jason clear up the SSETI Express table and the laminar flow bench, and 

move all the SSETI Express hardware across onto the bench. This is to allow for the other 

users of the cleanroom to use the large table. 

MODIFICATION 70: ESA_Neil and ESA_Marcel mill the lid of the S-Band FM enclosure 

so that it fits around the power connector properly. ESA_Neil performs a fit check on the 

hardware and reinstalls the TNC into the lid. 

219






ESA_Neil removes the CanX mass dummy from the +z T-Pod and performs a fit check with 

the Xi-III engineering model. It is a very tight fit (the lid), but the safe-bolt just manages to 

hold the lid down so ESA_Neil closes and prepares the T-Pod. 

220






ESA_Neil integrates the UHF FM Antenna to the flight structure. It fits fine. 

PROBLEM 136: The UHF EM Antenna leans the wrong way compared to the FM. 

ESA_Neil performs a fit check with the –x side protector on the EM, including the lateral 

panel (first time in this configuration). It fits fine. 

221






ARRIVAL 59: The flight battery arrives and looks fine. 

ESA_Neil removes the UWE-1 mass dummy from the –x T-Pod and performs a fit check 

with the UWE-1 engineering model. It fits fine so ESA_Neil closes and prepares the T-Pod. 

222






NOTE: The springs in the two horizontal T-Pods have a propensity to drop slightly so that 

they interfere with the pushing-plate, therefore not allowing the Cubesat to slide back fully 

into place. Careful manipulation of the spring through the charging port to push it back into 

place solves the problem. 

31 st January 2005 

ESA_Neil dismantles the FM CAM box (including cycling the connectors), unplugs the 

camera itself and takes the two PCBs apart. 

ESA_Neil cleans the CAM PCBs and the MAGIC PCBs with IPA. 

ESA_Neil applies conformal coating (CV1152) to the “upper” sides of both camera PCBs, all 

three MAGIC PCBs, the BCR and the top of the EM battery. 

ESA_Neil places the CAM I/F board into the vacuum chamber to test the removal of the 

bubbles from the conformal coating. 

223






PROBLEM 137: The conformal coating appears to dry faster than it should, this is 

presumably because it is past its shelf life. Because it was too tacky there were air bubbles on 

the CAM I/F board that did not burst or deflate properly – the situation is not serious though. 

SOLUTION: It still seems fine, but in future we should place the PCBs into the vacuum 

chamber sooner and for a shorter time. This particular coating is very good at settling without 

air bubbles anyway. This is the only stock we have and there is no possibility to replace it. 

ESA_Neil leaves the work pieces to dry. Tomorrow he will apply the coating to the other 

sides of the same boards. 

1 st February 2005 

ESA_Neil touches up the conformal coating on the upper side of both main CAM boards. 

ESA_Neil applies conformal coating to the other side of all three MAGIC PCBs. 

ESA_Neil applies conformal coating to the board on the base of the camera. 

ESA_Neil places MAGIC and CAM into the vacuum oven to remove air bubbles from the 

conformal coating. 

ESA_Neil tidies the cleanroom, leaves MAGIC ‘exploded’ to dry, and sections-off the area, 

since he is going to be away for a week or two. 

ESA_Neil requests a list of cables to be made by ESA_Jason and ESA_Bas. Although we 

don’t know the exact lengths, it is useful to make one end already, especially as they need to 

be soldered since ESA_Jason does not have enough gold connectors. 

224






8 th February 

ESA_Neil and MCC_Green meet in Oslo, Norway. 

ESA_Neil is introduced to the MCC and given a tour by MCC_Green. ESA_Neil is suitably 

impressed, and the only feature that he can come up with that we might need but is not 

implemented is a ‘time critical flight plan’ feature, such that, if such a flight plan is not 

successfully uploaded during a particular pass, then an attempt is made to flush it before lossof-signal. 

MCC_Green and ESA_Neil travel to Tromsø, to meet with KSAT_Børre. 

MCC_Green and ESA_Neil iterate the design of the ground segment, proposing that the copy 

of the MCC DB in Vienna be an exact replica which handles all the queries. There would 

then be a second database, containing only the radio amateur telemetry (and friendly 

telecommands), located on the Vienna server. The team and public mission data interface 

pages would then query both of these databases together. 

9 th February 

ESA_Neil, MCC_Green and KSAT_Børre arrive at Svalsat, Longyearbyen, and inspect the 

Yagi antenna and associated equipment. 

225






PROBLEM 137: The PC in the radome controlling the antenna requires a login password 

and no-one in K-Sat or NCube can remember it. There is also a large time pressure, as the 

weather is closing in and the road back down from the plateau will soon be impassable. 

SOLUTION: We disconnect the locked PC and take it down to K-Sat Svalbard HQ in 

Longyearbyen. After several fruitless phone calls and attempts in vain to identify the correct 

password, ESA_Neil hacks into Windows 2000, resetting the Administrator password and 

gaining access. 

226






PROBLEM 138: It seems that the only software on the PC is Nova and NetOp. The MCC 

team are not keen to interface to either of these programs, rather than creating their own 

system. 

PROBLEM 139: It is not clear how the NCube guys are planning to control the radio, or 

how they are planning to stream the data back to Andøya. These two pieces of information 

are important to make sure that the SSETI Express MCC / GND interfacing solution does not 

interfere with the NCube setup, and vice versa. 

10 th February 

ESA_Neil, MCC_Green and KSAT_Børre in Svalsat, working with MCC_Krage and 

GND_Kreyst in Aalborg, attempt to establish a remote connection from Aalborg, using 

NetOp to command the PC controlling the Yagi antenna. 

PROBLEM 140: The various layers of firewall are getting in the way, and connection is not 

possible. The software at both ends is demonstrated to be correct by local connections. 

MCC_Green and ESA_Neil ‘play’ with Nova. 

The Nova application running on the Yagi PC can control the Yagi antenna automatically for 

every pass of a given satellite using standard two-line-element sets. ESA_Neil uses Nova and 

the Yagi to track AO-40 across the Svalbardian sky. 

PROBLEM 141: Although Nova can calculate the appropriate Doppler compensations 

during a pass, it does not appear to be able to control the radio directly. We therefore assume 

that NCube are not planning to adjust for Doppler, which leaves the SSETI Express teams 

with the requirement to come up with a solution. Subsequent telephone calls with the NCube 

team confirm this assumption. 

227






In the absence of a working NetOp solution, ESA_Neil, MCC_Green and KSAT_Børre in 

Svalsat, working with MCC_Krage and GND_Kreyst in Aalborg, attempt to establish a 

remote connection from Aalborg, using VNC to command the PC controlling the Yagi 

antenna. 

PROBLEM 142: The firewall at K-Sat is also removing the possibility of using VNC for the 

Aalborg – K-Sat connection. 

SOLUTION: K-Sat open a wide hole in the firewall for Aalborg. 

A remote connection between Aalborg and Svalsat is established, and the PC controlling the 

Yagi antenna can be fully controlled from Aalborg. 

KSAT_Børre advises us to leave the station, as the weather is threatening our route down to 

the town again. Heeding these wise words, we relocate to K-Sat HQ in Longyearbyen. 

ESA_Neil googles the Kenwood TS 2000 radio and Nova, and discovers the existence of the 

‘W6IHG Radio Tuner’, which can interface to Nova and control a Kenwood TS 2000 to adjust 

for the Doppler shift during a pass. This software is not immediately available however, and 

will have to be ordered from the US if required. 

MCC_Green phones the NCube team to ask how they are planning to stream data from the 

station back to Andøya. It turns out that they are simply planning to log data from the serial 

port (connected to the radio) to the hard drive using Hyperterminal, and will then FTP the data 

back home. 

PROBLEM 143: SSETI Express cannot use the same solution as the NCube team, as we 

require control of the radio during a pass to correct for the Doppler shift, but Hyperterminal 

would not allow such control of a port it is connected to. This point is somewhat moot 

however, since we also require the data to be streamed real-time back and forth. A solution 

therefore needs to be found that can switch between the NCube software setup and the SSETI 

Express software setup. 

MCC_Green discusses the various issues with his team. The main options appear to be the 

following: 

1) We use Nova just like the NCube guys do to control the antenna, but we also use the 

‘W6IHG Radio Tuner’ to correct for Doppler and control the radio. New software 

would need to be written to stream the downlink back to Aalborg in real-time. At the 

start and end of a SSETI Express communications session the TLE that Nova uses 

could be swapped, changing it from tracking NCube, to tracking SSETI Express. A 

similar file-swap may be able to control the ‘W6IHG Radio Tuner’ so switch between 

the two project settings. 

228






2) We use Nova just like the NCube guys do to control the antenna, but we write new 

software to correct for the Doppler shift, control the radio, and stream the data. This 

involves a lot of work, but avoids the port conflict problems of (1). Again, at the start 

and end of a SSETI Express communications session the TLE that Nova uses could be 

swapped, changing it from tracking NCube, to tracking SSETI Express. 

3) We could ignore Nova completely and write the whole lot with home-grown 

Aalborgian software. This, however, is a lot of work. 

The most promising idea seems to be number (2), but there is still much to discuss. 

ESA_Neil, SYS_Jörg, ACDS_Lars and EPS_Fulvio agree on the thresholds for the safe mode 

entry and exit levels. Various pending items arise regarding numbers for the efficiency of the 

battery, and several of the EPS components. 

MCC_Green, KSAT_Børre and ESA_Neil enjoy blended Norwegian beers and the worlds 

northern-most blues concert. 

11 th February 

In order to enable the continuation of configuration and testing work at the interface of the 

Yagi antenna in Svalsat and the Aalborg MCC, given their impending departure, ESA_Neil 

and MCC_Green decide to implement a webcam through which Aalborgians with appropriate 

control of the Yagi PC can view the control panels of the Yagi antenna rotator and the 

Kenwood TS 2000. 

KSAT_Børre obligingly provides an Axis networking webcam, while ESA_Neil makes an 

Ethernet cable. 

MCC_Krage, GND_Kreyst, ESA_Neil and MCC_Green test and demonstrate the upload of 

files to the Yagi PC via NetOp FTP. 

MCC_Krage, GND_Kreyst, ESA_Neil and MCC_Green test and demonstrate that the PC can 

be restarted remotely and that control can be required upon Windows startup. 

ESA_Neil and MCC_Green install, mount and carefully position the webcam so that the 

controls of the Yagi antenna rotator and the Kenwood TS 2000 are readable on the resulting 

picture (viewed via a web browser form any machine within the ksat.no domain, and therefore 

accessible via NetOp-ing to the Yagi PC). 

229






MCC_Krage, GND_Kreyst, ESA_Neil and MCC_Green all agree that the combination of full 

remote control and webcam views of the hardware should be sufficient to continue, and 

potentially complete, the GND / MCC work at Svalsat entirely from Aalborg. 

ESA_Neil, MCC_Krage and KSAT_Børre run for their plane back to the Norwegian 

mainland, and back to the sunlight. 

14 th February 2005 

ARRIVAL 60: The EPS external BCR arrives, with a spare. It seems fine (physically). 

ARRIVAL 61: A spare laptop for the launch campaign is generously donated by 

AMS_Graham. 

230







AMS_Graham and AMS_Sam arrive and inspect the modifications done on the S-Band unit. 

They are happy with the results. 

MODIFICATION 71: AMS_Sam opens the S-band FM box and de-solders the 

programming switch link. 

MODIFICATION 72: OBC_Karl uploads the latest version of the OBC software, including 

the S-Band / Debug hack to get around the data bottleneck problem. 

ESA_Neil reassembles the test groundstation. 

ESA_Neil boots up the OBC to test the latest software. He has to change the port speed of the 

terminal on the debugging laptop after running the new code, as it adjusts the port speed to 

become compatible with S-Band. 

PROBLEM 144: With the new OBC software no TM is generated, and no downlink at all. 

231






OBC_Karl fixes it and uploads again. 

PROBLEM 145: ESA_Neil boots up the OBC to test the latest software. The downlink 

works this time, but there is no TCS or ACDS TM. 

OBC_Karl enables the TCS and ACDS threads. 

ESA_Neil boots up the OBC to test the latest software. This time it seems to work fine, so 

the OBC debug port baud rate change from 38.4kb/s to 56.7kb/s seems to be successful. 

PROBLEM 146: There seem to be a lot of “Fetch Queue Timeouts SID (4)” and “Hex 

dump” errors coming from the OBC to the debugger. 

ESA_Neil prepares a spacecraft-to-laptop serial cable, and AMS_Graham attempts to 

reprogram the S-Band TNC to 56.7kb/s. 

PROBLEM 147: Two-way communication with the S-band TNC cannot be established from 

an external laptop. 

SOLUTION: We try a different laptop and a different USB-to-RS232 adapter. 

AMS_Graham successfully reprograms the TNC to 56.7kb/s. (As reported by the TNC 

itself.) 

MODIFICATION 73: AMS_Sam re-solders a link across the programming switch on the S- 

Band TNC. 

ESA_Neil boots up the OBC using the debugging port to upload software, then switches to S- 

Band on that port to test the data downlink. 

232






PROBLEM 148: After the baud rate and port change there is no data downlink from the S- 

Band unit. 

PROBLEM 149: After the baud rate and port change we are only getting every second 

beacon from UHF (every 36 seconds). 

PROBLEM 150: The uplink on UHF is extremely unreliable. 

SOLUTION (to 148): ESA_Neil switches to a wide filter on the I-COM radio used for S- 

Band downlink (via VHF conversion). 

S-Band data downlink seems fine, implying that the S-Band baud rate change hack has 

worked perfectly. 

SOLUTION (to 149): After much troubleshooting AMS_Sam suggests that the external 

power supply might be current limiting when both S-Band and UHF are trying to send 

simultaneously on every second beacon. A quick adjustment of the current limit confirms 

this, and every second beacon is then received via S-band and UHF. 

ESA_Neil, AMS_Graham, AMS_Sam and OBC_Karl test the data rate from S-Band 

downlink by downloading picture chunks from the satellite. Almost every single time 100 

packets out of 100 packets are received, and the raw Bps suggests that we a re close to the 

intended 38.4kbs data rate. 

PROBLEM 151: The occasional packet in the picture chunk downlink is erroneous, 

therefore forcing the test groundstation software to automatically request the entire chunk 

again. The result is that it is usually 5 or 6 cycles before the entire chunk comes down trouble 

free. Suggestion is to only re-request chunks once and then combine all good data. 

AMS_Sam conformal coats the RF side of the S-Band box. 

233






ESA_Neil and AMS_Graham attempt to troubleshoot the unreliability of the UHF uplink. 

The problem seems too severe to be caused by multi-pathing, the groundstation seems to be 

transmitting properly (confirmed by handhelds), and the spacecraft is receiving something, but 

not the correct data (“Received TM too long” alarm each time in AL queue). 

PROBLEM 152: The spacecraft stops responding entirely to the UHF uplink. 

SOLUTION: The UHF TNC has locked up, cycling the power fixes it. However, the uplink 

is still unreliable. 

AMS_Graham twiddles the focus on the cleanroom webcam, improving the quality of the 

image dramatically. 

AMS_Graham and AMS_Howard (on phone) investigate the possibility that we have fiddled 

with the ground UHF radio, but it seems to be fine. 

ESA_Neil skips both ends of the UHF and debugs by putting the groundstation laptop direct 

to the UHF port on the OBC. This not only works fine (so it is not the computer), but the 

“Fetch Queue Timeout SID (4)” errors stop. This strongly implies that there is something 

wrong with the TNC in the UHF box. 

AMS_Graham and AMS_Sam re-jig the test groundstation (taking the attenuation off of the 

radio and making sure that we are using radio A instead of B) while ESA_Neil rebuilds the 

UHF-OBC RS232 cable. When we plug everything back in again it works much better, with 

a reliability of around 80% on uplink, and something significantly lower than 1% error on 

downlink. 

AMS_Sam applies conformal coating to the other side of the S-Band box. 

After each set of conformal coating we have been using the vacuum chamber to remove air 

from the CV1152. AMS_Graham also uses the opportunity to subject the EM s-band patch 

antenna to a couple of small vacuum cycles. 

234







AMS_Graham, AMS_Sam and ESA_Neil power up the S-band unit for the first time after the 

conformal coating and thermal vacuum cycling. Initially everything seems fine. 

We test the data downlink via S-Band, which seems just as good as yesterday. 

PROBLEM 153: S-band is producing almost no power to the antennas. 

AMS_Sam opens the box and troubleshoots, finally deciding that there is a problem with the 

exciter, namely an MGA 82563 appears to be damaged (no output, and no input bias voltage 

present). 

AMS_Graham, AMS_Sam and ESA_Neil drive across to BFI Optilas BV in Alphen aan den 

Rijn to procure three replacement MGA 82563s. 

MODIFICATION 74: ESA_Jason removes the old MGA 82563 and replaces it with one of 

the new ones. This does not solve the problem. 

AMS_Graham submits the EM antenna to several thermal-vacuum cycles, giving a total of 

one vacuum cycle, and four thermal-vacuum cycles for that particular unit. It suffers no 

discernable damage. 

AMS_Sam and AMS_Graham continue to troubleshoot the S-Band unit. 

ESA_Neil applies conformal coating to the lower sides of both camera boards, and to the 

upper side of the ACDS coil driver. 

235






ESA_Neil fires up the EPS PCU for the first time in a long time – it works with no issue. 

AMS_Graham and AMS_Sam conclude that the filter in the exciter has probably been 

detuned by the conformal coating. They decide that they must take the unit back to the UK to 

replace the filter or the whole exciter, then they will return. 

236






However, the current consumption is almost unaffected, so we can still test if the PCU current 

limits the S-Band unit when it comes up from carrier-off to transmitting in one go (a worry 

that has been nagging at ESA_Neil since before the end of last year). 

PROBLEM 154: When the S-band unit attempts to go from carrier down to transmitting in 

one go the EPS PCU appears to cycle the power as it trips the current limiting circuit. 

Therefore the transmission is not successful, and the unit ends up in an unpredictable state 

(since the power down is so short that the unit does not fully reset). This doesn’t make much 

sense though, as no real extra current is used when transmitting as when the carrier comes up 

alone – no explanation is forthcoming for some time. 

AMS_Sam measures the frequency of the S-Band unit – it appears to not have been affected 

by the conformal coating. 

AMS_Graham points out that maybe the PCU actually IS limiting the current every time – but 

we wouldn’t notice with a normal “carrier up” command, since we are not actually 

transmitting anything. This explains a little of the mystery. 

AMS_Graham and ESA_Neil dismantle the +y lateral panel from the EM and remove the –x 

EM patch antenna and attach the EM back-shield to it. 

AMS_Sam tests the match of the antenna, back-shield, attenuation cap and a large metal 

reflector with the following results: 

Antenna: -20.8 dB 

Antenna and attenuation cap: -15.2 dB 

237






Antenna, attenuation cap and proximity of large metal reflector: 

-16.2 dB 

The conclusion is that the attenuation caps are very effective and will allow the S-band system 

to be turned on in the thermal-vacuum chamber. 

The following options, in order of preference, are identified for the resolution of the current 

limiting issue: 

1) The presence of a 100 microfarad capacitor across the power line between the 

MAX chip and the s-band unit is probably drawing some 300mA of inrush 

current unnecessarily. We might be able to remove this capacitor and solve the 

problem. (Maybe increasing noise, but this should not be a serious problem.) 

2) Add a choke, preferably in the harness, that can be glued to the wall of the 

spacecraft. It is a messy solution, but better than the next one. 

3) Modify the PSU, replacing the mosfet with an appropriate one. This would be 

very difficult as it is glued down heavily. 

4) Replace the PSU. 

ESA_Neil resolves to work on option (1) over the next few days, whereas the AMSAT-UK 

team will consider number (2) as a backup solution. 

DEPARTURE 9: The S-Band FM unit returns to the UK with AMS_Graham and 

AMS_Sam. 

238






ESA_Neil reintegrates the –x EM patch antenna and the +y lateral panel to the EM and tidies 

up the cleanroom. 


ESA_Neil reassembles the camera. 

ESA_Neil attempts to power the PCU up by one of the solar panel inputs, using a dummy 

load equal to the entire spacecraft powered on for nominal mode. 

PROBLEM 155: The PCU draws no current from the solar panel input. 

ESA_Neil reconfigures the groundstation laptop to act as an intermediary between the 

cleanroom groundstation and the Aalborg ground segment. 

239







ESA_Neil attempts to power up the PCU by a solar panel input again, this time using the 

large orange five-headed-monster from EPS. 

PROBLEM 156: The PCU draws power right to the limit, and powers the shunt resister even 

though the input is only 25V. 

MCC and ESA_Neil set up remote control of the groundstation in the cleanroom from 

Aalborg. It works fine. 

ESA_Neil powers up the camera. It draws the correct voltage. 

ESA_Neil commands the camera to add information to the housekeeping stack, it does so 

without a problem. 

ESA_Neil repeatedly sends the specified parameters to the camera, commands it to take a 

picture and commands it to transfer the thumbnails to the OBC. All of these activities seem to 

work fine. 

PROBLEM 157: The downlinked thumbnails from the OBC are entirely black. Perhaps the 

parameters are not being sent correctly. 

22 nd February 2005 

ESA_Neil tests the PCU with EPS_Tommy online: 

2) Connect a current source to one of the solar panel inputs (pin 6 of solar panel 

connector) 

3) Put voltmeter across it and adjust voltage to 35V 

4) Disconnect all other loads from EPS 

5) Connect the shunt resistor 

6) Turn on the current source, results: 

Current Source: up to 18.7V asymptotically at ~100mA 

Voltage source: 18.7V (even though it is off) 

Shunt resistor: 0V 

7) Add a load of 83 ohms to the UHF connection, results: 

Current Source: up to 7.17V at ~100mA 


Note: Timers not working! (EPS_Tommy suggests due to ramp-up.) 

8) Turn off the current source and apply 28V from a voltage source to pins 23 and 25 

of the solar panel connector (power feeding and power-before-timers respectively), 

as well, results: 

240






Voltage source: 85mA at 28V 


9) As (7) but now turn the current source on, results: 

Current source: 17.31V at ~100mA 


Shunt resistor: 2.2V 

Note: This looks as if the voltage regulator is not working properly 

10) As (8), but add load of 83 ohms, results: 

Current source: Oscillating from 9V to 15V at ~100mA, at about 2Hz 


Shunt resistor: Oscillating so that multimeter cannot read accurately 

11) Turn everything off, remove loads and disconnect voltage source from pins 23 

and 25. Reapply current source. (This differs from (5) because in (5) the voltage 

source was just turned off but was still connected.) Results: 

Current source: 26.4V at ~100mA 

Shunt resistor: 0.39V 

12) As in (10) but applying a load of 83 ohms, results: 

Current source: Oscillating as in (9) but at higher voltage 

Shunt resistor: Oscillating so that multimeter cannot read accurately 

PROBLEM 158: Timers and voltage regulator may not be working properly, but we need a 

bigger current source to check for sure (which we don’t have). 

ESA_Neil tracks down a bigger current source. It is huge, and requires re-wiring on the back 

plane to get it over 10V… 

ESA_Neil powers up the spacecraft and the groundstation so that MCC in Aalborg can ‘play’ 

via a remote connection. 


ESA_Neil finally gets the Keithley current source working properly. It turns out that the 

position of the decimal point on the readout determines the range (i.e.: entering “1.000” will 

give 1V in the 1V range, entering “01.00” will give 1V in the 10V range and entering “001.0” 

will give 1V in the 100V range). 

ARRIVAL 62: NCube-2 flight model arrives with Bjørn Pedersen. 

241






PROBLEM 159: The antennas of NCube-2 have deployed during transit. This is a onetime-only 

physical mechanism, controlled by a one-time-only hardware timer, and therefore 

cannot be reset easily. It appears that the remove-before-flight pin is faulty, and the 

spacecraft must have activated during transit. 

NCUBE_Bjørn stows the antennas and secures the deployment boxes temporarily with tape. 

He also opens the Cubesat to investigate the problem – it seems the deployment was via the 

software, so the spacecraft must have activated. 

ESA_Neil opens the +x T-Pod. 

ESA_Neil and NCUBE_Bjørn attempt a fit-check with NCube-2 into the +x T-Pod. 

PROBLEM 160: The remove-before-flight pin on NCube-2 protrudes too much and 

interferes with the wall of the T-Pod, therefore not allowing integration. 

242






MODIFICATION 75: NCUBE_Bjørn cuts the end of the remove-before-flight pin off to 

shorten it. 

ESA_Neil and NCUBE_Bjørn attempt a fit-check with NCube-2 into the +x T-Pod. The fit is 

good on all sides, and the access port does allow removal of the RBF pin and connection of 

the checkout and power umbilicals. Perhaps the pads on the T-Pod door should be made a 

little thicker in order to ensure that the kill-switches are fully depressed. 

NCUBE_Bjørn packs NCube-2 back into the transit case while ESA_Neil closes the +x T- 

Pod and restores the side protector. 

DEPARTURE 10: NCube-2 departs with NCUBE_Bjørn in order for the antenna 

deployment mechanisms to be reset, the flight pin re-worked and replaced, and one crack in 

the solar panels examined and corrected if necessary. 

243






Note: A teleconference with the NCube-2 team concludes that they will return the flight 

model before the 14 th of March. 


ESA_Neil and EPS_Tommy (remote) perform the following testing on the PCU: 

1) Power feeding, power-before-timers and power-after-timers are all 

disconnected 

2) Pin 23 of the panel connector is grounded as usual 

3) The shunt resistor is connected with a voltmeter across it 

4) The current source is set up to provide 250mA, along with a compliance 

voltage of 33V 

5) A dummy load of 83 ohms is connected to the current source and it is turned 

on. Results: 

Current source current: 248mA 

Current source potential: 20.6V 

Comments: This implies that the current source is working fine 

6) The current source is turned off, disconnected from the dummy load and 

connected to pin 6 of the panels connector. Then the current source is turned 

on. Results: 



Shunt resistor potential: 4.08V 

Comments: EPS_Tommy: “?!” 

7) While still running the current source is raised to 300mA. Results: 



Shunt resistor potential: 5.27 

Comments: EPS_Tommy: “?!?!” 

244






8) While still running the current source is lowered to 200mA. Results: 



Shunt resistor potential: 3.03V 

Comments: EPS_Tommy: “?!?!?!” 

ESA_Neil: “So?” 

EPS_Tommy: “It definitively doesn’t work!” 

ESA_Neil: “Get on the next plane.” 

ESA_Neil boots up the spacecraft, boots up the groundstation, connects the groundstation to 

the LAN, runs the test groundstation software, and enables the ‘Aalborg server’ connection. 

ESA_Neil connects the camera to the PIN and the OBC so that CAM_Morten can perform a 

checkout via MCC in Aalborg if he is available. 

1 st March 2005 

ARRIVAL 63: The S-Band FM arrives (again) with AMS_David and AMS_Sam 

ESA_Neil powers up the groundstation, the UHF and the OBC for the purposes of testing the 

S-Band FM with AMS_David and AMS_Sam. 

- The initial power consumption of S-Band FM seems good 

- The carrier is brought up successfully with the usual command 

- The telemetry is turned on and received without issue 

- The unit is switched to data config and data is transmitted without issue 

- The carrier is brought back down 



245






AMS_Sam torques all of the bolts on the RF side of the unit and glues them into place, along 

with a couple of remaining components, including the new filter. 

The S-Band FM unit is re-tested as above and it performs fine, so it is left overnight for the 

glue to dry. 

2 nd March 2005 

ESA_Neil powers up the groundstation, the UHF and the OBC for the purposes of testing the 

S-Band FM with AMS_David and AMS_Sam. 

The initial power consumption of S-Band FM seems good 

The carrier is brought up successfully with the usual command 

The telemetry is turned on and received without issue 

The unit is switched to data config and data is transmitted without issue 

The carrier is brought back down 

The telemetry is turned on and received without issue 

The unit is switched to data config and data is transmitted without issue 

We can therefore conclude that the glue has not damaged the filter at all. 

EPS_Tommy and ESA_Neil puts the PCU into the loop and re-power the UHF, OBC, PCU 

and S-BAND in order to replicate the current limiting problem (#132, 17/12/2004). 

Having successfully replicated problem 132 AMS_David, AMS_Sam, EPS_Tommy and 

ESA_Neil use a digital storing oscilloscope to measure exactly the parameters of the current 

spike that is causing the problem. 

246






The initial spike and the 150ms current drop-out can clearly be seen. We “zoom in” to try to 

see the exact nature of the spike, and we find that the peak is right up at around 2A. Below on 

the left is the spike when it is current limited, and on the right is without the PCU in the loop 

(where the peak is only 1.75A): 

AMS_David prepares the first test of a 15mH choke that might relieve the problem. 

247






Below on the left is the new system tested with current limiting, and on the right is it tested 

without current limiting. It can clearly be seen that the noise has been reduced dramatically, 

but the peak has not reduced very much. The initial peak, with current limiting, is about 

1.75A with a width of around 700 microseconds. Without, it is around 1.5A. 

Although we are fairly sure that this solution is becoming impractical it is still important to 

characterise it, therefore AMS_David and AMS_Sam prepare a bigger, 51mH choke. 

However, the inductance now has gone so high that the switch-mode power supply in the S- 

Band unit starts to fight with it and only wild oscillation is achieved. This demonstrates that a 

choke is not a viable solution, and, since we cannot change the power supply in S-Band, the 

only remaining solution is to modify the EPS PSU. 

248






MODIFICATION 76: EPS_Tommy cuts the legs off of the original mosfet that is limiting 

the s-band current and temporarily solders a replacement mosfet with a lower resistance for 

the purposes of testing. 

A current limit of 4A is set (which is high enough for us to be sure that it is not restrictive, 

but in the case of a short in s-band would only be active for 4 microseconds every 150ms (due 

to the max chip) so we are not worried about the potential power drain). 

The results of the new current limit are demonstrated and tested with S-Band. Now, the 

carrier can come up without a problem, telemetry is transmitted and data is transmitted, both 

with the carrier initially down. There is still a lot of noise on the line, but AMS_Sam and 

AMS_David are not concerned. This temporary modification therefore appears to be a 

success. 

249






EPS_Tommy takes the PCU and begins to troubleshoot the voltage regulator. 

AMS_Sam removes various pieces of debris from the S-Band box with a pair of tweezers, and 

then torques and glues all the bolts on the digital side. 

MODIFICATION 77: EPS_Tommy removes the temporary mosfet by de-soldering the 

original pins, prepares a new mosfet by bending the legs into position and glues it into place 

on top of the old one. (There is no possibility to remove the old one from the board.) 

3 rd March 2005 

AMS_David uses a scalpel to trim off the edge of the harness clamp that is slightly proud 

from the top of the S-Band enclosure. 

AMS_David and AMS_Sam torque all the SMA connectors in the box. 

250






ESA_Neil applies small amounts (to avoid flow) of glue to the SMA connectors (rotator to 

thread), between the semi-rigid cables and the enclosure, to the small bolts on the interior of 

the SMA connectors and to the screw locks on the d-sub connectors. 

MODIFICATION 78: EPS_Tommy replaces a potentially faulty operational amplifier on 

the EPS FM PSU. 

EPS_Tommy tests the modified voltage regulator. 

ESA_Neil and AMS_David clean the bolts using an ultrasound bath, and the base plate of the 

S-Band unit, and begin to integrate. 

PROBLEM 161: There are not quite enough bolts to integrate the S-Band baseplate (four 

M2x6mm short), or the S-Band lid (one M2.5x6mm damaged). 

SOLUTION: Five cross-head replacement bolts are used. These are placed symmetrically 

nearest the four corners on the short side of the enclosure on the baseplate, and near one end 

of the lid. 

Due to no defined “optimum” torque, and the absence of a sensitive enough torque wrench, 

AMS_David tightens the bolts by hand. No glue can be applied to these ones as it is 

important for the surface of the box to sit flush onto the spacecraft structure. 

The box is overturned and AMS_David applies glue to the sole remaining “loose” coil, to the 

two remaining bolt heads, and to a couple of wires as they leave the PSU. 

ESA_Neil applies the tie wrap to the lid and the TNC wiring, then, with AMS_David, puts the 

lid on and tightens the bolts. 

ESA_Neil applies small amounts of glue to each of the bolts on the sides of the lid, to the 

small bolts on the exterior SMA connectors, and to the TNC bolts on the top of the lid. 

251






ESA_Neil weighs the S-Band FM unit at 1459g, including two 9-pin d-sub savers, one 15-pin 

d-sub saver, one SMA saver, and two SMA dummy loads. 

ESA_Neil, AMS_David and AMS_Sam power up the OBC, UHF and S-BAND for the 

purposes of testing. 

- The initial power consumption of S-Band FM seems good 

- The carrier is brought up successfully with the usual command 



- The carrier is brought back down 



- Transponding is tested without issue (remember narrow filter and increase gain!) 

MILESTONE 25: The S-Band sub-system is declared flight-ready. 

252






AMS_Sam prepares a T7F as a possible replacement of the test groundstation radio, but we 

can’t seem to get the cabling correct and the AMSAT-UK team run out of time. 

AMS_Sam tunes the transmission output level of port one on the test groundstation TNC. 

This does seem to alter the uplink reliability, so fine tuning is required at a later date. The pot 

is left rotated by half a turn anti-clockwise 

AMS_David and AMS_Sam depart for the United Kingdom. 

EPS_Tommy declares the voltage regulator to be fixed. 

EPS_Tommy demonstrates the voltage regulator to ESA_Neil using various currents from a 

constant current source being fed into panel number 1 (pin 7 of the panel connector). At each 

current above 57mA (the current needed to power the timers, which are currently the only 

“load”) the output from the current source is around 29V. Below 57mA the voltage drops to 

zero, as there is not enough power to run the “load”. As the current input increases beyond 

57mA the corresponding excess power is dissipated in the shunt resistor. The demonstration 

is taken up to 800mA. 

NOTE: The protection diodes are still in the PCU, which makes these readings about 0.5V 

higher than they will be finally. 

As per discussions with ALCATEL_Samson, ESA_Neil asks EPS_Tommy to remove the 

protection diodes from the solar panels (lateral panels) in the PCU, as we will place protection 

diodes at the top of each string and there is no need to waste unnecessary power by doing so 

twice. 

MODIFICATION 79 EPS_Tommy removes the diodes and solders wire shorts across their 

previous locations instead. 

EPS_Tommy and ESA_Neil prepare a simple list of tests to perform on the “completed” EPS 

system. However, these tests are not possible without the timer modification (defined on the 

14 th December 2004), and the latest PDU software. 

MODIFICATION 80: ESA_Neil and EPS_Tommy upload version 6.8 of the PDU software 

to the PDU PIC. 

EPS_Tommy tests the timer modification (defined on the 14 th December 2004) on the EM. 

4 th March 2005 

EPS_Tommy tests the timer modification (14/12/2004) on an engineering model. 

253






PROBLEM 162: The power cut during eclipse is too long for the timers (tested for 40 

minutes), and the capacitor discharges, meaning that the timers would run again once in 

sunlight. 

SOLUTION: A larger capacitor should fix the problem. 

EPS_Tommy retests the timer power-off delay. After 40 minutes the timers still do not 

restart. This is about the same length as an eclipse, so we can conclude that the modification 

is successful. 

MODIFICATION 81: EPS_Tommy modifies the timers on the flight board. 

254






PROBLEM 163: The duration of the timers has been affected by the modification, it is now 

around 85 minutes instead of 74 minutes. 

EPS_Tommy sets about modifying the resistor that controls the duration of the timer. 

NOTE: It is dangerous to touch the timer reset wire to ground while the PCU is powered. 

This could damage the board irreparably. 

PENDING MODIFICATION: EPS_Tommy will put a resistor in the timer reset wire. 

EPS_Tommy gives a demonstration of the new timer functionality to ESA_Neil: 

- Upon activation of the PCU the timers start counting (nothing perceptible 

happens) 

- The PCU is powered down 

- Upon re-activation of the PCU the T-Pods fire immediately (the buzzer in place of 

them sounds) 



them sounds) 


- The timer reset wire is touched to ground 

- Upon re-activation of the PCU the timers start counting (nothing perceptible 

happens) 



them sounds) 

CONCLUSION: The timer modification works fine. 

ESA_Neil takes ex-Nuna-II solar panels up to the cleanroom for EPS_Tommy to fashion a 

test SSETI Express solar panel from. 

ESA_Neil and EPS_Tommy discuss the optimum location for the reset wire. It is decided 

that it should be located on one of the spare pins from the solar panel / external feeding 

connector. Then it will be harnessed through to the external power supply connector on the 

FPP for ease of access. 


EPS_Tommy tests the timer duration. It is still out, by four minutes too fast, so the resistor 

must be changed again. 

255






MODIFICATION 82: EPS_Tommy adds a safety resistor to the timer reset wire, and then 

crimps it to pin 1 of the solar panel connector, from which it can run to the FPP. 

EPS_Tommy and ESA_Neil make a few new cables and connect up the PCU, BATT, EM E- 

BOX, OBC, CAM, PIN, UHF and S-BAND. The Keithley current source is used to simulate 

solar panels, the battery output current is monitored on an ammeter and the shunt resistor 

voltage on a voltmeter. 

Test 

1) EPS_Tommy connects the “activation 

switch” (a pair of d-subs right now) to turn 

the spacecraft on 

2) EPS_Tommy toggles the activation switch 

3) ESA_Neil resets the e-box and 

EPS_Tommy toggles the activation switch 

4) EPS_Tommy disconnects the activation 

switch and uses a spare wire to discharge the 

capacitor. Then he re-connects the activation 

switch 

5) EPS_Tommy toggles the activation switch 

6) ESA_Neil attempts to listen to the 

recovery mode beacon on a handheld radio 

with the squelch open 

7) ESA_Neil powers up the fast digital scope 

to try to view the beacon on the PTT line 

from EPS to UHF 

8) EPS_Tommy and ESA_Neil disconnect 

the measurement cable from the battery box 

in order to force a drop to safe-mode 

9) In an effort to remove the noise on the line 

EPS_Tommy powers the cable with 3V on 

pin 1 to simulate a known battery level 

Results 

Everything seems fine, the spacecraft timers 

run and the current consumption is normal. 

The spacecraft powers up and bypasses the 

timers as it should. The e-box is activated, 

then the spacecraft enters recovery mode 




The spacecraft timers run and the current 

consumption is normal. 




Something a lot like the beacon can be heard, 

but it is not conclusive since it is not easy to 

hold the squelch open for long periods of 

time. 

The beacon does appear on the scope, but the 

pattern does not seem correct or consistent. 

The beacon does appear on the scope, but the 

pattern does not seem correct or consistent. 

Looks ok initially but then things start to go a 

bit strange 

PROBLEM 164: The PCU collapses from recovery mode and consumes 150mA without 

powering any loads at all. Perhaps a loose cable touched a damaging potential. 

In an attempt to troubleshoot the problem EPS_Tommy toggles the activation switch, but 

nothing changes. 

256






EPS_Tommy tests voltages at several points inside the PCU. 

PROBLEM 165: EPS_Tommy discovers that the linear voltage regulator powering the PDU 

is extremely hot. ESA_Neil switches off the power immediately. 

MODIFICATION 83: EPS_Tommy removes the linear voltage regulator powering the PDU 

from the PSU. 

EPS_Tommy feeds the PDU directly with 5V. 

PROBLEM 166: The PDU is consuming 235mA (it should be around 30mA). EPS_Tommy 

turns the power off. This implies that something on the PDU has failed, probably the PIC. 

MODIFICATION 84: EPS_Tommy removes the PIC. 

EPS_Tommy feeds the PDU directly with 5V, it consumes almost nothing, but is presenting 

the correct voltages at the correct places. 

ESA_Neil and EPS_Tommy search for the spare PICs and find them under the EM. 

EPS_Tommy investigates to see if the spare PICs already have the bootloader on them. 

PROBLEM 167: The spare PICs do not have the bootloader on them. 

EPS_Tommy starts to design his own prom burner. 


PROBLEM 168: EPS_Tommy continues to attempt to burn the PIC, but has no luck. 


OBC_Karl arrives with a prom burner, which is eagerly leapt upon by EPS_Tommy. 

OBC_Karl and EPS_Tommy burn the bootloader onto a PIC. 

EPS_Tommy uploads the PDU software version 6.8 on the PDU and powers it up. The 

current consumption seems ok. 

257






MODIFICATION 85: EPS_Tommy re-solders the linear voltage regulator on the PSU for 

power feeding of the PDU. 

Timer duration is tested and found to be within a few seconds of 74 minutes. 

The timer skip functionality is tested and found to work fine. A resistor is used between the 

enclosure and pin 1 of the panels connector to discharge the capacitor. Care should be taken 

not to discharge it while the PDU is powered. 

The T-Pod pulse is activating the EM E-Box with no problem. 

PROBLEM 169: We cannot test the safe-mode functionality as EPS team are not totally 

sure how to force and simulate safe mode at present. 

We measure the recovery mode beacon using a fast digital oscilloscope and the bits are 

clearly visible as 100ms pulses. Each beacon should start with ON, ON, OFF, then proceed to 

give a 10-bit number for the battery voltage, and then give an ON for safe-mode or an OFF 

for recovery mode – therefore giving a total of 14 bits. (NOTE: For the flight version it will 

be a 12-bit number.) 

258






With the following input voltages (simulated via the BATTPCU measurement connector 

and a power supply) we obtained the following corresponding beacons: 

Voltage 

Beacon 

4.992 11011111111111 

4.487 11011000111111 

4.847 11001011011111 

4.847 11011011011111 

4.576 11010110101111 

4.576 11000100101111 

4.576 11001000101111 

4.576 11011100101111 

4.576 11000100101111 

3.001 11010100110011 

3.001 11010100110011 

2.000 11010100101101 

1.198 11011110111001 

1.302 11000001000101 

NOTE: The spacecraft is not entering safe mode because it assumed from a zero 

measurement at turn on that the batteries were defective. 

It seems that the four least significant bits (chronologically first) are random noise. This 

corresponds to 16 parts in 1024, which is approximately 1.5% noise on the reading, which is 

perfectly acceptable. 

ACDS_Lars tests the fit of the –y lateral panel to spacecraft when it has the coil and harness 

taped into place. It fits well. 

259






ACDS_Lars glues coil number 1 to the –y lateral panel. 

PROBLEM 170: The recovery mode beacon is not being transmitted by the radio (there are 

no power spikes and it cannot be heard via a handheld). 

PROBLEM 171: It turns out that problem 170 is due to an incorrect pinout on the PTT 

connector on the UHF FM enclosure. Tests show that pins 1 and 6 are ground and 2 and 7 are 

positive. Instead it should be 1 and 2 are positive and 6 and 7 are ground. This means that 

the PCU has been short circuiting itself every time it pushed the PTT and this could be why 

the PIC died. 

MODIFICATION 86: Rather than modify the hard-to-access UHF box, the harness for the 

PTT link is redesigned. The new pinout uses ground on pins 1 and 6 and positive 5V on pins 

2 and 7. The new cable is symmetrical, and it is clearly labelled. 

ESA_Neil tests activating the PTT directly with a 5V power supply. The results are positive 

as verified via a handheld radio. 

ESA_Neil applies the new test harness to PCU and UHF boxes and waits for the beacon on 

the oscilloscope and radio. The results are positive. 

PROLEM 172: The +x lateral does not fit right, even when we loosen the corner profile. 

The same two bolts as the EM are misaligned (centre +z) and the also +y corner. 

260






PENDING MODIFICATION: The bolt holes in the +x lateral panel must be slotted to 

allow them to be bolted to the top-plate. 

ACDS_Lars torques and glues the bolts on the coil driver. 

The ping delays between the PCU and the OBC are demonstrated to be as designed. (Power 

on, wait one minute, ping, wait one minute for pong, ping, wait one minute for pong, 

shutdown, wait 5 seconds, power off, wait one minute, power on… ad infinitum (hopefully 

not)). 

ESA_Neil applies conformal coating to the back of the ACDS coil driver and removes air 

with the vacuum oven. 

The following strange sequence of results are obtained when testing with the OBC and EPS: 

261






OBC up, ping received and pong sent, ping received and pong sent, shutdown from EPS 


OBC up, ping received and pong sent, TC to EPS to shut down S-band, s-band shutdown, 

ping received and pong sent, shutdown from EPS 

TEST: We emulate the PCU with a laptop – the OBC is responding to pings properly. 

The sequence is repeated for confirmation: 



OBC up, ping received and pong sent, TC to EPS to shut down S-band, s-band shutdown, 

ping received and pong sent, shutdown from EPS 

PROBLEM 173: Although two-way communication between the PCU and OBC is working 

(pings received by OBC and TC received by EPS), the PCU is not “hearing” the pongs sent 

by the OBC. This implies that the EPS software is incorrect. 


After looking at s/w EPS_Tommy manages to get it into safe mode (must turn both the PCU 

and the battery measurement simulation on, simultaneously, then lower the voltage below the 

safe mode entry threshold for 30 seconds). 

ESA_Neil and EPS_Tommy use a scope and a handheld radio to review the safe mode and 

recovery mode beacons. The following results are obtained: 

Voltage Beacon Interval Conclusions 

3.631 11010100111010 30 

3.631 11000100111010 30 

3.631 11011100111010 30 

3.631 11011100111010 30 

4.001 11011001100110 30 

4.001 11010001100110 30 Safe mode beacon is ok 

4.200 11000011010110 120 

4.200 11011011010111 120 Safe mode exit level is below 4.2V 

4.200 11010111010111 120 

4.100 11010010010111 120 Safe mode exit level is below 4.1V 

4.050 11001011100111 120 Recovery mode beacon is ok 

3.798 11000010000111 120 

3.753 11001111111011 120 

262






3.703 11000101111010 30 3.7V < Safe mode entry level < 3.75V 

3.703 11000101111010 30 

4.000 11000001100110 30 

4.000 11010001100110 30 

4.050 11001011100110 30 

4.050 11011011100110 30 

4.100 11010100010110 2m 

4.100 11000010010111 2m 4.05V < safe mode exit level < 4.1V 

ESA_Neil powers the OBC, UHF and S-BAND using the Alternate Power Supply (APS, that 

25-pin connector that powers the PIN directly from a power supply). 

OBC_Karl and ESA_Neil bring up the latest version of the OBC software and attempt a data 

downlink on S-Band, monitoring the speed using OBC_Karl’s term.exe. 

The top speed “raw” read by term.exe is 4500 Bps, which is equivalent to 36 kbps. 

The entry and exit safe mode levels measured are confirmed by EPS_Fulvio. 

PROBLEM 174: We appear to be loosing packets on the S-Band downlink, implying that 

the OBC is streaming data to the S-BAND unit too fast and we are overflowing the ‘leaking 

bucket algorithm’. 

TEST: OBC_Karl slows down the OBC->S-BAND rate to 35/38 th s of what it was, but we 

still lose packets. 

OBC_Karl and ESA_Neil set up a direct connection to the S-Band port on the OBC using a 

laptop and term.exe with the ‘direct’ connection check-box ticked. 

SOLUTION (to 174): The OBC is transmitting data to SBAND at approximately 45 kbps. 

This is too fast and the ‘bucket’ is overflowing. 

OBC_Karl attempts to slow the speed down to just under 38k4 bps. 

PROBLEM 175: The OBC cannot give data to S-Band at “any” speed, because the possible 

sleep times in between packets are only integer multiples of 10ms (which is too long). 

EPS_Tommy uploads PDU software version 6.3 to the PCU to test the watchdog. ESA_Neil 

boots up the spacecraft and it works fine: 4 pings -> stable. This demonstrates that some 

rework needs to be done on version 6.8 to make the pings more like 6.3. 

OBC_Karl attempts the transmission of a picture from the camera to the OBC whilst 

emulating the PCU with a laptop. 

263






PROBLEM 176: The pings from the PCU are not answered by the OBC while it is busy 

storing the newly transferred picture into flash memory. 

MODIFICATION 87: OBC_Karl fixes problem 176 by creating a ‘flash storage’ thread 

which loads the picture into flash memory but sleeps occasionally to allow the other threads 

to deal with the pings from the PCU. 

EPS_Tommy tests some new software. It does the following: 

Safe mode 

Recovery mode 

OBC power up 

ping pong 

ping pong 

OBC shutdown 

OBC power off 

OBC power up 

ping pong 

ping pong 

OBC power off (NO shutdown) 

PROBLEM 177: ACDS_Lars needs an unobscured OBC tomorrow, but OBC cannot be 

unobscured until radio tunings are complete, but radio tunings cannot be completed while 

EPS is using the system. 

ACDS_Lars ports ACDS code to a PC and makes special test harness to interface the 

magnetometer to the code through the PC. This is due to the unavailability of the OBC for 

testing. Also develops "back-door" to allow tele-commands to be issued to ACDS during runtime 

on the PC. 

ESA_Neil and ACDS_Lars convince the workshop to produce a L-profile to mount the 

ACDS extra connector to the panel. 

ACDS_Lars tests for correct alignment between magnetometer axes and the spacecraft. 

Correct, however all signs were changed in the software to adopt a "North pole is positive" 

convention. 

ACDS_Lars estimates the magnetometer bias induced by the passive magnet by doing 

measurements with the magnet in place and without. Results: 

X_bias = -3934 nT 

Y_bias = 483 nT 

Z_bias = 22 nT 

264






However, as the spacecraft is not near to being fully integrated and because of heavy 

electrical activity near to the spacecraft (OBC+EPS) then this test should be redone by 

MAN/SYS at a more appropriate time. 

A procedure to determine X, Y satellite self biases is planed but postponed until the spacecraft 

is properly integrated. 

ACDS_Lards implements a new telecommand to adjust the parameters of the control filter. 

The coefficients of the y[k-1] coefficient in the difference equation is now adjustable by a 

commandable number of (the parameter in telecommand as a 16 bit signed integer) 1/32678 

increments/decrements. The x[k] coefficient is calculated on-line by preserving a unity-sum 

of the two coefficients. Each adjustment will generate an alarm message with the current 

value (after update) of the y[k-1] coefficient of the filter. 

All ACDS telecommads were tested (once again). All worked as supposed to. 

ESA_Neil comes up with cunning way to split the spacecraft so that the Tommy-test-plan and 

the Karl-test-plan are orthogonal and commutative. Five laptops are used: one controlling the 

groundstation, one uploading software to the OBC, one for writing software changes for the 

OBC, one connected to the PCU and emulating the OBC, and one connected to the OBC and 

emulating the PCU. The PIN, OBC, UHF and SBAND are powered up off of one power 

source. The PCU is powered from the battery, and a second power source is used to simulate 

any specific battery voltage to the PCU. 

OBC_Karl attempts to tune the leaking bucket algorithm. 

EPS_Tommy and EPS_Fulvio attempt to write and upload some new software that listens to 

pongs and keeps the OBC alive. 

PROBLEM 178: Utility processor bottleneck cannot drive UHF past about 8kbps 

TEST: OBC_Karl speeds up the computer bus, but this results in internal data errors. 

However, with transmission speed at 11 kbps the UHF carrier stays up permanently. 

PROBLEM 179: With transmission speed at just below 9k6 there is a significant error rate, 

probably partly from the internal conflicts and partly from the UHF loosing packets. 

OBC_Karl and ESA_Neil investigate possibilities to change the various TNC settings in the 

UHF FM unit. 

MODIFICATION 87: OBC_Karl adds telecommand 32, which passes command 

parameters to the TNC in the UHF FM. 

265






ESA_Neil and OBC_Karl alter the TX_delay, the TX_tail, the Persistence and the Slot-time 

in an attempt to get the carrier up full time while transmitting. Although a setting of p=1 will 

give the desired effect, it also seems to create unpredictable delays. 

All settings are returned to their "defaults" (TX_delay=50ms, TX_tail=0, p=63, Slot-time=10) 

PROBLEM 180: We keep missing the acknowledge when we send a TC. 

SOLUTION: ESA_Neil sets the TX_delay to 150ms, which seems to help. 

ESA_Neil, EPS_Tommy and OBC_Karl go and eat dinner in ESCAPE. 

EPS_Tommy tests the no-charge-for-12-mins battery failure. 

OBC_Karl adds nominal mode beacon decoding routines into term.exe so that the EPS battery 

voltage and PCU temperatures are displayed in a human-readable format. 

OBC_Karl tries to solve the s-band leaking bucket problem by inter-packet-counting (busy 

loop) for 10 packets, then sleeps for 10ms to let the other threads catch up. 

ESA_Neil downlinks large thumbnails on the UHF. 1280 packets takes 150seconds, which is 

equivalent to only 5734 bps. More tuning is needed! 

We pass midnight into the… 


EPS_Tommy reports success on the no-charge-for-12-mins battery failure. 

Over several hours OBC_Karl tweaks the timing parameters on the S-Band data link and the 

iterations are tested with ESA_Neil operating the groundstation, with the following results. 

NOTE: A large thumbnail is 1280 packets, but either 5 or 6 beacons will also be received 

during the transmission, so a packet count of 1285 or 1286 is acceptable. 

125 seconds, 1280 packets, 7k4, 69 errors :(((((((((((((((( 

125 seconds, 1280 packets, 7k4, 11 errors :((((((((((((( 

110 seconds, 1280 packets, 8k3, 11 errors :(((((((((( 

104 seconds, 1280 packets, 8k9, 14 errors :((((((((( 

100 seconds, 1225 packets, 8k8, 3 errors :(((((((( DROPPPED PACKETS 

105seconds, 1285 packets, 8k8, 9 errors :(((((( 

ESA_Neil puts the TX_tail = 80ms 

266






105seconds, 1286 packets, 8k8, 11 errors :(((((( 

We don't believe it but: OBC_Karl leans forward, 2/5 uplink reliability. OBC_Karl leans 

back, 5/5 TEN TIME IN A ROW!!! 

PROBLEM 181: The UHF uplink reliability is extremely sensitive to other objects in the 

clean room. 

ESA_Neil tests the TX delay longer and longer until we get a reliable acknowledge again, 

finally setting it to 100ms. 

100seconds, 1286 packets, 9k3, 14 errors :(((( 

103seconds, 1285 packets, 9k0, 11 errors :(((( 

Karl makes OBC set TX_delay to 100ms 3secs after each boot up. 

103seconds, 1281 packets, 9k0, 7 errors :((( 

103seconds, 1282 packets, 9k0, 6 errors :(( 

103seconds, 1282 packets, 9k0, 9 errors :(( 

103seconds, 1282 packets, 9k0, 4 errors :( 

103seconds, 1238 packets, blah, 7 errors :((((((( too far... 

103seconds, 1285 packets, 9k0, 9 errors :) 

This corresponds to no packet loss and less than a 1% erroneous packet rate (the actual bit 

error rate would of course be much lower). 

EPS_Tommy has fallen asleep on the cleanroom floor (it is about 3am). 

267






OBC_Karl and ESA_Neil set about testing S-Band in the same manner. 

PROBLEM 182: Part-way through a large downlink from S-Band the transmission stops 

and no more data can be sent without cycling the S-Band TNC. This is probably because the 

leaking bucket algorithm is overflowing the buffer. 

We downlink a large thumbnail on S-Band in 24.1 seconds, 1283 packets, 38k3 bps with 

ZERO errors. 


ZERO errors. 


ZERO errors. 

ESA_Neil and OBC_Karl demonstrate, using a laptop to emulate the EPS, that the OBC still 

responds to pings while doing a large downlink via UHF. 

ESA_Neil and OBC_Karl attempt to demonstrate, using a laptop to emulate the EPS, that the 

OBC still responds to pings while doing a large downlink via S-BAND. 

PROBLEM 183: The OBC does not respond to the PCU pings while downlinking a picture 

on S-Band. 

TEST: OBC_Karl slows down the leaking bucket filling by 10%, then OBC pongs the pings 

with no problem. So OBC_Karl speeds it up a bit again. 

24.1s, 1275 packets, 38k3, 0 errors, pinging... (packets dropped) 

OBC_Karl and ESA_Neil test the OBC -> S-Band link by a direct connection to a laptop 

instead of the RF link. No packets are dropped, implying that they must be lost in the TNC or 

RF, not in the OBC. 

TEST: OBC_Karl tunes the timing by separating 8 packets with busy-loop counting, and 

then sleeping for 10ms before the next 8. (Before it was double both timings.) No packets 

are lost. 

24.3s, 1282 packets, 38k0, 0 errors, pinging. Perfect! 

OBC_Karl removes the ‘picture data received’ message from term.exe so that a laptop can 

cope with a full picture download. 

24.3s, 1282 packets, 38k0, 0 errors, pinging. 

268






OBC_Karl and ESA_Neil downlink a full picture on S-band. It consists of 20480 packets, 

which is 1,843,200 bytes of raw data, which is transmitted in 6 minutes and 31 seconds, with 

no errors. This corresponds to an effective speed of 37k7 bps. 

OBC_Karl realises that the debugger is still send “ping received” to S-Band. He disables that 

in the code. 

It is 5:55am. ESA_Neil wakes up EPS_Tommy and asks him to prepare for EPS testing, 

while ESA_Neil takes a break and lets OBC_Karl into the office so that he can sleep in the 

hammock. 

ESA_Neil and EPS_Tommy begin testing the latest EPS software (version 6.9). 

We power up the spacecraft, progress through the modes and bring up the OBC. It remains 

up for four pings, so the situation is declared stable. 

NOTE: The reading the PCU performs on the battery is from 0-5V. This is being simulated 

in these tests by a power supply. The number must be multiplied by 5 to obtain the actual 

battery voltage. 

The nominal mode beacon reads 24V on battery (simulated currently at 4.9) and temperature 

at 24 deg C (could be ok, room is 22). 

EPS_Tommy set 4V simulated on battery reading. Nominal mode beacon shows 19V (~20, 

so ok) and temp rises to 27 deg C. 

PENDING MODIFICATION: OBC_Karl should increase the decimal accuracy in the 

decoding of the EPS data in the nominal mode beacon in term.exe. 

ESA_Neil and EPS_Tommy test the EPS telecommands. (Switching on and off each other 

unit, and power-cycling UHF). They work fine. 

Since the software has changed it is necessary to re-inspect the beacons. 

ESA_Neil kills the watchdog via TC 00 0b -- --, after two pings (still shows as received on 

the OBC debugger) the power to the OBC is cut off. 

PROBLEM 184: If the OBC does not respond to two consecutive pings the EPS should send 

a “shutdown” command and THEN power it off. This shutdown command was not sent. 

ESA_Neil and EPS_Tommy test the beacons. 

Voltage 

Beacon 

4.008 11001000010111 

4.008 11011000010111 

269






The beacons still seem to be working fine. 

4.008 11011111100111 

4.008 11000000010111 

3.594 11001111011010 

3.594 11001111011010 

3.594 11010111011010 

We restart the spacecraft and EPS_Tommy drops the battery voltage below the “dead” 

threshold (2.62V). The PCU goes into recovery mode – ignoring the battery as it should. 

We want to show that the battery-not-charging failure only occurs if it is the first time in safemode 

(as it will be if battery is dead and spacecraft powers up as it comes out of eclipse): so 

power up with a simulated battery voltage of 3.3V, and wait for 12 mins 

Meanwhile… 

Voltage 

Beacon 

3.300 11010000101010 

3.300 11010000101010 

3.300 11001111001010 

3.300 11010000101010 

After approximately 14 minutes of no charging the PCU enters recovery mode, implying that 

it assumes the battery is dead. This is perfect. (A little long, but we need to only do this if we 

are SURE that the batteries are dead.) 

Now we reboot the spacecraft and bring it up into recovery mode. WE then drop back down 

to safe mode and lower the battery voltage below the “dead” threshold. We wait for 15 

minutes and there is no change. This is as it should be (since it is not the first time in safe 

mode since it is booted up, therefore the satellite has not been powered off by an eclipse). 

EPS_Tommy and ESA_Neil simulate a solar panel using the Keithley current source running 

at 550mA and taking 28.7 volts. 

Battery initially at 4.20V, nothing on shunt. 

We raise solar panel to 580mA (28.7V), shunt starts to output. 

We raise solar panel to 850mA (28.9V), shunt at 6.02V. 

After 4mins battery is reading 4.25V -> charging. 

Charging threshold on panel is 21.5V, which is equal to 5 times the battery voltage. This is as 

it should be. 

270






We raise solar panel to 800mA (at 28.9V), the battery reads 4.29V. 

3 mins later, 4.32V on the battery, 800mA on panels (28.9V) -> batteries are charging. 

ESA_Neil demonstrates that all modes of the S-Band unit are working when powered off of 

the PCU. 

NOTE: The next tests are simulating a failed battery and a slow entry into eclipse. We do 

not expect the system to do particularly sensible things to start with. 

We power down, disconnect battery (activation switch) and simulate 0V on battery. 

ESA_Neil turns on panels with low current and slowly raises it: 

The voltage initially oscillates as PDU tries to use it but it is not enough. Then the voltage is 

constant at 9.7V once the panel is up to 290mA. However, the T-pods never fired. 

The system is only JUST stable at 430mA and 28.5V (any lower and the voltage drops). It 

gets more stable towards 490mA. 

At 500mA 28.7V we eventually get what we have been waiting for: a recovery mode beacon 

with all zeros (11010000000001) 

At 550mA the OBC finally comes up, UHF receiving (can see on the debugger), but the 

spacecraft cannot transmit anything because of the current peaks. We also see a lot of fetch 

queue timeouts with the UHF. 

The UHF can only transmit when the panel is above...710mA. This implies that we need 

three strings in order to transmit with no battery. 

We drop below 660 and the OBC powered off (a spike must have killed it) 

Much to EPS_Tommy’s relief we power off. 

EPS_Tommy and ESA_Neil power the spacecraft up using the battery until we get to nominal 

mode. Then we drop the simulated battery power until we exit nominal mode and go to safe 

mode because of the low power. This is as it should be. 

PROBLEM 185: Suddenly the spacecraft powers off for no apparent reason! Battery low? 

Battery protected?! Then the battery is reading out at 18V, and seems to charge quite rapidly 

to 20V or so but no more. 

SOLUTION: The batteries are just empty (they have been used for several days now). They 

seem to work fine again after some charging. 

271






TEST: Is the occasional “fetch queue timeout” EPS’ fault or a cable fault? Last packet was 

wrong – was it breaking connection or the software? Need a re-test at some point. 

EPS_Tommy leaves the batteries charging from the Keithley and goes home to Italy. 

ESA_Neil and OBC_Karl power up the spacecraft from the battery and a panel. 

PROBLEM 186: The nominal mode beacon data was not being sent correctly from EPS. 

SOLUTION: Karl was snooping on the line with a laptop, the resulting lower current is 

affecting communications reliability. 

ACDS_Lars re-ports the ACDS code back to the OBC hardware. Times the execution of the 

ACDS thread to verify that it was executed at the proper frequency. Main loop was found to 

be executed at 0.6Hz as expected. 

ACDS_Lars tests dataflow from magnetometer through OBC to ACDS software. Identified 

and corrected a "fence-post-error" in the interfacing code. Verified proper treatment of 

floating point numbers by the run-time environment on the OBC. Verified (again) the correct 

alignment between magnetometer axes and spacecraft. 

ACDS_Lars connects the coils to the coil driver and made temporary changes in the ACDS 

code to allow use of the on-board magnetometer to measure the activity in the coils (not the 

default behaviour). 

Using the above changes to the code; ACDS_Lars verified that the signs in the control 

algorithm are correct - They are (btw. Coil1 is a synonym for X and coil2 for Y). He also 

implements a work-around ("ignore "data if all zeros") to compensate for OBC bus problems, 

specifically in times with a lot of bus activity (TM broadcast). Then he changes the code back 

to its default behaviour and tests that measurements are now insensitive to coil action. 

Using a big bar-magnet ACDS_Lars verifies that the on-board passive magnet is mounted 

correctly, i.e magnetic-north pole aligned z+. 

All ACDS harness is thoroughly marked with descriptive stickers (replacing the old ones 

which had a tendency to fall off). This should help ESA_Neil finish the flight harness for 

ACDS when the spacecraft is integrated to a level where it is possible to (finally) asses the 

length of each piece of harness. 

The new L profile was glued to the x+ panel above the ACDS coil driver. 

272






ESA_Neil and OBC_Karl commence testing of the OBC software by the list of 

telecommands: 

Telecommand Parameters Result 

Get HK One packet, radio 4 Acknowledge and 1 down UHF 

Get HK Two packets, radio 2 Acknowledge and 2 down UHF 

Get HK All packets, radio 2 Acknowledge and all down UHF 

Get HK All packers, radio ff Acknowledge and all down UHF 

Get HK 256 packets, radio 0 Acknowledge and 256 down UHF 

Get HK All packets, radio 4 Acknowledge 23191 bytes in 20s = 9.276 kbps 

Get HK – Verify One packet, radio 4 Acknowledge and 1 down UHF 

Get AL All packets, radio 4 Acknowledge + measurement 95 and CID 81 

Get HK Nine packets, radio 4 Acknowledge and 9 down UHF 

Get HK – Delay Nine packets, radio 4 Acknowledge, delay, 9 down UHF 

Get HK 256 packets, radio S Acknowledge U, 256 down S-Band @ 20 kbps 

Get AL All packets, radio 0 Acknowledge and all down UHF 

Get AL All packets, radio S Acknowledge and all down S-Band 

Get AL – Verify All packers, radio 1 Acknowledge and all, including itself 

Get PIC First chunk, radio 0 Acknowledge and 100 packets @ 9.322 kbps 

Get PIC 9 th chunk, radio ft Acknowledge and 100 packets 

Get PIC All chunks, radio S Acknowledge and 20120 packets in 394s 

OBC_Karl tunes the s/w as some packets were dropped. 

OBC_Karl, ACDS_Lars and ESA_Neil go and eat. 

Flush HK No packets, radio 0 Acknowledge and stack report 

Flush HK No packets, radio 4 Acknowledge and stack report 

Flush HK 256 packets, radio 0 Acknowledge and correct flush 

Flush HK 256 packets, radio S Acknowledge and correct flush 

273






Flush AL No packets, radio 0 Acknowledge and stack report 

Flush AL 256 packets, radio 0 Acknowledge and correct flush 

Flush AL – Ver. 256 packets, radio 0 Acknowledge and correct flush (inc itself) 

Flush AL 256 packets, radio S Acknowledge and correct flush 

FP Flush No items, radio 0 Acknowledge and stack report 

FP Flush One item, radio 0 Acknowledge and correct flush 

FP Flush – delay One item, radio 0 Acknowledge and correct flush (inc. itself) 

WD Toggle Nada OBC gets shutdown and cycled by EPS 

Get THUMB1 All chunks, radio 4 Acknowledge and all down at 9k0 bps 

Get THUMB1 All chunks, radio S Acknowledge and all but three down at 36k4 bps 

Shutdown Nada Does exactly what it says on the tin 

Live stream Radio 4 Does exactly what it says on the tin 

Live stream Radio S Does exactly what it says on the tin 

Friendly Toggle True Enables friendly telecommands 

Get HK – Friend 10 packets, radio 4 F-acknowledge and 10 packets down UHF 

Get HK – Friend 10 packets, radio S F-acknowledge and 10 packets down S-Band 

Get AL – Friend 10 packets, radio 4 F-acknowledge and 10 packets down UHF 

Get PIC – Friend Chunk 0, radio 4 F-acknowledge and 100 packets down UHF 

Friendly Toggle False Disables friendly telecommands 

Check flash True Acknowledge and reports flash is ok 

Get up-time Nada Acknowledge and reports 33.56 minutes 

MODIFICATION 88: OBC_Karl removes the flash erasure commands. (They are replaced 

now by the filing system.) 

We pass midnight (again) and enter into… 

274







Open file File 27 Exactly what it says on the tin 

Open file File 26 Exactly what it says on the tin (and closes 27) 

Close file File 26 Exactly what it says on the tin 

Open file File 27 Exactly what it says on the tin 

Get file 28 packets, radio 4 Acknowledge and 28 packets down 

File seek 28 packets Moves marker by 28 packets 

Delete file File 27 Exactly what it says on the tin 

Protect file File 25 Exactly what it says on the tin 

Unprotect file File 25 Exactly what it says on the tin 

Generate CRC Nada Exactly what it says on the tin 

ACK Target Radio S Acknowledges on S-Band from here on 

INIT modem TX_delay, 1000 Sets the TX-Delay to 1 second 

MODIFICATION 89: OBC_Karl takes TX_delay set out of the OBC boot-up sequence. 

MODIFICATION 90: OBC_Karl makes “whole picture” commands unfriendly. 

OBC_Karl tweaks the radio delays to fix the timing problems, and changes the password of 

the spacecraft. 

In dialogue with ESA_Neil; ACDS_Lars decides to have ACDS detumbling *OFF* by 

default, meaning that detumbling must be a "ground assisted operation". Reason is to KISS, 

alternatively more autonomy should have been included on the OBC to do an "out of the box 

test" of the ACDS autonomy. 

ESA_Neil declares the OBC software FLIGHT READY. He then goes home to sleep, as he 

has been working almost non-stop for a 40 hour shift. 

OBC_Karl burns flight proms, opens the OBC and inserts them. 

275






PROBLEM 187: The OBC flight proms do not work as they should – there are some serious 

timing issues due to an unexpected difference between the flight and non-flight chips. 

ESA_Neil declares the OBC software NOT FLIGHT READY. 

The majority of Aalborg university sets about troubleshooting the flight prom issue. 

[5:30AM] ACDS_Lars plugs in ACDS driver plus sun-sensors. Objective: Map telemetry to 

physical directions on the spacecraft. Results follows: 

Sensor on the "Y-"-panel: 

-------------------------- 

sun_11: (~1525 dark). Horizontal. + is LEFT 

sun_12: (~1525 dark). Vertical. + is UP (down has no sensitivity :( ) 

sun_ref1: Reference area. 800-900 with lamp illuminating, 0 dark 

temp_sun1: NOT EXISTING in HW! 

Sensor on the "Y+"-panel: 

-------------------------- 

sun_21: (~1515 dark). Horizontal. + is LEFT 

sun_22: (~1515 dark). Vertical. + is UP 

sun_ref2: Reference area. 800-900 with lamp illuminating, 0 dark 

temp_sun2: NON-COMMITTED ADC INPUT (floating) 

[6:40AM] Test concluded! :) 

OBC_Karl replaces the non-flight proms. 

ACDS_Lars and OBC_Karl leave for Denmark to complete the troubleshooting. 

ESA_Neil meets with the testing division and schedules the vibration tests for the 11 th April, 

and the EMC tests for the 6 th of April. 

ALCATEL_Samson cannot perform his visit, so ESA_Neil and ESA_Philippe arrange for an 

emergency backup from Dutchspace. 

DS_Roland visits the cleanroom and appraises the solar cell situation. Then he returns to 

Dutchspace and comes back with DS_Fons. 

DS_Fons gives ESA_Jason, ESA_Bas, ESA_Neil, SOL_Nico and SOL_Yoann a solar cell 

soldering course, using two of the class 7 cells and the SSTL rear-solder jig. 

276






PROBLEM 188: The SSTL rear-solder jig is the wrong size and shape for the Tecstar cells. 

ARRIVAL 64: DS_Roland and DS_Fons generously donate some RTV and special solder to 

SSETI for the purposes of panel laydown. 

ESA_Neil designs the SSETI Express Solar Cell Rear Solder Jig and requests manufacture. 

STRU_Melro designs solar panel substrates. 


ARRIVAL 65: DS_Roland and DS_Fons generously donate 64 Tecstar solar cell interconnectors. 

ARRIVAL 66: The SSETI Express Solar Cell Rear Solder Jig arrives. A simple fit check 

proves that it will be perfect for the job. 

277






ESA_Neil gives several tours of the SSETI Express cleanroom to participants of the ESEO 

workshop. 

ARRIVAL 67: The first of the solar panel substrates arrives. 

ARRIVAL 68: The last batch of solar cells arrive. 


ESA_Neil and ACDS_Lars discuss the ACDS LEOPs autonomy. 

MODIFICATION 91: It is decided that for successive periods of 24 hours after the launch 

without reception of a valid TC the following configurations of ACDS will be used 

respectively: normal, off, off, off, off, off, off, inverted, off forever. 

In this manner if the algorithms are wrong, the satellite spins UP instead of down, and 

commanding the spacecraft is impossible, then one week later the algorithms will be inverted 

and re-implemented to detumble properly. The week-long delay is to make sure that we have 

had time to distinguish the spacecraft from the other 7 passengers of the launch, and to 

establish two-way communication. 

MODIFICATION 92: OBC_Karl implements a byte stored in flash memory that controls 

what configuration will be used for ACDS after boot-up. A new telecommand to reset the 

byte is added, and the ACDS reading of the byte is implemented. 


ESA_Neil removes the side protectors from the spacecraft. 

278






MODIFICATION 93: The teflon stand-offs for the side protectors are shortened by 3mm to 

make space for M4 nuts that will hold the bolts in place. 

MODIFICATION 94: The EM dummy antenna is dismantled, the edges of the base ground 

inwards by 0.5mm, the holes drilled out to 4.2mm and reassembled. 

ESA_Neil prepares the bolts for MAGIC. 

ESA_Bas solders string 0a. 

DS_Roland and DS_Wouter give a demonstration to the SOL team (and ESA_Neil) on string 

laydown with string 0a and substrate panel 7. 

ARRIVAL 69: Dutchspace generously donate three pots of RTV, two vials of hardener and 

a bottle of primer. 

The SOL team write up and develop the laydown procedure while string 0a is left to dry. 


SSTL_Andy informs that the fit-check will take place on the 18 th April and that the hardware 

has to be shipped THIS WEEK. 

ESA_Neil scrubs the rust off of the EM lifting frame. 

279






MODIFICATION 95: ESA_Neil shortens the side protector spacers by 3mm and adds an 

M4 nut to each of the bolts to hold the spacers on. 

ESA_Neil cuts bolts for Magic to a length of 88mm from M4 thread. 

ESA_Neil attaches the EM antenna to the EM structure, taking care to make sure the angle is 

similar to the FM. 

280






ESA_Neil and ESA_Iñaki drive to Lisse to pick up the first flight substrate panel (1a). 

ESA_Neil purchases mixing bowls and sticks for the RTV preparation. 

ESA_Neil gives several tours of the cleanroom to VIPs. 

The SOL team develop techniques to lay the kapton flat on the panels without any air bubbles 

underneath. 

ESA_Neil starts to collate the tools required for fit-check. 

ESA_Bas solders strings 1a, 1b, 6a, 6b, 7a and 7b. 

The SOL team mix sample 1 of the RTV and use it to lay down string 1a on panel 1. It goes 

very well. 

281






ESA_Neil commences software testing with OBC and EPS. 

PROBLEM 189: There is no nominal mode beacon data being received by the OBC from 

EPS. Instead there is an occasional MID 04 in the TM, and the occasional fetch queue 

timeout on SID 2. 

282






MODIFICATION 96: EPS_Fulvio adds delays between packets as well as between bytes, 

as some packets are being sent very close to each other. Version 7.1. 


EPS. Instead there is an occasional MID 10 in the TM, and the occasional fetch queue 

timeout on SID 2. 

MODIFICATION 97: EPS_Fulvio extends the inter-packet and inter-byte delay to 1ms 

instead of 1 us. Version 7.2. 

MODIFICATION 98: EPS_Fulvio changes the period on the nominal mode beacon data to 

59 seconds instead of 60 so that it doesn’t collide with the telemetry packet so often. This is 

version 7.3. 

ESA_Neil verifies the presence of nominal mode beacon data from EPS by emulating the 

OBC with a laptop and hex terminal. 


EPS, even though it is being sent. (Although there are no timeouts now.) 

ESA_Neil emulates the PCU with a laptop and a hex terminal and sends nominal mode 

beacon data to the OBC manually. It is not received unless the initial “99” is replaces by 

another number. He informs Karl of this. 

The EPS batteries run out and the spacecraft powers down. 

ESA_Neil turns the solar panel simulator on to charge the EPS battery, and powers the 

spacecraft back up again via the Alternate Power System. 

MOIDIFICATION 99: OBC_Karl corrects the OBC software to rectify the “switch case fall 

through error” that was causing the EPS nominal mode beacon data to register as OBC data. 

ESA_Neil tests the latest OBC software with a large thumbnail downlink on UHF. The 

results are as follows: 

- 1286 packets on UHF in 106 seconds, 2 errors 

- 1281, 1 error (four packets dropped) 

283






- 1284, 1 error (two dropped) 

- 1285, 2 errors (one dropped) 

- 1286, 3 errors (none dropped) 

MODIFICATION 100: OBC_Karl tweaks the UHF downlink dripping bucket. 

ESA_Neil reboots the OBC and uploads the new software. 

ESA_Neil manually sends nominal mode beacon data from an emulated EPS and it is entered 

into the telemetry stack as it should be. 

ESA_Neil tests the latest OBC software with a large thumbnail downlink on UHF. The 

results are as follows: 

- 1286 packets, 101 seconds, no errors 

- 1286 packets, 101 seconds, 1 error 

- 1286 packets, 101 seconds, 1 error 

- 1282 packets, 101 seconds, 1 error, FOUR packets dropped, but there was a fetch 

queue timeout on the UHF 

UHF downlink speed equates to 9k2 bps with a very small error rate. The OBC was also 

responding well to emulated pings during the downlink. This is fine. 

ESA_Neil snoops on the UHF to OBC RS232 line to troubleshoot the fetch queue timeout: 

000001 22:02:29.235 7E C0 00 AA 92 40 40 40 40 60 A6 A6 8A A8 92 64 

000002 22:02:29.295 61 03 F0 00 12 FF 04 01 00 00 00 00 4B B4 4B 80 

000003 22:02:29.295 C0 

This is fine and the fetch queue timeout cannot be repeated. This would cause the OBC – 

UHF link to go down for four seconds though, which would result in packet loss. 

ESA_Neil compares the snoop on the OBC – UHF line to the debugger. It can clearly be seen 

that all uplink attempts are leaving the UHF TNC correctly, or not at all. This demonstrates 

that it is not the OBC that is at fault. It is also important to know that gibberish is not being 

sent by the UHF on a failed attempt. 

ESA_Neil tests telemetry downlink on UHF. By comparing the number of packets 

downlinked to the number of packets shown on the debugger it is clear that no packets are 

lost. 3968 bytes are received in about 4 seconds, which is equivalent to 7k9 bps. 

ESA_Neil tests S-Band downlink of large thumbnails with the following results: 

- 1273 packets (10 dropped), 25 seconds, no errors 

- 1254 packets (29 dropped), 25 seconds, no errors 

284






MODIFICATION 101: OBC_Karl tweaks the S-Band downlink dripping bucket. 

- 1283 packets (none dropped), 24.77 seconds, no errors 




ESA_Neil performs a full picture downlink on S-Band (note, there should be about 33 

beacons during one picture downlink, so a total of 20513 packets is expected): 

- 20513 packets (none dropped), 1855941 bytes, 394 seconds 

This is equivalent to 37k7 bps. 

ESA_Neil tests TM downlink via S-Band: 

- 192 packets, 3.03 seconds, 8055 bytes, 21k48 

ESA_Neil turns on the CAM and sends it a ping to check it is working. “CAM” appears in 

the telemetry stack as it should. 

ESA_Neil transmits a large thumbnail from the camera to the OBC. The OBC keeps 

responding well to emulated pings throughout. 

ESA_Neil transmits an entire picture from the camera to the OBC. The OBC does keep 

responding to emulated pings throughout (including the flash memory manipulation), but does 

so very slowly. The longest it ever takes though is about 20 seconds, which, although slow, is 

not a problem as the timeout is 60 seconds. 

ESA_Neil shuts down the spacecraft and powers it back up again via the EPS. 

285






The nominal mode beacon data from EPS is now being received by the OBC without a 

problem. There are no fetch queue timeouts on SID 2. 

ESA_Neil kills the EPS-OBC watchdog and listens to the recovery mode beacon. It is fine. 

ESA_Neil powers down the spacecraft (00:30am). 

MILESTONE 26: The OBC and EPS software are declared FLIGHT READY 

ESA_Neil asks EPS_Fulvio and OBC_Karl to burn flight proms and send them to ESTEC as 

soon as possible. 


ESA_Jason solders wires onto the 0a string. 

ESA_Neil tests the 0a panel in the vacuum chamber, it suffers no damage. 

ESA_Jason and ESA_Neil find diodes: IN 5819 schottky 40V dc blocking, 1A. 

ESA_Neil powers up spacecraft and connects the groundstation to the LAN so that it can be 

remotely operated from Aalborg. 

PROBLEM 192: MCC are not ready and need some software modifications. 

ESA_Neil prepares the EM for shipment to Omsk for the fit check and collates the items and 

tools that will be needed. 

After checking with SSTL_Andy it is agreed that the side protectors can be carried separately 

by the fit-check personnel, therefore allowing us to keep them with the flight model in 

ESTEC in the meantime. 

ESA_Jason solders wires onto the 1a solar panel. 

ESA_Neil and ESA_Marie pack the fit-check shipment. Pictures are taken primarily for the 

Russian import papers. 

00 - Transport container 

01 - Shape dummy 

286






02 - Packing foam 

03 - Toolbox 

04 - Lifting frame 

05 - Bolts and spacers 

287






06 - Dummy antenna 

07 - Antenna bolts 

08 - Straight torque wrench 

09- Angled torque wrench 

288






11 - Wrench bits 

12, 13, 14 - RBF plugs 

15 - Spanner 

16 - Electrolube wipes 

17 - Prosat wipes 

18 - Vinyl anti-static gloves 

289






19 - Stahlwille 400 

20 - PTFE wire 

ARRIVAL 70: Three solar panel substrates arrive, 1 FM, 1 EM and 7 FM. 

ESA_Neil does a fit-check with the new substrate panels. 

PROBLEM 193: Solar panel substrate 1 FM and 1 EM interfere with the sub-sensor, and do 

not reflect the ACDS coil driver location modifications. 

290






MODIFICATION 102: ESA_Neil and ESA_Andre remove a cut-out from the 1 FM 

substrate panel and drill two new holes for the coil driver bolts (7mm down and 1mm left). 

LESSON LEARNED 20: Any modifications to the hardware (or software) should be “back 

implemented” into the design immediately. Otherwise future actions derived from the design 

may not fit with the current reality. 

NOTE: It is necessary to disintegrate the FM ACDS coil driver from the –y lateral panel 

temporarily in order to perform a satisfactory fit-check of the solar panel. The unit is replaced 

immediately afterwards. 

PROBLEM 194: Solar panel substrate 7 FM has been dropped and has some damage at one 

corner. 

MODIFICATION 103: ESA_Neil files off the erroneous damaged corner of the solar panel 

substrate 7 FM 

291






PROBLEM 195: There is no attenuation on S-Band. This could be causing some TX 

reflections and radiation problems for personnel. 

SOL team start laying down strings 1a, 1b, 2a and 2b. 

ESA_Neil and the SOL team enjoy dinner and beer in ESCAPE. 

The SOL team lay down strings 1a, 1b, 2a and 2b. 


ESA_Neil puts the attenuator back on the s-band unit. 

The SOL team set about cleaning the panels they have made and removing kapton tape from 

the bolt holes. 

292






ARRIVAL 71: The rest of the solar panel substrates arrive. 

ESA_Bas solders strings 4, 8, 9a and 9b. 

ESA_Jason solders wires onto solar panel 2. 

ESA_Neil and SOL_Nicolas perform fit checks of the solar panel substrates onto the FM 


ESA_Neil disintegrates both sun sensors, and the coil driver, from the lateral panels. Care is 

taken to label the sun sensors so that they go back on the right way around. They are stored in 

the low-humidity cupboard. 

PROBLEM 196: During disintegration the head of one of the cross-head bolts used to 

mount the sun sensors is partially stripped. 

NOTE: In order to perform the fit checks several of the magnetorquer coils bracket bolts had 

to be removed. These had previously been torqued and glued, so some glue remains on the 

items and care should be taken when using them (get the nuts the right way around and they 

still work). 

PROBLEM 197: Solar panel substrate number 3 does not have the holes for the –y sun 

sensor mounting in the right place, and does not have the harness hole for the sun sensor 

harness at all. 

PROBLEM 198: Solar cell string 3a interferes with the sub sensor since the original location 

has been modified by ACDS but not reported back to STRU. 

MODIFICATION 104: ESA_Neil relocates the –y sun sensor holes in the lateral panel, 

using solar panel substrate 3 as a guide. 

293






MODIFICATION 105: ESA_Neil adds the –y sun sensor harness hole to solar panel 

substrate 3 and to the –y lateral panel, using the position of original harness hole relative to 

the original fixation holes in the lateral panel as a guide. 

MODIFICATION 106: ESA_Neil drills 8 holes into the –y lateral panel for the new 

interfaces with the first and third solar panel substrates. 

This completes a positive fit-check of substrates 1, 2 and 3 on the –y lateral panel. 

MODIFICATION 107: ESA_Neil drills 2 holes into the +x lateral panel for the new 

interfaces with the fifth solar panel substrate. 

294






This completes a positive fit-check of substrates 4 and 5 on the +x lateral panel. 

MODIFICATION 108: ESA_André enlarges the cut-out in the top of the +y panel for the 

heating block and harness of the +y T-Pod. 

MODIFICATION 109: ESA_Neil drills 10 holes into the +y lateral panel for the new 

interfaces with the sixth and eighth solar panel substrate. 

295






PENDING MODIFCATION: There is probably no harness hole for the +y sun sensor in 

the sixth solar panel substrate. 

Other than the above, this completes a positive fit-check of substrates 6, 7 and 8 on the +y 


MODIFICATION 110: ESA_Neil drills 2 holes into the -x lateral panel for the new 

interfaces with the ninth solar panel substrate. 

This completes a positive fit-check of substrates 9 and 10 on the -x lateral panel. 

296






PROBLEM 199: When STRU_Melro designed the solar panel substrates we did not know 

that there should be 1mm gaps between the cells, or that there were 5mm connectors on each 

end. Subsequently several of the strings are actually slightly too long for the panels. 

MODIFICATION 111: It is decided that problem 199 can be alleviated by adding the final 

connectors to the side of the last cell instead of the end. Strings 4 and 8 are adapted to reflect 

this, and strings 5a, 6a and 9a are planned to be implemented with this change already in 

place. Similar modifications possibly remain to strings 3b, 3c and 6b. 

ESA_Neil powers up UHF and S-Band and brings the carrier of the S-Band unit up for 10 

minutes. The attenuator is then found to be warm – implying that no damage was done to the 

unit by leaving off the attenuator. 

ESA_Neil performs a vacuum test on panel 1a. It suffers no discernable damage and still 

reports a voltage of 9.3 under the cleanroom lights (as it did before). 

The SOL team proceed to lay down the remaining strings. 

ESA_Bas solders string 10. 

297







MODIFICATION 112: ESA_Neil adds a harness hole for the +y sun sensor to solar panel 

substrate 6 and the +y lateral panel. It is positioned relative to the fixation points in the same 

pattern as for the –y sun sensor. 

MODIFICATION 113: ESA_Bas solders an extra cell onto 3b to make it into a two cell 

string – hereafter dubbed 3d. 

SOL_Vincent solders wires onto solar panel 7. 

MODIFICATION 114: ESA_Bas replaces 6b with a cell that has connectors on the side. 

MODIFICATION 115: For ease of configuration single cell strings 3b and 3c are to be 

relocated to the second solar panel substrate as a two cell string, 3d. 

298






The SOL team lay down the remaining strings. 

DEPARTURE 11: The fit-check hardware leaves for Omsk. 

SOL_Yoann and SOL_Vincent leave for Toulouse. SOL_Nicolas stays for the soldering and 

wire routing. 

299






The bolts specifications for the MAGIC box cannot be fulfilled, as no such standard bolt 

exists. Therefore a thread was cut to length with the intention of having a nut on each end. 

These threads and nuts are cleaned in the ultrasound bath by ESA_Neil. 

PROBLEM 200: The holes in the base of the MAGIC box are not large enough to 

accommodate an M4 nut, only a bolt head. 

MODIFICATION 116: ESA_Neil uses a milling bit on the Dremmel to enlarge the holes in 

the baseplate of MAGIC such that M4 nuts will fit. 

ESA_Neil torques / tightens and glues all bolts and screw locks in the MAGIC box. 

PROBLEM 201: The +z-y bolt hole in the top-plate of MAGIC is slight misaligned (or the 

bolt is slightly bent). 

MODIFICATION 117: ESA_Neil uses the Dremmel to enlarge the hole. 

ESA_Neil closes the MAGIC box and tightens and glues the nuts in place. 

300






NOTE: During these manipulations all connectors on the top plate were necessarily cycled. 

MAGIC weighs in at 1647g, including two 9-pin savers and one 15-pin saver. 

ESA_Neil attempts to integrate MAGIC to the FM structure. 

PROBLEM 202: The bolts are too tight to integrate the MAGIC box into the FM structure. 

The combination of several tolerances must have led to a misalignment. (Probably the insert 

potting.) 

MODIFICATION 118: ESA_Neil uses the Dremmel to slightly enlarge the four corner 

mounting holes of the MAGIC box. 

ESA_Neil integrates MAGIC to the FM structure, and tapes all savers in place appropriately. 

A fit check with the CAM is performed. All bolts go in easily. 

ESA_Neil performs a fit check with the OBC baseplate. All bolts go in easily. 

301






Aluminium tape is carefully applied around the edges of the harness holes to protect the 

harness from the sharp skins of the honeycomb, which could easy cut wires during vibration. 

SOL_Nicolas completes the soldering of the solar panels and proceeds to glue the wires in 

place using RTV. 

302






ESA_Neil assembles the flight activation switch plate and battery charge stud. 

PROBLEM 203: The nut on the batter charge stud is too tight and cannot be applied. 

SOLUTION: It is replaced by another M5 nut. However, this one is stainless steel instead 

of gold plated like the original one. 

The activation switch plate is put in place but not yet bolted, as it will need to be removed for 

soldering to the switches and the charge stud. 

PROBLEM 204: A bolt from one of the mid height brackets is protruding through the 

honeycomb wall into the back of UHF (or it would be if it was there). 

MODIFICATION 119: ESA_Neil grinds the end of the bolt flush with the shear panel, 

using the Dremmel. 

ESA_Neil disintegrates the magnetometer box and the battery box. 

303






ESA_Neil applies conformal coating to the rear side of both sun sensors, to the top of the 

magnetometer board and to the front of the battery charge regulator. 

NOTE: The modifications on the sun sensors (yellow wires) were not secure at all. 

SOL_Nicolas completes the gluing down of all the wires on the solar panels. 

304






SOL_Nicolas and ESA_Neil leave for their respective homes. It is 6am. 


ESA_Neil performs the following tasks: 

PROBLEM 205: A fit check with the –y lateral panel reveals that one of the bolts is a little 

too tight for comfort, it is marked for future modification. 

PROBLEM 206: It is not possible to integrate the +x lateral panel after the –y lateral panel 

because the +x magnetorquer coil needs to slide behind the +x-y corner profile (attached to 

the –y panel), and the +x panel cannot move like that because the +x T-Pod is in the way. 

The –y lateral panel is removed. 

PROBLEM 207: A fit check with the +x panel reveals that an extra bolt is too tight, in 

addition to the three already noted above (problem 172), they are marked for future 

modification. 

305






The –y panel is placed again to test the interface with the +x coil – it is tricky but possible. It 

would be a lot easier with two people. 

A fit check with the +y lateral panel reveals no problems. 

PROBLEM 208: A fit check with the –x lateral panel reveals one bolt that is too tight (the 

one in the –y-z corner), it is marked for future modification. 

306






An attempt is made to integrate the battery box. 

PROBLEM 209: The battery, once covered in kapton, is too big for the battery box. It is not 

possible to close it. 

MODIFICATION 120: The various layers of kapton tape are removed from the interior 

walls of the battery box and replaced with a single layer of wider kapton tape. The battery 

box is then just possible to close, although two or three bolts still will not go in. 

MODIFICATION 121: The appropriate bolt holes in the lateral panels are milled out 

slightly in the appropriate directions to fix problems 205, 207 and 208. 

307






The solar panels and the lateral panels are integrated to the flight structure as a fit check. For 

the bolt-through holes that interface the solar panel to the lateral panels M4x8 bolts and M4 

nuts are used. No washers are used (must check this with STRU). The following sequence is 

identified as an optimum one and should be followed in future: 

1) Integrate the +x patch antenna to the +x lateral panel and the +x+y corner 


PROBLEM 210: Cannot find the fixation bolts for the patch antennas. Using standard bolts 

for now as placeholders. 

2) Integrate solar panels 4 and 5 to the +x lateral panel and the +x+y corner 

profile. Use only those bolt holes that do not also interface with the primary 

structure. Do not tighten the bolts fully. 

3) Integrate the +x lateral panel to the primary structure. 

4) Tighten all the bolts on the +x lateral panel, apart from those on the +x+y 


5) Integrate the –x patch antenna to the –x lateral panel and the –x+y corner 


6) Integrate solar panel 9 to the –x lateral panel. Use only those bolt holes that do 


7) Integrate the –x lateral panel to the primary structure. Do not tighten the bolts 

fully. 

8) Integrate solar panel 10 to the –x lateral panel and primary structure. 

9) Tighten all of the bolts on the –x lateral panel, apart from those on the –x+y 


Solar panel 6 is cleaned using the standard procedure. 



11) Integrate the +y sun sensor to the +y lateral panel and solar panel 6. 

i. Pass the trailing sun sensor harness through the 4mm hole adjacent to 

the fixation holes. 

ii. Remove the protective cover from the sun sensor and take care at all 

times not to touch the delicate soldering near the window on the front. 

iii. Tighten the two M2 securing bolts until the head of each is touching 

the PCB 

iv. Tighten the M2 nuts on the reverse side until they are touching the 

PCB. 

v. Place two M2.5 washers on top of each of the fixation holes on solar 

panel 6. 

vi. Locate the sun sensor so that the bolts line up with the holes and the 

harness passes between the bolts underneath and into the harness hole. 

308






vii. Gently push the bolts through the washers, solar panel 6, and the +y 


viii. Holding the sun sensor in place attach M2 nuts to the bolts on the rear 

side of the +y lateral panel. 

ix. Tighten the nuts with a spanner. 

x. Replace the sun sensor protective cover. 



13) Integrate the +y lateral panel to the primary structure and to both the +y corner 

profiles. Do not tighten the bolts fully. 

14) Integrate solar panel 7 to the +y lateral panel and the primary structure. 

15) Tighten all the bolts on the +y lateral panel. 

16) Tighten all the bolts on the +x and –x sides of the +y corner profiles. 

17) Locate the coil driver correctly on the inside of the –y lateral panel, place M3 

bolts through the fixation holes and tape the heads in place so that they will not 

fall when the panel is turned. 

The 25-pin connector from the coil driver is integrated to the l-profile on the inside of the –y 

lateral panel. This necessarily involved cycling the saver. 

309






18) Turn the –y lateral panel over and carefully place solar panel 2 over the coil 

driver fixation bolts. Add nuts to these bolts to hold them in place. The 

kapton on the reverse side can then be removed. 

19) Integrate solar panel 2 to the –y lateral panel. Use only those bolt holes that do 


Solar panel 3 is cleaned using the standard procedure. 

20) Integrate solar panel 3 to the –y lateral panel and the –y+x corner profile. Use 



21) Integrate the –y sun sensor to the –y lateral panel. For this, repeat the 

instructions given in step 11 above (replacing +y with –y and solar panel 6 

with solar panel 3 throughout). 

22) Integrate solar panel 1 to the –y lateral panel and the –x-x corner profile. Use 



310






PROBLEM 211: The RTV is not properly securing the wires on the last point before they 

leave solar panel 1. 

23) Integrate the –y lateral panel to the primary structure, taking care to slide the 

+x-y corner profile under the +x magnetorquer coil. 

24) Tighten all the bolts on the –y lateral panel. 

25) Tighten all the bolts on the +x and –x sides of the –y corner profiles. 

Conformal coating is applied to the reverse side of the magnetometer. 

311






The side protectors are integrated to the spacecraft. 


In order to remove an antenna bolt for measuring and to allow access for re-integration of the 

magnetometer the –x panel must be taken off. 

The –x and +y side protectors are removed, the +y bolts on the –x+y corner profile are 

removed, and then solar panel 10 is removed. 

Lateral panel –x is removed and laid down using the side-protector as a stand. This works 

rather well. 

The +y side protector is replaced. The required antenna bolt is removed and measured so that 

details can be given to COMM_PL to manufacture new bolts. 

312






ESA_Neil closes the magnetometer, noting the cadmium bolts and screw locks and checking 

them with ESA_Jason and ACDS_Lars. They should be fine as long as they are well 

shielded. 

PROBLEM 212: STRU_Antonio notices that there are only four bolts holding MAGIC 

together, we need two more. 

ESA_Neil, MCC and CAM_Morten attempt to perform functional checkout of the camera via 

a remote link with Aalborg. 

313






PROBLEM 213: We are still getting nominal mode beacons, but no TC are making it to the 

OBC. We have no idea why. 

PROBLEM 214: The OBC has been shutdown by EPS and we are in recovery mode. This 

implies that two pings were missed. Why? 

PROBLEM 215: After problem 214 ESA_Neil brings the OBC back up – but no pings are 

being received from EPS at all - why?! 

PROBLEM 216: The uplink is so unreliable that it is unusable. 

SOLUTION: We set up a direct connection between the camera and the linux box, powered 

off of the APS (alternate power supply). 

PROBLEM 217: The linux box finally dies. It sounds like it would work well as a coffee 

grinder though. 

ESA_Neil turns the linux box off. 

ESA_Neil, MCC and CAM_Morten give up and go home (it is late). 

22 nd March 2005 

ESA_Neil tries to reboot the linux box, it actually works and sounds better than it did last 

night. In parallel we try to procure a new one also… 

ESA_Neil turns on the CAM, directly connects it to Aalborg and leaves Morten to play. 

ESA_Neil manufactures bolts for Magic and S-Band. 

314






CAM_Morten reports that the camera is fine, but it seems that the optimum parameters have 

shifted significantly for no apparent reason. He takes this picture with it (it is pointing up into 

the far corner of the cleanroom). This is a “full” picture. 

ESA_Neil and CAM_Morten perform the following tasks: 

An angle test reveals that the pictures need to be rotated by 121 degrees anticlockwise. After 

this rotation the “up” direction in the picture will be aligned with the –y axis of the spacecraft. 

These images are “large thumbnails” which were both taken in and out of context (using 

OBC, EPS and RF). 

315






CAM_Morten is worried about the focus, so we point the camera out of the window and try 

again. It is MUCH brighter, so the exposure time has to be lowered 

PROBLEM 218: Problem 213 rears its UGLY head again. We are getting nominal mode 

beacons, but cannot uplink at all. 

TEST: Without rebooting the OBC we try power cycling the UHF unit – it makes no 

difference. 

TEST: We power cycle the groundstation and restart the ground software – it makes no 

difference. 

TEST: We kill the watchdog, go into recovery mode and then reboot the OBC – it makes no 

difference. 

TEST: We do a direct connect between the groundstation and the OBC (missing out both 

radios and the RF link). This works fine. 

TEST: We put the RF link back in the loop. It works again! But intermittently… (the worst 

kind of problem). 

The satellite batteries run dry and the spacecraft shuts down. 

ESA_Neil puts the CAM on the APS, disconnects the PCU from the loads and charges the 

battery via a simulated solar panel. 

CAM_Morten takes a full picture of ESA_Neil through the cleanroom window. NOTE: 

There are blinds, panes of glass and a pane of plastic in the way. Also, the distance is only 

50m instead of 700km, and the lighting is not the same as in space. 

316






CAM_Morten and ESA_Neil declare the camera checkout complete. 

ESA_Neil torques and glues all the bolts and cables inside the camera. 

317






PROBLEM 219: One of the small bolts securing the CAM PCBs to the enclosure snaps 

upon tightening. 

SOULTION: ESA_Neil repositions it. It is just long enough to still function properly. 

ESA_Neil then closes the box and torques and glues those bolts too. 

ESA_Neil integrates the FM CAM to the flight model. 

318






ESA_Neil disintegrates MAGIC from the FM, grinds the two remaining bolt holes wider (as 

per modification 116), adds the two new bolts and reintegrates it. 

MODIFICATION 122: MAGIC is turned around to make room for the CAM connectors 

next to the relay (the original 3D model of MAGIC was incorrect in terms of the top-plate 

orientation). 

ESA_Neil buzzes through the ground connection between MAGIC and CAM. It seems to be 

fine, as it is to the rest of the structure. 

ESA_Neil buzzes through the CAM power connector to the enclosure, making sure that the 

electronics are isolated as they should be. They are. 

23 rd March 2005 

ARRIVAL 72: The OBC flight proms arrive. 

319






ARRIVAL 73: The EPS flight PIC arrives. 

ESA_Neil follows the following instructions supplied by OBC_Karl: 

1) Before doing anything, be sure you have PLCC-32 compatible chip extractor. 

Jason has one. 

2) Wear anti static gloves during all the following steps. 

3) Take the gray bags out of the transparent bag and throw the transparent one in 

the bin (it is not anti static). Note that the grey bags are marked "1" and "2". 

4) Disassemble the OBC box such that the three sides holding PCBs and 

connectors are still attached. Note how the plates are oriented w.r.t. each other 

so you can put it back together correctly later. 

5) Think for a second about how to disassemble the last alu parts without 

stressing wire solder joint on the bottom PCB, then unscrew the last two sides 

from the plate holding the PCBs. 

6) Unscrew the screws holding the PCBs without removing the screws from the 

alu plate and the PCBs. You just want the top PCB to come lose. One of the 

320






nuts is glued to the interface card. This is perfectly normal. Do not pry the nut 

lose. 

7) Carefully move the top PCB (interface card) away to expose the main PCB 

underneath. 

8) Identify the PROM sockets. There are three: ROM1, ROM2 and the utility 

processor PROM. Number 1 and 2 currently hold Redboot and are up for 

replacement by flight PROMs containing eCos and OBDH. PROM 1 and 2 are 

located side by side with a white designators on the green PCB. 

9) Extract PROM 1 using the PLCC-32 chip extractor. Be very careful not to 

apply any torque to the socket - it may shatter. The two gripping fingers of the 

extractor are inserted into the two grooves in opposite corners of the socket and 

the extractor is squeezed together forcing the gripping fingers together and 

upwards thus extracting the chip. 

10) Having successfully extracted PROM 1, place the PROM in an anti static bag 

marked "Redboot 1". 

11) Open the gray bag marked "1" and take the item contained in the bag out. This 

item is from hereon out referred to as "flight PROM 1". 

12) Clean flight PROM 1 with IPA tissue. Pay special attention to the pins and 

scrub them thoroughly. Make sure no tissue residue is left on the chip 

(especially the pins) before continuing. 

13) Insert flight PROM 1 into the empty socket 1 with text side up (pins down). 

Don't press the flight PROM down until you are sure it is level on top of the 

socket with all pins lining up with the contact groves of the socket. Be sure the 

PROM is oriented correctly with the cut-off corner of the PROM matching the 

cut-off corner of the socket. 

14) Apply pressure using you index finger vertically onto the flight PROM until it 

clicks into place. Post-press "a little bit" around the top surface of the flight 

PROM to ensure proper insertion contact all the way around. 

15) Repeat step 9 to 14 with PROM 2. Be sure to place the old PROM in an anti 

static bag marked "Redboot 2" before opening the gray bag marked "2" 

containing flight PROM 2. 

321






16) With flight PROM 2 successfully inserted reassemble the OBC by first putting 

the interface card back on top of the main board. Be ware of the wires running 

between the boards. Screw the PCBs together. 

17) Making sure not wires are trapped between the alu plates, screw the connector 

plates onto the PCB alu plate with all wires from the bottom board coming up 

behind the cut-off corner of the interface card. 

18) All went well and you are ready to test the OBC. hook it up to the APS and 

debugging computer. 

19) Open Minicom and set the baud rate to 57600 (this should be default from 

now). 

20) Power up the OBC with 28V. 

21) Type on the debugging interface and watch the 

debugging help screen scroll down. 

22) If(21 is not successful) go to step 23. Else go to step 24 

23) Go to "Fiji". 

24) Clap your little hands and put the rest of the OBC box back together. 

322






These steps all proceeded well. Although the box has not been put back together since the job 

of conformal coating remains. 

After a successful stand-alone test the OBC is put back into context and the spacecraft is 

booted up. The following checks were performed and passed: 

- Nominal mode was held 

- Killing the watchdog resulted in collapse to recovery mode 

- Nominal mode resumed and held 

- Telecommand was as good (or bad) as usual 

- A large thumbnail was downlinked on UHF 

- A large thumbnail was downlinked on S-BAND 

- The spacecraft was turned off, the battery measurement connecter unplugged, and 

then the spacecraft was booted up again 

- After a pause in safe mode to ascertain that the battery was “broken” the spacecraft 

proceeded to nominal mode as usual and ignored the battery voltage as a result 


ESA_Neil powers down the spacecraft and disconnects the PCU. The PCU is then opened 

and the test PIC is removed and replaced by the “flight” PIC. 

323






The PCU is then reconnected and the spacecraft is booted back up again. 


- Killing the watchdog resulted in collapse to recovery mode 

- Nominal mode resumed and held 

- Telecommand was as good (or bad) as usual 

- A large thumbnail was downlinked on UHF 

- A large thumbnail was downlinked on S-BAND 

- The spacecraft was turned off, the battery measurement connecter unplugged, and 

then the spacecraft was booted up again 

- After a pause in safe mode to ascertain that the battery was “broken” the spacecraft 

proceeded to nominal mode as usual and ignored the battery voltage as a result 


ESA_Neil sets the groundstation up such that it can be controlled from his office. 

ESA_Neil successfully downlinks a whole picture on S-BAND. It is identical to the one 

received via the CAM yesterday, but this time it is in full context. 

324






PROBLEM 220: Problems 214 and 215 manifest themselves again. The OBC is power 

cycled regularly as if the ping/pongs are not functioning correctly. (This happened during a 

UHF full picture downlink, but previous experience suggests that this is not the reason.) 

TEMPORARY SOLUTION: Sending the “reset PDU” command allows nominal mode to 

be held next time. 

ESA_Neil suspects that the problem is time related: it always seems to happen after around 

two hours of up-time. We therefore wait for it to happen again. 

A full picture (the same as above) is successfully downlinked via UHF. This takes about 30 

minutes with a total loss of 27 packets out of over 20000. The losses can be seen on the 

picture – the errors are dropping entire chunks in the term.exe code (must check with 

OBC_Karl). The large line across the picture common to S-Band and UHF must be in the 

memory on the OBC. 

SUGGESTION: OBC_Karl suggests that maybe the EPS “ticket” in the pings is 

overflowing, such that the pong reply does not match anymore. 

This seems like a rather good suggestion. OBC_Karl, ESA_Neil and EPS_Fulvio jump on the 

EPS code to check it out. It certainly seems like it could be the problem, since the ticket is 

325






stored as long variable, but passed to OBC in a single byte, which will not be representing the 

variable accurately after 256 pings. 

In testing this would have taken over four and a half hours to manifest itself, which is about as 

long as we have ever had the s/c up for. It has been much easier to detect now because the 

pings are every 20 seconds instead of 1 minute, and the OBC does not need assistance to be 

rebooted, since the flight proms are installed. 

The problem is recreated (OBC being cycled in recovery mode and pongs are being ignored) 

and ESA_Neil emulates the OBC so that we can see what EPS is sending. Sure enough the 

ticket is low, implying that it has “wrapped around”. 

The time for the problem to manifest itself is around 90 minutes. This does correspond to 256 

pings at around 21 seconds each. 

ESA_Neil opens the PCU, takes out the “flight” prom (version OTP 1.2) and puts the 

bootloader PIC back in again. 

EPS_Fulvio makes the necessary s/w change and uploads it to the FTP as version 7.4. 

ESA_Neil uploads 7.4 on to the PDU and starts to test again. The pings are about every 22 

seconds (takes some response time) so we expect around 94 minutes. 

PROBLEM 221: EPS software version 7.4 still resets the OBC when the keepcount variable 

rolls over. After that it is stable though. This must mean it has been reduced to a boundary 

problem. 

MODIFICATION 123: EPS_Fulvio changes the software so that the ticket is always “42” 

(0x2A). This is done since it is impossible for the OBC to get into a loop where it responds to 

pings automatically without “listening” to them, so the content of the ticket is irrelevant. 

ESA_Neil uploads 7.5 on to the PDU and starts to test again. 


The OBC was reset overnight, however, this was because ESA_Neil forgot to put the solar 

panel simulator on, so the spacecraft battery ran low and the PCU entered safe mode, and then 

eventually ran dry and turned off totally. In the meantime the OBC was up for more than four 

hours, implying that modification 123 was successful in fixing problem 221. 

ESA_Neil declares the EPS software flight ready. Again. 

326






The TMTC log of the overnight session is on the FTP in EPS/ 2005-03- 

23_Ver_7.4_overnight_log.txt 

ARRIVAL 74: NCube-II arrives with NCUBE_Åge. This time it has a metal protection box 

that is depressing the kill switches. The antennas have not deployed, and the RBF pin appears 

to be smaller, as required. 

ESA_Neil and NCUBE_Åge load NCube-II into the +x T-Pod. No issues are encountered. 

327






ESA_Neil disintegrates the OBC and conformal coats the upper side of both PCBs. It is 

decided to not coat over the top of the chips in the their sockets because: 

- Trapping air underneath a chip would then produce a positive pressure which 

could potentially push the chip up out of the socket 

- Conformal coating could potentially run down between the pins and connectors 

and cause problems 

- The only reason to do so would be to prevent a little out-gassing from the plastic 

socket. This is not critical. 

- It makes it much easier to replace them if we ever need to 

328






After coating the boards the OBC is placed in the vacuum chamber to remove all the trapped 

air. This works without a problem. 

EPS_Fulvio reports that the flight PIC will not be in ESTEC until Tuesday (29 th ). 


ESA_Neil attempts again to close the battery box. It is not possible to get all the bolts in. 

(This is problem 209.) 

MODIFICATION 124: ESA_Neil slightly widens all the holes on the side panels of the 

battery box, and slightly shortens three of the bolts. It is then possible to close it properly. 

All bolts inside the battery box are tightened and glued. The battery box is closed, then all 

bolts on the outside are torqued and glued. 

329






The missing screw lock is added to the PIN box, then all bolts on the PIN are tightened and 

glued. (This means cycling the UHF connector on the PIN.) 

ESA_Neil disintegrates the OBC, carefully reverses the direction of the PCB securing bolts, 

turns the PCBs upside down allowing them to sit naturally without stressing any connections, 

and then uses kapton tape to secure the sides together and the PCBs level. 

The “upper” (computer) PCB is removed from the mounting bolts so that conformal coating 

can be applied to the reverse of the interface card. Some paper is present on top of two of the 

chips, but rather than risking removing it, it is carefully coated over. The computer is then 

replaced and the reverse of that is coated also. 

The bolts (all but one, which is too hard to get out without disturbing the wet coating) are 

carefully reversed again back to their original mounting positions and the OBC baseplate is 

secured on the “top” as normal. The two connector-laden sides are then replaced to support 

the baseplate and PCBs in the drying position. 

The OBC is placed in a vacuum for 15 minutes to remove air bubbles from the coating. 

330






The OBC is powered up to check that it has survived its ordeal. Everything seems fine. 

The flight battery, and associated equipment (PTFE wire, solder, flux, gloves, IPA tissues, lint 

free tissues, kapton) is prepared for shipment to Naples for integration. 


AMS_Howard arrives to troubleshoot the uplink reliability problem and to run some 

functional tests on the S-Band unit. He sets up the test gear while ESA_Neil boots up the 

spacecraft. 

ESA_Neil demonstrates the uplink problem for AMS_Howard, who then manages to replicate 

it on a separate test groundstation. 

These notes are from AMS_Howard: 

331






o UHF uplink unreliable 

Firstly an alternative UHF groundstation segment was assembled comprising of another 

Symek TNC-3S TNC and a Yaesu FT-817 radio that allows frequency adjustment in FM 

down to 100Hz. On the nominal 437.250MHz commanding was still unreliable. 

The transmit frequency of the alternate groundstation was checked with a frequency meter 

and it was within about 300Hz. 

The alternative groundstation frequency was dropped down 2.5kHz to 437.2475MHz and 

uplinking became 100% reliable. 

On the ESA groundstation the frequency was adjusted 5kHz down to 437.245 and uplinking 

was still 100% reliable. 

A probe was placed inside the UHF box to look at the receiver's eye pattern and connected to 

an oscilloscope. With the groundstation operating at the 437.250 'nominal' frequency this 

showed the eye pattern spending a lot more time in the high state to the low state, and there 

was some overshoot in the low state. It was easily demonstrated that by adjusting the uplink 

frequency down to 437.245, the eye pattern became far more symmetrical and less distorted. 

The uplink frequency was further adjusted up and down for the best eye pattern, with the best 

results still resulting from using 427.245MHz. 

The opportunity for measuring the UHF downlink frequency was taken and this was shown to 

be at 437.252MHz. 

Although no problems were apparent by using 437.245MHz as the groundstation uplink and 

downlink frequencies, the ESA groundstation transceiver was adjusted so that it would 

operate to compensate for the differing uplink and downlink frequencies. Memory 0 of the 

Kenwood TH-F6A was set to transmit on 437.245MHz and receive on 437.250MHz. The 

band in use on the radio was set to 5kHz steps - it may be possible to more finely tune the 

radio using the second band, although this was not tried and as there were no apparent 

problems no further adjustments were made. 

o UHF Receiver Sensitivity tests 

The object of this test is to ensure that the spacecraft's UHF receiver is within acceptable 

limits. At this stage of the project lifecycle, unless this test shows up a serious problem there 

will be no attempt to make any corrections. 

An RF signal generator is connected to a TNC and packets are repeatedly sent at varying 

power levels. 

332






The output of the RF signal generator combined with the losses of the jig's RF cables are 

calculated with a measuring receiver, confirming signal power and deviation. 

By adjusting the output of the RF signal generator, a 'pinch point' can be found within two or 

three dB where the number of successful packets drop from almost 100% to 10% or less. The 

AX.25 packet sizes being sent for this test were 200 bits. 

Ten packets are sent at each power level and a count taken of the number of successful 

packets received. Although statistically ten packets is a small sample size, each packet has to 

be sent manually and there is a limited amount of time available. Even with only ten packets 

per sample, the trend is clear. [**** Neil can you confirm with Karl the total AX.25 packet 

size including AX.25 headers??], 

More usually a full BER test would specifically detect failures at the bit level, in terms of 

number of bit errors per thousand, rather than the packet level. However as a true BER test 

would require re-blowing the spacecraft TNC firmware, something not to be recommended at 

this stage of integration. 

Under a separate controlled environment that can simulate both packet and BER loss, it will 

be possible to repeat the test by producing the same packet loss and then, measuring the BER. 

at -96 dBm: 10/10 

at -110 dBm: 3/10 

at -105 dBm: 10/10 

at -106 dBm: 10/10 

at -107 dBm: 9/10 

at -108 dBm: 9/10 

at -109 dBm: 7/10 

at -110 dBm: 4/10 

As the increase in loss of packets is dramatic over a fairly small loss of carrier power, this will 

probably coincide to within a dB or so the 10^-3 BER. 

The results we achieved are perhaps 10dB or so shy of what would be expected from a truly 

optimised configuration. Given time this could be improved, however at this stage it is easier 

to increase groundstation power than to start attempting to debug and potentially re-engineer 

the UHF section. 

Typical improvements that affect performance are the choice of FM IF filter(s) and the noise 

figure of the front end: the combined total of these alone can easily be 10dB, such as when the 

S band groundstation BER tests have been conducted. Previous tests when resolving the off 

frequency receiver during integration of the UHF receiver revealed that the IF filter is quite 

narrow, so minor frequency deviations or even group delay could be part of the 10dB 

degradation. 

333






A typical groundstation would be able to overcome this 10dB with little difficulty. Most 

groundstations will be operating of the order of 15dB more power than the satellite downlink. 

ESA_Neil glues the toroids, screw locks, bolts and potentiometers in the UHF box. 

ESA_Neil conformal coats the FM UHF, avoiding the radio compartment so as not to de-tune 

anything. The lid is then closed and the bolts torqued and glued. 

ESA_Neil carefully glues the proms into their sockets in the OBC. So as not to restrict any 

future removal of the chips the glue does not cover the pins or the extraction holes, but simply 

secures the surface of the chips to the two remaining corners of the sockets. 

ESA_Neil tightens and glues the screw locks in the OBC, then closes the box. We are 

temporarily missing two ‘torx’ screws, so integration cannot be completed yet. 

We plug the UHF and OBC in to test that they still work. 

334






PROBLEM 222: The downlink on UHF is very dodgy all of a sudden. Whenever 

significant UHF downlink reaches the term.exe it crashes. 

SOLUTION: ESA_Neil changes the UHF-OBC cable, this fixes the problem. 

o Commanding unreliable 

It is believed that a connection problem existed between the 9 pin D connectors between some 

of the units leading to faults where the OBC would receive data from the TNC that was not 

recognised as a valid command, for example being too short. In KISS mode, the TNC will not 

send any data to the OBC if the AX.25 CRC is invalid - these frames are dropped. 

Since switching some cables on the bench test bed and subsequently assembling the UHF and 

S Band units into the spacecraft this fault has not recurred. 


ESA_Neil cleans and conformal coats the underneath of the PCU boards, then puts them in 

the vacuum chamber to remove air from the coating. 

335






o S Band Bit Error Rate tests 

It is not possible to directly conduct BER tests on the transmitter because the firmware 

version in the S band TNC does not support direct BER testing. For a traditional BER test, the 

transmitter is set up to send a continuous stream of ones and/or zeros. Due to the pseudo 

random scrambling of the modem, it is not necessary to test both one's and zero's, either will 

be sufficient. 

For testing the S band link a similar method is used to that of testing the UHF receive 

sensitivity, except that rather than sending ten packets, 1280 were sent with an average data 

payload size of 91 bytes each [**** Neil can you confirm with Karl the total AX.25 packet 

size including AX.25 headers??], close to 10^3 bits per packet. Unlike sending packets on the 

UHF uplink, it is very easy to generate a large number of packets by requesting a file 

download from the OBC. 

Initially it was clear that the receiver would need to be placed away from the transmitter as the 

leakage was very high. The receiver (an AOR8600 with a DB6NT MKU232A2 30dB gain 

336






0.6dB noise figure preamp, and 10.7MHz Symek IFD with the narrower 80kHz filters) was 

placed in a Faraday cage. 

To confirm that there was no leakage in the Faraday cage, a short test with the S band 

transmitter on and the receiver listening in the Faraday cage resulted in no discernable signal 

being received. 

The 1.5W port of the S band transmitter was attached directly to a 20dB pad. This in turn 

went to a variable attenuator (0 to 64dB in 0.25dB steps) and further optional 50dB and 30dB 

pads. This was attached to a cable that fed over several meters to the receiver inside the 

Faraday cage. 

With the S band transmitter switched on, the combination of the 20dB pad, variable attenuator 

(set to 0dB) and cable resulted in s signal of -23dBm being measured at the end of the cable in 

the Faraday cage using an HP435A power meter, so the cable itself had very significant 

losses, of the order of 33dB. 

In order to conduct the actual packet rate loss tests, an additional 80dB was placed in line at 

the S and transmitter. The following results were obtained by inserting additional attenuation 

from the programmable attenuator: 

RxPower PktsReceived 

-108dBm 1277 

-113dBm 1250 

-114dBm 1261 

-115dBm 1027 

-116dBm 1084 

-117dBm 1051 

337






-118dBm 931 

-119dBm 1 

The 'pinch point' of -118dBm is about 4dB better than those achieved in controlled tests 

showing real BER of 10^-3. 

It should be noted that there was some unexplained variation in signal strength of about 2dB 

in the Faraday cage during these long packet transmissions. It is possible that under low signal 

conditions, S band RF could be leaking into the cable outside the Faraday cage, although 

earlier leakage tests would seem to dispute this. 

Despite the above concern, the result above is certainly of merit and it would seem that there 

is no problem in the S band data link at all. 

In the future a further test will be performed using the same receive equipment and the 

engineering model to generate a BER and packet loss rate equivalent. 

AMS_Howard and ESA_Neil integrate S-Band to the FM. An SMA saver is placed on the 

connector of the +z patch antenna, and an attenuation cap over the top. 

338






AMS_Howard and ESA_Neil integrate UHF to the FM. 

PROBLEM 223: There is not really enough space for the RF cable or the RS232 cable from 

UHF, as it is too close to the shear panel. 

SOLUTION: We remove the saver from the RS232 port and add the flight cable (other end 

is temporary). 

MODIFICATION 125: We add a right-angled SMA saver to the RF port. 

339






PROBLEM 224: The saver on the audio port (to S-Band) on the UHF is in the way of 

integration of the PIN. 

SOLUTION: We add the flight cable to the UHF audio port (other end is temporary). 

ESA_Neil integrates the PIN. 

PROBLEM 225: The holes in the PIN are large than they should be, causing potential load 

bearing problems. 

MODIFICATION 126: For the PIN ESA_Neil uses M4x12mm bolts and two M4 washers 

instead of M4x10mm bolts with no washers. 

340






o Post integration testing of UHF and S band units 

PROBLEM 226: Once the UHF and S band units had been installed, it no longer seemed to 

be possible for the groundstation to receive the UHF downlink (the S band downlink was 

OK). 

On listening to the UHF downlink with an audio receiver the transmissions from the UHF 

downlink seemed to be predominantly a tone of approximately 1kHz, totally unlike the 

expected pseudo random noise normally associated with such signals. 

As the cabling and been substantially reworked, an investigation ensued into whether it may 

be RF inadvertently entering the UHF unit and creating the incorrect modulation. 

A dummy load was placed on the UHF RF port, and the fault cleared up. 

It was noted that on the bench prior to integration into the spacecraft, there were a number of 

inter-series adapters and a 6dB attenuator before a simple monopole antenna. 

A programmable attenuator was placed in line with the UHF RF port and the antenna. At 

around 6dB of attenuation the 1kHz tone stopped and the link seemed perfect again. 

A fixed 6dB attenuator was placed in line but this did not resolve the problem. Suspecting 

additional attenuation in the cables used to attach the programmable attenuator, a fixed 10dB 

pad was attached and the problem was resolved. 

Despite moving various interconnecting cabled it is as yet unclear where the RF is re-entering 

the UHF box. It is hoped that once the aluminium side panels are attached the problem will 

resolve itself. Possible alternatives are the use of ferrite on some of the leads entering the 

UHF box or the use of screened cables. If the RF is entering via the power leads, screened 

cables may not help. 

341






S-Band commanding works, S-Band data works, S-Band telemetry works. 

The test groundstation laptop dies. 

AMS_Howard departs. 


The MCC_Martins come to see the spacecraft and end up staying to help all day. Firstly they 

label the wires on the half-made flight harness. 

ESA_Neil demonstrates the spacecraft for SYS_Jörg, mainly the RF functionality. We 

measure the UHF transmission as 506 mA (including power up). 

ESA_Neil, SYS_Jörg and the MCC_Martins cover the exposed lateral panels and honeycomb 

core with kapton tape, taking a lot of time and care not to get any air bubbles underneath. The 

most successful results are obtained by applying the kapton while the lateral panels are still 

mounted. 

342







ARRIVAL 75: The flight PIC (ver 1.3) arrives. 

ARRIVAL 76: The S-Band patch antenna bolts arrive. 

ARRIVAL 77: The new shunt resistor arrives. 

343






ESA_Neil inserts the new PIC into the PDU. 

The rest of the hardware is integrated into the spacecraft, so just the OBC, the BATT and the 

PCU are powered up for testing. They seem to work fine. 

ESA_Neil checks the ping ticket by emulating the OBC using a laptop connected to EPS. 

The ping ticket is 2A as, of course, it should be. The pings look stable, the telemetry seems 

fine. However, the nominal mode beacon looks a little odd, since it looks as if the formulas to 

convert the 12-bit measurements from the EPS PIC into human-readable telemetry have not 

been correctly specified. 

Some details are recorded in order to troubleshoot this issue. The measurements are taken 

using a voltmeter on pins 1 (battery voltage) and 7 (battery temp) of the 9-pin measurement 

connector on the battery box. Then some voltages are simulated using a power source. 

Beacon data Measurement (0-5) Actual 

Temp 473.2 deg C 2.043 V >= 23 deg C 

1193.75 deg C 0 V >= 23 deg C 

Voltage 28.25 V 4.72 V 23.38 V 

27.09 V 4.5 V 22.5 V (simulated) 

23.92 V 4.0 V 20 V (simulated) 

22.76V 3.8 V 19 V (simulated) 

30 V 5 V 25 V (simulated) 

27.7 V 4.62 V 22.92 

The safe mode entry level is found to be 3.7 V measured (equivalent to 18.5 V actual). 

The safe mode exit level is found to be 4.05 V measured (equivalent to 20 V actual). 

PROBLEM 227: The PCU software is coded such that it is assumed that the 0-5V reading of 

the battery voltage must be multiplied by 6 in order to get the actual value. In fact the BCR is 

344






only dividing the voltage by 5 in order to give the reading. Therefore the safe mode entry 

threshold, the safe mode exit threshold, and the battery broken threshold are all incorrect. 

This is not acceptable. 

The EPS team is out of PICs and would take a while to get any. However, it is critical to the 

schedule that we have a replacement within two days. An alternative must be found. 

ESA_Neil removes the lateral panels. This involves also removing solar panels 7 and 10. 

PROBLEM 228: The new antenna bolts are too narrow and do not securely hold the clips 

that fix the attenuation caps on. The old bolts were 7.67mm in diameter, but the new ones are 

only 6.88mm. 

The OBC has a final weight of 1221.4 Kg, including ten 9-pin savers and one 15-pin saver. 

(A 9-pin saver weighs 12.2 g). 

The MCC_Martins and ESA_Neil integrate the magnetometer and the OBC to the flight 

model. 

345






The MCC_Martins and ESA_Neil commence the harness routing. Preliminary power harness 

routing is done from the PIN to OBC, CAM, PIC, S-BAND and UHF. Only one end of each 

cable has a connector, and no connectors are cycled at this point. 


ESA_Jason reviews the preliminary harness routing and advises on the best way to continue. 

ESA_Neil disintegrates the UHF, upgrades the right-angled saver to a flight adapter, and then 

cycles the antenna side of the flight adapter by adding the flight coax cable. 

346






ARRIVAL 78: The L-profile for the PCU is manufactured. 

ESA_Neil routes power harness from the PIN to UHF, PIC, CAM, OBC, S-BAND and 

ACDS. The PIN ends are all connected, cycling all the PIN 15-pin sockets once as the savers 

are removed. The other ends of the cables are left too long, empty, and labelled. 

347






ESA_Neil routes power harness to the three T-Pods. The Pod ends are all integrated, cycling 

all the PIN 15-pin sockets once as the savers are removed. The other ends of the cables are 

left too long, empty, and labelled. 

ESA_Neil routes the Pyro power harness line from the PCU to the PIC, the shunt resistor 

cable to the PCU, and the PTT cable to the PCU. The PCU connectors are added (although 

not integrated since the PCU is not ready), but the other ends of the cables are left too long, 

empty, and labelled. 

ESA_Neil routes data cables to the OBC from UHF, EPS, PIC and CAM. These cables are 

connected at the non-OBC ends (apart from EPS since the PCU isn’t ready). The other ends 

of the cables are left too long, empty, and labelled. 

ESA_Neil glues harness clamps to the spacecraft structure and system as necessary, moving 

the routed harness aside carefully as he does so. For particularly heavy sections double 

clamps are placed. A total of 73 clamps are glued, taped into place securely and then left 

overnight to dry. The following pictures show their positions, starting in the –x0y 

compartment and progressing anticlockwise around the spacecraft: 

348






349






350






351






352







ESA_Neil performs the following harness integration work, with some help from the 

MCC_Martins: 

The t-pod harness is rerouted to the FPP, but the -x pod does not reach and has to be replaced 

(this means cycling the connector once on the pod). 

The debug port on the OBC is committed and routed to the FPP. 

The S-Band RS232 port is committed and routed to the FPP. 

353






A T-Pod/FPP/PCU cable is made passing pins 1,2,5,6,9,10,,14,15,18,19,22,23 to the PCU 

from the FPP. 

Commit UHF power harness at both ends. Add attenuation to UHF (20dB) and saver to 

antenna end of cable (use test cable as saver). The UHF power harness buzzes through ok, so 

the PIN is powered up with the alternative power supply: 155 mA = ok (UHF + PIN) 

Commit CAM power harness at both ends. It buzzes through fine and powers up using APS 

to consume 165mA = ok (UHF, PIN, CAM). 

Commit OBC power harness at both ends. It buzzes through fine and powers up using APS to 

consume 188mA = ok (UHF, PIN, CAM, OBC). 

Commit S-BAND power harness at both ends. It buzzes through fine and powers up using 

APS to consume 280mA = ok (UHF, PIN, CAM, OBC, S-BAND). 

Commit audio line between UHF and S-BAND. It buzzes through ok. Power up with APS 

and then command S-BAND carrier up: 705mA (UHF, PIN, CAM, OBC, SBAND carrier 

up). Commanding the TNC up adds an extra 21mA. 

Commit RS232 from UHF to OBC. Buzzes through ok. Power up groundstation, power up 

spacecraft. TC and TM work fine. 

Commit RS232 from EPS to OBC (other end left as PCU is not in yet). 

Commit RS232 from CAM to OBC. Power up groundstation and spacecraft, send a ping to 

CAM, retrieve TM, get the pong from CAM. 

354






Commit CAN from MAGIC to OBC. Buzzes through ok. 

T-Pod 1 = +X, pins 1,2,9,10 from pod go to pins 3,4,16,17 on FPP, should then loop back 

through arm plug to 1,2,14,15 on FPP to go to POD-1 on PCU 

T-Pod 2 = +Y, pins 1,2,9,10 from pod go to pins 7,8,20,21 on FPP, should then loop back 


T-Pod 3 = -X, pins 1,2,9,10 from pod go to pins 11,12,24,25 on FPP, should then loop back 


ESA_Bas solders the solar panel harness, including the protection diodes. 

1 st April 2005 

ARRIVAL 79: The fourth attempt at the flight EPS PIC arrives. 

ESA_Neil inserts OTP 1.4 then powers up the PCU and BATT. OBC is emulated using a 

laptop, and the battery measurement connector is left detached so that the battery voltage 

reading can be simulated with an external power supply. A voltmeter is connected across the 

load pins for S-Band to see the transition between safe and recovery mode. 

A lot of testing reveals that the drop to safe mode occurs at a measurement of around 4.5V, 

which is equivalent to 22.5V on the battery. The transition back to recovery mode occurs at 

slightly less than 4.9V, which is equivalent to 22.3 on the battery. At a measurement of 3.1V 

it is a assumed that the battery is broken. 

EPS_Fulvio confirms the thresholds above are correct. 

The PCU seems to ping and command ok. Since the rest of the functionality was tested with 

the previous (unchanged) code, ESA_Neil declares the software ready and glues the chip into 

the socket with AY138. 

355






ESA_Neil glue more cable clamps down, including those for the routing on the -y lateral 

panel. The sun sensor and magnetorquer coil wires are also routed at this time, specifically 

along the coils themselves. 

ESA_Neil does the following harness tasks: 

Commit the debug port to the FPP, using pins 2, 3 and 5 as usual. 

Commit the S-Band RS232 to the debug port on the FPP, pin 2 of s-band to pin 8 of debug 

FPP, pin 3 of s-band to pin 9 of debug FPP. 

The spacecraft is powered up and the debugger tested. It works fine. Then the OBC is told to 

go into S-Band data mode, and the S-Band TNC is commanded up. 

A loop-back connector is applied to the debug port, passing pin 2 to pin 8 and pin 3 to pin 9. 

The groundstation is powered up and data is downlinked on S-Band without a problem. 

All the connectors are attached to the FPP, and the screw-locks are tightened: 

- rightmost is EPS safe / arm 

- next is t-pod safe / arm 

- next is pyro safe / arm 

- next is external power 

- 9-pin is debug / s-band 

The battery box weighs in at 1955g. 

ESA_Neil and ESA_Marie integrate the battery box to the structure, using two washers 

underneath each foot to stand it off from the panel slightly (for thermal reasons). Washers are 

also added on the top, since the feet have M5 holes for the M4 bolts. 

356






Commit MAGIC high power end (other end still changeable if necessary, since PCU is not in 

place). 

Commit MAGIC low power from PIN. It buzzes through fine. On power up using the APS 

the total consumption is 313mA (UHF, PIN, CAM, OBC, S-BAND) = 33mA for Magic = ok. 

Commit –x thruster cluster, it buzzes through fine. 

Commit the MAGIC -> FPP -> PMS (Pyro) connection. It buzzes through fine. 

Commit the branch valve and main PMS MAGIC line. They buzz through fine. 

Commit +x thruster cluster, it buzzes through fine. 

357






Redo the routing and ty-wraps on the PMS, covering the old (sharp) clamps with aluminium 

tape to protect the cabling. 

ESA_Neil adds savers to the FPP, and bolts to hold them on. 

ESA_Neil conformal coats the PCU. 

ESA_Neil glues the FPP to the structure. 

2 nd April 2005 

ESA_Neil tightens and glues all screw locks and bolts in the PCU. At the request of EPS the 

boards are only bolts in at one side. However, a cable tie is run through the bolt holes on the 

other side of the board, which will add a flexible protection against the boards “falling out” of 

the rails there. 

358






The PCU is then closed, torqued and glued. 

PROBLEM 229: The PCU is closer to the walls than we realised and it doesn’t fit because 

there are harness clamps in the way. 

MODIFICATION 127: ESA_Neil cuts and Dremmels the harness clamps by the PCU flat. 

PROBLEM 230: The holes in the PCU L-Profile do not match those in the lateral panel. 

359






MODIFICATION 128: ESA_Neil slots the holes in the PCU L-profile with a Dremmel. 

The PCU weighs in at 2398 kg. 

ESA_Neil integrates the PCU – it now fits fine. 

SYS_Jörg and ESA_Neil glue the thermistors and route the TCS harness. 

360






ESA_Marie and ESA_Tor glue harness clamps and cut, strip and crimp in order to prepare the 

solar panels for harness integration. 

SYS_Jörg makes “JLI” (Jörg Layered Insulation) out of aluminium foil and kapton tape. We 

then test a segment of it in the vacuum chamber to see if there are bubbles and how they fare. 

There are a few small ones, which burst eventually, but this is no serious problem. 

361






The PCU weighs in at 2398g. 

ESA_Marie and ESA_Tor test the solar strings with a simple halogen lamp about 50cm away 

from the cells. All strings present around 30 volts in this configuration, implying that none of 

them are broken. They then proceed to complete the harness clamp gluing and solar cell 

harness crimping on the lateral panels. 

ESA_Neil and SYS_Jörg slowly wire up the flight version of the “five headed orange 

monster” from EPS. This takes several hours. 

ESA_Neil integrates the five-headed-cable and the load cable while SYS_Jörg prepares 

thermal blankets. 

We power up the spacecraft properly for the first time, while monitoring the current running 

through the ABF plug. 

362






PROBLEM 231: The PCU is consuming a very strange current (190mA), even when the 

load cable is not plugged in. This implies that it is not functioning correctly – which is 

practically irreversible after the conformal coating. 

ESA_Neil buzzes through the five-headed-cable one more time. He discovers that two pins 

have been swapped, but these are on a redundancy (pins 3 and 14 on the battery connector), sp 

this should not make a difference. 

ESA_Neil integrates the test harness to the FM instead of the flight harness (using savers of 

course), but this does not fix the problem. This implies then that the fault is a physical one 

with the PCU itself, maybe it did not survive conformal coating. 

ESA_Neil tries charging the spacecraft using the solar panel simulator. There seems to be 

500mA running through the ABF, so he powers it back off again, gives up and goes home. 

4 th April 2005 

After discussions with EPS the suggestion arises that measuring the current consumption 

through the ABF is not accurate, since this is the current from the battery, not from the BDR – 

therefore it could well read “incorrectly”, since it is not at 28V. 

ESA_Neil reconfigures and measures the current consumption after the BDR. The spacecraft 

powers up fine. 

ESA_Bas solders all the positives terminals of the TCS thermistors into one wire, and then the 

same for the ground. 

ESA_Neil glues the microwave cable clamps into position, and then charges the spacecraft 

battery. 

363






The EPS-UHF is buzzed through then commited. The s/c is powered up and a handheld radio 

is used to detect the recovery mode beacon (RS232 is not in yet, so nominal mode cannot be 

achieved). The beacon comes up fine. 

The BATT-PCU measurement cable is committed. The s/c is powered up and a handheld 

radio is used to detect the recovery mode beacon (RS232 is not in yet, so nominal mode 

cannot be achieved). The beacon comes up fine and sounds “high”, as it should. 

The PCU-EPS RS232 cable is committed, but cannot test it as the battery is too low to leave 

safe mode after boot-up. Therefore the spacecraft is left charging for a while. 

The MAGIC high power and shunt resistor cables are committed. 

364






Once the battery voltage is high enough the spacecraft is booted up, as is the groundstation. 

The nominal mode beacon contains EPS data, this implies that the RS232 cable is working 

fine. 

The charging battery levels out at 24.62V, and is only consuming about 90mA (instead of the 

500mA it consumes at lower charge depth). 

Powering up the high power half of magic results in the expected current spike, then nothing. 

ESA_Neil commits the T-Pod cables from the PCU to the FPP. They buzz through ok, and 

the cabling from the FPP to the pods themselves is checked also. A triple-headed T-Pod test 

cable is made to connect the EM e-box to any of the three T-Pod ports on the PCU (via the 

FPP). They are tested in turn and all work fine. 

Pin 

Destination 

1 PCU, T-Pod 1, pin 1 


3 +X T-Pod, pin 1 


5 PCU T-Pod 2, pin 1 


7 +Y T-Pod, pin 1 




11 -X T-Pod, pin 1 


13 Blank 



365






The TCS harness is committed to the OBC. 











PROBLEM 232: Once the thermistors are plugged in the OBC will not boot up. 

With the thermistors unplugged the OBC comes up fine. 

ESA_Neil powers the thermistors form an external supply and is surprised to find that they 

consume 2.8 amps. OBC_Karl checks the datasheet and ESA_Neil checks the cables: it turns 

out that they were labelled, and therefore wired, with the power pins inverted. We have 

therefore fried them 

It might be possible to cut the wires by the thermistors, solder new ones on, and glue them 

down – perhaps also to remove the old ones. However, there is no time to do that during this 

“spacecraft open” session. 

The solar panel harness is connected to the panels using crimp connectors, heat shrink, and a 

strain relief loop for each. Then the lateral panels are laid out alongside the spacecraft, each 

resting on its side protector, which each rest on solar cell boxes to make them the right height. 

The S-Band patch antennas are connected with the new bolts, they fit fine, and so do the 

attenuation caps. 

366






ESA_Neil subjects one of the s-band attenuation caps to a vacuum, just out of interest. A few 

small bubbles appear and then burst – nothing serious. 

The microwave cables are integrated to the structure using the cable clamps and bolts. 

ESA_Neil commits the OBC-ACDS RS232 and MGM harness, copying the ACDS test 

harness supplied by ACDS_Lars. The “remote” ends of the ACDS MUX, coil, MGM power 

and MGM ground are committed and the cables routed, along with the +y sun sensor cables 

and the +y solar panel harness. 

PROBLEM 233: The +y sun sensor harness is not long enough. 

367







ESA_Neil integrates the shunt resistor to the +y lateral pnel using M5x12mm bolts and M5 

nuts and washers. 

ESA_Bas starts work soldering the activation switches. While ESA_Neil glues some more 

harness clamps, and glues the s-band patch antenna connectors to the back shields for 

strength. 

MODIFICATION 129: ESA_Bas extends the +y sun sensor cables. 

ESA_Neil wires up the solar panel connector on the PCU: 

Panel connector pin 

Destination description 

1 FPP external power connector pin 1, for timer reset 

6 Positive side of –Y solar panels 

7 Positive side of –Y solar panels 

8 Positive side of +X solar panels 

368






9 Positive side of +X solar panels 

10 Positive side of +Y solar panels 

11 Positive side of +Y solar panels 

12 Positive side of –X solar panels 

13 Positive side of –X solar panels 

14 Blank 

15 Negative side of –Y solar panels 

16 Negative side of –Y solar panels 

17 Negative side of +X solar panels 

18 Negative side of +X solar panels 

19 Negative side of +Y solar panels 

20 Negative side of +Y solar panels 

21 Negative side of –X solar panels 

22 Negative side of –X solar panels 

23 FPP external power connector pin 13, power feeding 

24 FPP external power connector pin 14, after timers 

25 FPP external power connector pin 15, before timers 

The positive side of the solar panels are located on the EPS safe / arm connector on the FPP 

as pins 17-24. 

ESA_Bas completes the activation switches, ESA_Neil buzzes them through while toggling 

them, they seem perfect. 

PROBLEM 234: ESA_Jason’s glove melts onto the washer he is applying while bolting in 

the activation switch plate. This is because he accidentally connected the charge stud with the 

ground plate, shorting across the battery terminals. 

The spacecraft won’t power us as the battery is locked by the protection circuit. The battery 

voltage reads as 7 volts or so. However, charging the battery opens the protection again, and 

everything returns to normal. 

369






ESA_Neil commits the ACDS-OBC harness, copying the test harness, as follows (digital coil 

connector on OBC): OBC pin 5 goes to ACDS pin 1, OBC pin 1 goes to ACDS pin 4, OBC 

pin 4 goes to ACDS pin 5, OBC pin 2 goes to ACDS pin 2. The MUX cables is committed, it 

is simply straight. 

The external OBC TCS pins are located as pins 4 and 5 of the MUX analogue connector on 

OBC. The thermistor voltage signal wires are committed to them. 

The 25-pin internal ACDS connector is committed, using the labelled wires reaching it from 

all far corners of the spacecraft. 

The spacecraft is powered up, watching the current consumption through the ACDS system. 

It fluctuates initially, between about 40 and 120mA, for about 20 seconds, then it settles down 

at 44mA. The data received in the nominal mode beacon seems sensible (field strength static 

and field rate zero). 

ESA_Neil commits the power harness to ACDS, and attaches the microwave cables to the 

lateral patch antennas - they are tight, but they just fit. 

ESA_Neil removes the attenuator in the UHF system and then, with ESA_Iñaki, the lateral 

panels are integrated in the standard sequence. 

The spacecraft is powered up and an attempt to establish two-way communication is made. 

PROBLEM 235: The spacecraft still cannot transmit properly, and there is the “beep” 

detectable on a handheld radio when it tries to, just as in problem 226. This implies that the 

lateral panels do not add enough attenuation to fix the problem. 

370






TEST: Placing a large metal object next to the antenna alters the pitch of the audio tone 

heard on the handhelds, and proper placement resolves the problem entirely. 

TEST: The +z T-Pod door is opened to see if its presence is enough to fix the problem. It is 

not, the problem persists. 


The battery is put on to charge. 

ESA_Jason and ESA_Neil prepare a plastic bag for transporting the spacecraft outside of the 

cleanroom. 

ESA_Neil prepares a ground support trolley, containing all of the charging equipment and the 

test groundstation. 

371






The spacecraft is lifted (by hand, since we don’t have a small enough crane) onto a trolley 

beside the integration table. It is then covered by the bag and carefully transported to the 

EMC facilities on a trolley with soft air wheels. 

The spacecraft is installed in the anechoic chamber, the remove before flight items are 

removed, and the launch configuration and flight configuration frequency sweeps are 

performed. 

13:30 Arrival of specimen in facility 

Installation of SSETI on copper table. 

15:00 Plot 1 Narrowband emission vertical polarisation 10 kHz to 1 GHz unit OFF. 

Antenna on (–X) position 

15:30 Plot 2 Broadband emission vertical polarisation 10 kHz to 1 GHz unit OFF. 


Plot 3 Broadband emission vertical polarisation 10 kHz to 1 GHz unit OFF. 

Antenna on (–X) position recalculated to 1 MHz bandwidth 

16:00 Plot 4 Narrowband emission horizontal polarisation 30 MHz to 1 GHz unit OFF. 

372







16:15 Plot 5 Broadband emission vertical polarisation 30 MHz to 1 GHz unit OFF. 


Plot 6 Broadband emission vertical polarisation 30 MHz to 1 GHz unit OFF. 


16:25 Plot 7 Emission vertical polarisation 1 GHz to 15 GHz unit OFF. 


Plot 8 Broadband emission vertical polarisation 1 GHz to 15 GHz unit OFF. 


16:43 Plot 9 Emission horizontal polarisation 1 GHz to 15 GHz unit OFF. 


Plot 10 Broadband emission horizontall polarisation 1 GHz to 15 GHz unit OFF. 


17:10 Plot 11 Narrowband emission vertical polarisation 10 kHz to 1 GHz unit ON. In nominal 

mode Antenna on (–X) position 

17:27 Plot 12 Narrowband emission horizontal polarisation 30 MHz to 1 GHz unit ON. In nominal 


17:45 Plot 13 Narrowband emission horizontal polarisation 1 GHz to 5 GHz unit ON. In nominal 


17:49 Plot 14 Narrowband emission vertical polarisation 1 GHz to 5 GHz unit ON. In nominal 


Plot 15 Narrowband emission vertical polarisation 1 GHz to 5 GHz unit ON. In nominal 

mode Antenna on (–X) position Zoom of the measured 2.3 GHz signal. 

17:55 Plot 16 Narrowband emission vertical polarisation 1 GHz to 5 GHz unit ON. In nominal 

mode Antenna on (+X) position 

18:05 Plot 17 Narrowband emission horizontal polarisation 1 GHz to 5 GHz unit ON. In nominal 


18:12 Plot 18 Narrowband emission vertical polarisation 10 kHz to 1 GHz unit ON. In nominal 


18:25 Plot 19 Narrowband emission horizontal polarisation 30 MHz to 1 GHz unit ON. In nominal 

mode 

Antenna on (+X) position 

373






This involves temporarily removing Xi-III from the upper T-Pod, and leaving the door open. 

TEST: The test groundstation and handheld radio are used to see if the external environment 

(anechoic chamber instead of cleanroom) fixes problem 235, but it does not. 


The EMC testing continues with sweeps across the centre frequencies for each antenna, with 

and without attenuation caps, and with and without modulation. 

09:30 Plot 20 S-band transmission in +X location vertical polarisation NO modulation applied. 

10:00 Plot 21 S-band transmission in +X location vertical polarisation WITH modulation. 

10:10 Plot 22 S-band transmission in +X location horizontal polarisation NO modulation applied. 

10:15 Plot 23 S-band transmission in +X location horizontal polarisation WITH modulation. 

Due to the circular polarisation no difference is measured between vertical and horizontal 

polarised receiving antenna. All other S-band measurements will be done using vertical 

polarised receiving antenna. 

10:20 Plot 24 S-band transmission in -X location vertical polarisation NO modulation applied. 

10:27 Plot 25 S-band transmission in -X location vertical polarisation WITH modulation. 

10:40 Plot 26 S-band transmission in +Z location vertical polarisation NO modulation applied. 

10:47 Plot 27 S-band transmission in +Z location vertical polarisation WITH modulation. 

10:50 Plot 28 S-band transmission in +Z location vertical polarisation NO modulation applied with 

anechoic caps covering the S-band antennas. 

374






11:00 Plot 29 S-band transmission in -X location vertical polarisation NO modulation applied with 


11:07 Plot 30 S-band transmission in +X location vertical polarisation NO modulation applied with 


11:25 Plot 31 UHF-band transmission in -X location vertical polarisation Carrier ON with 

Modulation. with anechoic caps covering the S-band antennas. 

11:35 Plot 32 UHF-band transmission in -X location Horizontal polarisation Carrier ON with 

Modulation. with anechoic caps covering the S-band antennas. 

12:00 Test completed 

The results of the EMC tests can be found on the FTP in the AIV folder in the file 

Express_E_ESA_EMC_results.pdf. 

In conclusion of the EMC tests: 

- The SSTL launcher requirements have been satisfied 

- All antennas are transmitting as expected 

- The upper s-band antenna is the required 3dB above the other two 

- The s-band attenuation caps perform well 

ESA_Neil reintegrates Xi-III and the remove before flight items, then the spacecraft is 

returned to the cleanroom and lifted back on to the integration table. 

The spacecraft weighs in at 58kg. However, the scales used were old and not calibrated, so 

we cannot be sure. 

375






The –y, +x and +y panels are removed and laid next to the spacecraft in order to allow for 

troubleshooting of problem 235. 

Removing the S-Band-UHF audio cable does not alter the problem, neither does the removal 

of the PTT line, or the RS232 connection. The power cable is also lengthened, to no avail. 

The flight coax cable is removed from the antenna and a saver is added. Then another 

antenna (telescopic BNC) is applied. Although this does not remove the problem altogether, 

it allows us to realise that the problem only occurs when the antenna is grounded to the topplate 

directly – implying that we have a ground loop problem. 

This works: 

This does not: 

NOTE: During these tests the strange audio tone is listened for an a hand-held radio. Actual 

data transmissions are not possible since the EMC testing has run the battery down past the 

safe mode exit threshold, so the spacecraft will not enter nominal mode until it is charged. 

376






ESA_Neil attempts to integrate the credits plate. 

PROBLEM 236: The plate does not fit properly because of a harness clamp being in the 

way. 

MODIFICATION 130: ESA_Neil dremmels a cut-out in the credits plate to fit around the 

harness clamp. 

The credits plate is then integrated, and the spacecraft is left charging overnight. 

377







ARRIVAL 80: Filter / capacitor connectors and savers arrive from AMS_Howard for the 

purposes of troubleshooting the RF problems. 

ESA_Neil performs the following tests, with the following results, evaluating the responses 

by listening for the tell-tale audio tone using a hand-held radio, and watching for a valid data 

connection to the groundstation: 

NOTE: The test begin when using a BNC telescopic antenna 

Test 

Add capacitor / filter saver (RS 238 7343) to 

the power connector on UHF 

Add capacitor / filter saver to the RS232 

Add capacitor / filter saver to the PTT line 

Replace power cable with a filtered 15-way 

connector (RS 449770) 

Isolate antenna from the top plate using 

kapton tape 

Remove filtered power cable and replace with 

original 

Remove capacitor filter from RS232 

Remove capacitor filter from PTT line 

See if the outer of the SMA connector on the 

antenna is connected to the spacecraft ground 

when the antenna is not connected to the 

Result 

Audio tone present and no data received at 

groundstation 


groundstation 


groundstation 


groundstation 

No audio tone present and data received at 

groundstation 


groundstation 


groundstation 


groundstation 

It is grounded fine 

378






structure 

The resulting temptation is simply to isolate the foot of the flight antenna from the top-plate 

using kapton tape and PTFE bolts. In order to test this should the plan be approved by those 

“in the know” some temporary nylon bolts are procured from the main workshop. 

There is a large time pressure to solve this problem, since the spacecraft should be closed on 

Sunday with all bolts torqued and glued ready to go to the vibration facilities on Monday 

(today is Friday). However, UHF_Holger is not contactable. 


ESA_Neil reproduces problem 235 to AMS_Howard. 

AMS_Howard clamps pieces of ferrite onto the cables as they enter the UHF box, but this 

does not help. 

Throughout the work today a saver is placed on the end of the UHF power line from the PIN, 

and a spare cable on the UHF power input, so that power can be toggled without cycling 

connectors or cycling the spacecraft. Savers are also placed on the antenna end of the coax 

and on the antenna itself. 

We run the UHF output through a power meter. The power output is fine when the antenna is 

isolated and the transmission is stable, but the drops to around one third of the nominal value 

when the antenna is connected to the top-plate and the problem manifests itself. This implies 

that the PA is folding back on itself and reducing the power so as not to be damaged. 

379






AMS_Howard attempts to cause EMI problems by connecting a dipole antenna and waving it 

around the internal cabling. It is not possible to cause interference. However, if the dipole 

grounded braid touches the spaceframe then the audio beep occurs and the power drops. This 

is found to be path length dependant: repeatedly, at specific locations of grounding of the 

dipole braid, the system responds normally, but at others the problem occurs. This certainly 

implies a ground loop problem. 

PROBLEM 236: Around 50% of the time when the UHF unit is power toggled manually it 

comes up already transmitting at low power on an unknown frequency, it is then not possible 

to receive or transmit normally. 

TEST: Working on the assumption that this problem could be due to a short duration bad 

connection while the plug is cycled, or the connection of the positive line before the ground 

line (which would never happen in normal operations), ESA_Neil uses the debugging line on 

the FPP to internally command a series of power cycles on the UHF unit. It comes up 

correctly 10 times out of 10, implying that the problem is one that will only occur during 

manual power cycling in testing, and not in nominal operations. 

ESA_Neil applies several layers of kapton tape to the foot of the antenna, and the mounting 

insert. The antenna is then integrated using nylon (temporary) bolts to secure it, savers are 

used on the antenna and the coax, as before. 

380






The spacecraft is powered up and the antenna seems to work fine, the power meter reading is 

good. The return loss is about -10dB instead of -12dB as it used to be, and the SWR is now 

about 2:1. (It is slightly better with lid of upper t-pod shut, about 1.7:1.) 

It seems very stable and it is hardly possible to force it to go wrong. (Have to ground the tip 

of the antenna.) 

So, the SWR meter and the savers are removed. 

PROBLEM 237: Problem 235 reoccurs. 

The length of the coax seems to matter, so we must find the appropriate length. We add 

savers to find a length that works. 

Then we add about 1/8 lambda and this causes the problem to reoccur, even with the savers 

attenuating significantly. This is a proof positive for the current theory. 

381






Tests are performed changing the electrical length of the “patch” (in addition to the flight 

cable) by the addition and subtraction of savers and lengths of coax. The electrical length is 

measured and/or estimated using a network analyser. 

Electrical length 

Does it work? 

0 ? 

3 No 

6 No 

9 No 

12 Yes 

15 Yes 

17 Yes 

20 Yes 

23 No 

26 No 

29 No 

32 No 

35 No 

38.5 Yes 

41.5 Yes 

44.5 Yes 

47.5 Yes 

50.5 Yes 

53.5 No 

382






We put in the flight cable and test it with one saver, it works fine. We therefore start to add 

savers to find out when it goes wrong – it continues to operate until three savers are added. 

This implies that we are on “the cusp”, but it should be ok with no savers. 

The last savers are removed and the flight cable tested alone. It does not work. 

A piece of semi-rigid cable that worked during testing is integrated, and it works fine with no 

savers. It is heat shrinked to protect the copper, but this is not really adequate and it will need 

to be changed before it can fly. At least we have a specimen of the correct electrical length 

though. 

All cables are recommitted and tested, they work fine. The antenna is grounded using one 

metal bolt instead of a nylon one, and this breaks the downlink again, as expected. The 

isolation is necessary. 

The spacecraft is shut down and charged. 

AMS_Howard departs. 


ESA_Neil demonstrates the spacecraft for STRU_Antonio. Then we plan the torquing and 

gluing of all the internal bolts. 

We start with the –x0y compartment and work around the spacecraft in an anticlockwise 

direction. The standard torquing pattern is employed (4,2,6,3,5,1). All screw locks are also 

383






torqued (or simply tightened depending upon access) and glued. For those screw locks whose 

female half rotates with the tightening, both halves are glued. 

M4 washers are added to the PCU to distribute the loads properly. 

The passive magnet is glued into its housing since it is slightly loose and could be damaged in 

the vibrations. 

M4 washers are added to the battery for the same reason. 

Too be torqued and glued: UHF, Credits Plate, Lateral panels, UHF antenna and coax. 

(These are not done today since the UHF coax cable needs to be replaced and this necessitates 

disintegration.) 

MODIFICATION 130: In order to provide better access to the charging port on Xi-V, the 

+y lateral panel has a L shape dremmeled from it. The edges are tidied up with a sander and 

then protected with aluminium tape. 

384






PROP_Sascha prepares the pyro connector, glues it, and integrates it. 

All bolts on the lateral panels that do not have interfaces to the primary structure are torqued 

and glued. 

The lateral panels are placed, but not torqued, since accelerometers will need to be added 

tomorrow. 


ESA_Neil closes the +y t-pod, then PROP_Sascha and STRU_Melro help to move the 

spacecraft onto the transport trolley. 

The following tools are taken with the spacecraft, the groundstation, and the charging station, 

to the vibration test floor: 

m4 driver 

m3 driver 

385






torque wrench 

flathead driver 

m8 allen key 

tie wrap gun 

tie wraps 

scalpel 

kapton 

clippers 

bags for clips 

abf items 

attenuation cap box 

m8 bolts 

groundstation 

charging station 

wrench bits (m4, m5, m8) 

size 10 spanner 

spacecraft - with side protectors, attenuation caps, lifting frame and bag 

We weigh the test adapter, it is 29.8 kg. Then the spacecraft is lifted with a crane onto the 

scales on top of the adapter. The integration mass of the spacecraft, including all RBF items 

and the test adapter, is 97.35kg. 

The side protectors are removed, giving a total mass of 89.1kg. The lifting frame is removed, 

giving a resulting mass of 87.5kg. The attenuation caps are removed, giving a final total mass 

of 86.8kg. 

The launch mass of the spacecraft is therefore 57.0kg, not including fuel. 

386






The lifting frame is replaced and the spacecraft is lifted off of the test adapter. The adapter is 

moved to the centre of mass machine and bolted in place. 

The adapter is carefully aligned with the machine, and the resulting error is a measurement of 

0001, which is equivalent to 0.02kgm, or to 0.6667mm in the +X of the table. 

The spacecraft is lifted to the centre of mass machine and bolted to the adapter. 

A 57kg spacecraft with a maximum distance from the geometric centre to the centre of mass 

of 10mm, should give a maximum turning moment of 0.57kgm = 0028.5 digits. 

The x-axis measurement is taken: -24 (with correction) = -8.42mm v (expected -8.38mm) 

This is within 40 microns of the prediction, and within 2.58mm of the maximum. 

The spacecraft is lifted and turned by 90 degrees to make the y-axis measurement. 

Y-axis = -22 (with correction) = -7.72mm (expected -3.1mm). This is over 4mm off from the 

prediction, but is still 3.28mm within the maximum, so there is no problem. This discrepancy 

can be the result of the highly non-homogenous harness, and the extra glue and conformal 

coating applied to certain units on the –y side of the spacecraft. 

The spacecraft is lifted to an `x` of tables in order to remove the lateral panels. The panels are 

laid on the solar cell boxes and side protectors as usual. 

387






ESA_Neil removes the UHF coax cable and takes it to ESA_Ron to measure the exact 

electrical length, it is found to be 110cm. ESA_Bas then manufactures a new cable of the 

same length. 

ESA_Neil integrates the new cable and tests it, it works fine and problem 235 is finally 

solved. 

STRU_Antonio and PROP_Sascha, under the guidance of the ETS testing team, apply all the 

accelerometers apart from those on the lateral panels. 

388






389






STRU_Melro torques all the UHF screw locks and mounting bolts. ESA_Neil glues the bolts. 

The sides of the spacecraft are closed, and STRU_Melro torques and glues all the bolts 

between the primary structure and the lateral panels. 

390






MILESTONE 27: The integration is completed. 


PROP_Sascha, STRU_Melro and ESA_Neil have a safety meeting with the ETS team to 

discuss the hazards presented by the pyro valve, the high pressure system, and the lithium ion 

batteries. 

The external accelerometers are applied. 

391






The three magnetorquer tie wraps are applied. NOTE: Applying these tie wraps is really 

hard, especially the top one. The three together took over an hour. 

One of the 300 bar gas bottles loaned from Air Liquide is delivered from the bottle storage to 

the test floor. 

The spacecraft is lifted and then moved to the vibration table and installed in the longditudinal 

direction. 

392






PROBLEM 238: PROP does not have a hose that is proof pressure tested to 450 bar, so 

there can be no connection from the ground support equipment to the spacecraft. 

SOLUTION: Instead PROP_Sascha uses the backup filling solution (a direct hose from the 

bottle to the spacecraft). During the filling the spacecraft and the groundstation are powered 

up, and a live stream of the housekeeping stack from OBC is used to view pressure and 

temperature values regularly requested from Magic by telecommand. 

PROBLEM 239: PROP_Sascha was not expecting such a lengthy decoding process from the 

MAGIC telemetry to actual data. The conversion done manually at each reading would 

necessitate an unacceptably long filling procedure. 

393






SOLTUION: We make an Excel spreadsheet to use as a lookup table to decode the 

temperatures and pressures quickly. 

PROBLEM 240: The high pressure system reads at 213 bar, but the gas bottle is fully open 

so we should expect around 267 bar. Also, the temperature is not climbing as high as 

expected. 

TEST: The bottle is disconnected and reconnected. The pressure rises to 256 bar almost 

immediately. This is good, but not understood, PROP_Sascha is thinking about it. 

TEST: The bottle is disconnected and hooked up to one half of the ground support 

equipment. The GSE reads the pressure in the bottle at 265 bar. This is just about as 

expected, and it must be identical to the pressure inside the tank, since the valves were fully 

open. Therefore the 9 bar discrepancy between the value measured directly and the value 

reported by MAGIC must be due to calibration error (approximately 3%). 

Pressurisation complete, propulsion GSE is disconnected. The pressure reported by MAGIC 

drops by 1.8 bar in 10 mins. This is probably just the fuel cooling down. 

Pre-vibration checklist: 

Side protectors are absent x4 

Lifting frame is absent x1 

Attenuation caps are absent x3 

Thruster caps do not exist x4 

Remove the T-pod safe bolts x3 

Remove the lens cap x1 

Remove the sun sensor covers x2 

The lower ASAP M8 bolts are torqued to 15Nm x12 

The pilot accelerometers are added. 

MILESTONE 28: The spacecraft is ready for vibration testing. 

For details of the vibration loads applied to the spacecraft, refer to the latest 

SSETI_Express_Test_Plan on the FTP server in the AIV folder. 

The entire longitudinal axis (Z) will be vibrated first. 

1615: Tap tests are performed to check all accelerometer responses. 

1635: The first low-level sine sweep is performed, run 1Z, 4 minutes 20 seconds. 

394






The results of the low-level sine sweep are presented to us by the ETS team, and analysed by 

STRU_Antonio. 

Antonio says it is ok to go on with the sine and will think about random vibes overnight 

1745: Sinusoidal vibrations 0-100Hz, run 2Z, 1 minute 5 seconds. 

The results are analysed and seem reasonable. A visual inspection is carried out. 

PROBLEM 241: One of the kill switches from UWE-1 is found laying next to the +x 

thrusters cluster. It must have detached from the spacecraft and fallen through the access hole 

in the T-Pod. 

NOTE: There is a crack in the glue on the +x side of the bracket in the +x compartment, but 

this was already present, as confirmed by looking through previous pictures of the 

compartment. 

We perform a simple functional checkout: 

395






- power up 

- safe mode beacon detected on handheld, implies PCU and UHF are ok 

- after 12 minutes the PCU assumes that the battery is broken and enters recovery mode 

- nominal mode beacon detected, implies OBC is ok 

- s-band commanded carrier up, implies s-band is ok 

- cam command on and pinged, pong received, implies cam is ok 

- magic powered up and CAN data received, implies magic ok 

- pressure transducer read out via MAGIC, OBC, UHF, g/s, is now at 244bar, implies PROP 

ok (10 bar drop in 5 hours can be temperature stabilising, a leak would be much worse) 

- magnetometer data is ok 

- S-BAND data received ok implies TNC is fine 


The battery has finished charging overnight and the spacecraft is powered up. The pressure is 

checked and found to be 244 bar still. The spacecraft is powered down. 

0930: Low sine sweep, run 3Z, 4 minutes 20 seconds. 

Comparison analysis (Z1 and Z3), looks ok. Some mechanical settling and increased 

damping. 

1115: Random vibrations, run 4Z. 

PROBLEM 242: The NCube-2 antennas and gravity boom have deployed through the 

access port into SSETI Express. Their RBF pin is still present 

PROBLEM 243: An accelerometer detached from the +Z antenna during the vibration, and 

has slightly dented the surface. This damage is probably superficial, but must be checked 

with COMMPL. 

396






PROBLEM 244: One of the +x T-Pod cable ties has broken and the Teflon block is loose in 

the base of the pod, also, one of the pads on the lid has detached. 

PROBLEM 245: One of the -x T-Pod cable ties has broken and the Teflon block is loose in 

the base of the pod. 

PROBLEM 246: The fill and drain valve cap has fallen from the PMS box. It had not been 

torqued properly. 

NCube_Åge removes NCube-2 from the -x T-Pod. 

ESA_Neil replaces NCube-2 with a mass dummy for the low level sine sweep. 

1417: Low level sine sweep, run 5Z. NOTE: The noise level here is significantly higher 

than the 3Z sweep. 

A teleconference is organised with the NCube team, CANX_Fred, SYS_Joerg, 

STRU_Antonio and ESA_Neil. The following conclusions are reached: 

397






- The best way to assess the consequences of the minor T-Pod failure are to compare the 3Z 

and 5Z runs on the +x T-Pod, since no change was made here between the runs. 

- NCube-2 are to assess data and report whether or not they were mistreated by the T-Pod 

- CANX_Fred to asses whether or not the t-pod mistreated NCube-2 

- STRU_Antonio to assess whether or not the T-pod mistreated NCube-2 

- The SSETI Express shake to continue with mass dummy instead of NCube-2 

- Minor modifications are possible tonight to strengthen t-pod cable ties 

- The teleconference reconvenes tomorrow at 2000 

NOTE: The third and fifth runs (low level sweeps), there are significant differences, which 

could be explained by the loose Teflon blocks in the lateral pods. 

1615: Quasi-static loads, run 6Z 

We perform a simple functional checkout: 

- power up 


- after 12 minutes the PCU assumes that the battery is broken and enters recovery mode 


- s-band commanded carrier up, implies s-band is ok 

- cam command on and pinged, pong received, implies cam is ok 

- magic powered up and CAN data received, implies magic ok 

- pressure transducer read out via MAGIC, OBC, UHF, g/s, is still at 244bar 



1625: Low-level sine sweep, run 7Z. NOTE: The noise level here is significantly higher 

than the 3Z sweep, but similar to the 5Z one. 

The lifting frame is applied and the spacecraft is moved to the table, the adapter is moved to 

the slip table, and then the spacecraft is placed on top of it. 

398






We open all three pods and remove all three cubes. 

PROBLEM 247: Xi-III has loose objects of significant mass inside it, it rattles loudly. It 

was probably this causing most of the noise in the last two low level sweeps. 

The following modifications are discussed over the phone with CANX_Fred and then 

implemented by PROP_Sascha, STRU_Antonio and ESA_Neil. They are considered minor 

enough to not invalidate the previous shakes. 

MODIFICATION 131: The +x t-pod is modified thusly: Teflon block roughed up with a 

screwdriver, cable ties cut to 80% width by hand with scalpel, two ties inserted and block 

glued to the back of the pod with AY138. Spring reattached and rails cleaned, cable ties 

tightened and UWE-1 (minus one kill switch) is re-loaded. 

399






MODIFICATION 132: The -x t-pod is modified thusly: Teflon block roughed up with a 

screwdriver, cable ties cut to 80% width by hand with scalpel, only one tie wrap inserted since 

one hole in the block was too narrow. Block glued to the back of pod with AY138. Spring 

reattached and rails cleaned, cable tie tightened and mass dummy (pretending to be ncube-2) 

is re-loaded. The detached pad is also re-glued. 

MODIFICATION 133: The +y t-pod is modified thusly: ay138 applied to bases of centres 

of sides of Teflon block to try and secure it in position on the back of the pod. Rails and push 

plate cleaned for easy sliding. 

PROBLEM 248: We cannot reinsert Xi-III as it is clearly damaged, but do not have any 

more mass dummies to replace it with. 


UWE-1 is loaded into the +x T-Pod. 

A mass dummy is loaded into the +Y T-Pod. 

The antenna cases are taped closed on NCube-2 and it is loaded, upside down (to protect 

SSETI Express from potential antenna deployment), into the –x T-Pod. 

400






The appropriate accelerometers are replaced on the lids of the T-Pods. 

0900: Low level sine sweep - ABORTED 

PROBLEM 249: There is a lot of noise during the low level sine sweep, it seems to be the 

mass dummy in the +y T-Pod. This could be throwing out the control curves, and could 

potentially add increased loads to the local shear panel. 

The low level sine sweep is aborted by ESA_Neil using the emergency stop button. 

The mass dummy is removed from the +y T-Pod and kapton tape is applied along the long 

edges and in layers over the top of the rails to ensure a tight fit. This then renders it a more 

accurate mass dummy for Xi-III. 

1100: Low level sine sweep, run 1X 

The sine sweep is much quieter, implying that the problem was indeed the mass dummy no 

being an accurate representation of Xi-III. 

401






PROBLEM 250: Low sine sweep reveals a resonance at 44Hz (expected more like 50Hz). 

Using Miles formula (equivalent static load from a random test) we would get 21g at the first 

resonance. This would almost certainly cause damage to the spacecraft. 

1130: Sinusoidal vibrations, run 2X 

Functional check: PCU ok, UHF ok, OBC does not come up. This could be due to battery 

voltage, but functional checkout abandoned for time reasons and will be completed after the 

next low level sine sweep. 


Functional check, wait for PCU to assume the battery is broken. Everything comes up fine. 


- after 12 minutes PCU assumes BATT broken and enters recovery mode 


- S-Band commanded carrier up, implies S-Band is ok 

- CAM command on and pinged, pong received, implies CAM is ok 

- MAGIC powered up and CAN data received, implies MAGIC is ok 

- pressure transducer read out via MAGIC, OBC, UHF, g/s, is now at 243bar 



1350: Quasi-static loads, run 4X 


All x axis low-level sine sweeps overlay each other almost perfectly. 

We attempt to discuss notching possibilities with SSTL, but are told that SSTL cannot make 

that decision, instead it should be Polyot. Ultimately SSTL do not respond fast enough. 

Therefore we decide to stop wasting time and turn the spacecraft to vibrate on the y axis for 

the sinusoidal and quasi-static vibrations. We will come back to the random vibrations later. 

1615: Low level sine sweep, y axis, run 1Y 

1630: Sinusoidal vibrations, run 2Y 

1700: Low level sine sweep, y axis, run 3Y 

Functional check, wait for PCU to assume the battery is broken. Everything comes up fine. 


- after 12 minutes PCU assumes BATT broken and enters recovery mode 

402













1725: Quasi-static loads, run 4Y 

1740: Low level sine sweep, run 5Y 

All y axis low-level sine sweeps overlay each other almost perfectly. 

SSTL report via telephone that there were mistakes in the recommendation document and that 

the protoflight levels should be at 1.9dB above the launch levels, not at 3.5dB. They also 

report that there should not be a problem notching to the launch levels around the first 

resonance. 

The SSETI and ETS teams discuss and the next run is prepared using the new data from 

SSTL. 

SSTL changes their minds and informs us that we are now “required” to vibrate at 3.5dB 

above the launch levels, and that we can only notch to 1.5 times the launch PSD at the first 

resonance. The relevance of, and obligation to fulfil, this very late requirement is in 

contention. 


SSTL are not available for a teleconference in time for our schedule. We therefore decide to 

run a random vibration with the profile they now “require” of us (for the first time as of 

yesterday), but only at 50% levels so that we can analyse the results and decide if we want to 

notch further down. This profile is at 3.5dB above the launch loads, but with a notch down to 

1.5 times the launch PSD for a 10kHz band across the first resonance. 

1040: Random vibration, run 6Y, 50% protoflight levels, with notch down to 50% acceptance 

levels over 10Hz around the 38Hz resonance. 

MODIFICATION 134: In order to protect the spacecraft it is decided to notch the input 

loads to the flight levels across the first resonance. This is not in accordance with the recent 

SSTL “regulations”, but that will be negotiated later. 

403






1200: Random vibration, run 7Y, 100% protoflight levels, with notch down to flight levels 

over 10Hz around the 38Hz resonance. 

PROBLEM 251: During the random vibration a bolt unscrews itself and falls from the 

spacecraft. It is the upper-most (+z) bolt connecting the +y lateral panel to the +y+x small 

shear panel. This is not serious. 

PROBLEM 252: Bolt a bit loose in the –y lateral panel / baseplate interface, below the +x–y 


Functional checkout: 









Analysis of the last two low level sine sweeps: POTENTIAL PROBLEM: It appears that 

the fundamental bending mode resonant frequency has decreased and the response level has 

dampened. The change is significant and implies that there is more “play” than there was 

before. It could be that the entire potting radius of the inserts for the battery have shifted 

slightly in the honeycomb, as in problem 84. Alternatively there could be some extensive 

mechanical settling, such as a loosening of the coupling between the shear panels and the base 

brackets, but this is unlikely. The only way to find out is to open the spacecraft, but we will 

not do that until the vibration tests are complete. 

PROBLEM 253: During run 7Y the fundamental resonance shifted lower and ended up 

halfway out of the notch. This is good in that the levels are more acceptable for SSTL, but 

404






bad in case some damage was done to the spacecraft. In the subsequent low-level sine sweep 

the fundamental rose slightly again, which is a good sign. 

The spacecraft is lifted slightly and turned back to the x-axis for the remaining random 

vibration test. All bolts are torqued on the ASAP ring, and RBF items are removed (lifting 

frame and lens cap). 


1510: Random vibration, run 7X, 50% protoflight levels, with notch down to 50% launch 

levels over 10Hz around the 38Hz resonance. 

STRU_Antonio advises that we are ok to continue by scaling up by 6dB. 

1545: Random vibration, run 8X, 100% protoflight levels, with notch down to launch levels 

over 10Hz around the 38Hz first resonance. 

NOTE: There were very high amplitude vibrations of the +y T-Pod and the +x+y shear panel 

in the +/-x direction. It was clearly bending from pivot points at the mid-height bracket on the 

+x side of +x+y compartment, and the –y side of the top of the panel where it interfaces with 

the top-bracket. 

PROBLEM 254: During the vibration a bolt fell out, second from the top (+z), on the +y 

lateral panel interface to the +x+y shear panel. This is just below the one that fell out last 

time (problem 253), and can be explained by the heavy vibrations of the +x+y shear panel and 

the +Y T-Pod. This is not serious but it should be well glued and torqued for the flight. 

PROBLEM 255: Bolt loose at the top of the +y lateral panel interface to the +x+y shear 

panel. This is the one that fell out last time and can be explained for the same reasons as 

problem 254. This is not serious but it should be well glued and torqued for the flight. 

PROBLEM 256: Bolt loose from the +y lateral to baseplate interface, just below the –x+y 


PROBLEM 257: The –x side of the +y T-Pod is scoured where it repeatedly impacted on the 

upper skin of the top-plate during the vibration. The +x side is also scoured at the +y corner 

for the same reason, but not at the –y corner, where the top bracket is holding it. 


405






Functional checkout: 









NOTE: The OBC boot attempts report as “1” in the nominal mode beacon, and a flash 

integrity check returns “0000”. Perhaps the flash chip has suffered damage. 

PROBLEM 258: The fundamental mode of the spacecraft on the x axis has divided into a 

38Hz resonance and a 32Hz resonance. 

The lifting frame is applied and the spacecraft is moved to the table. 

PROP_Sascha vents the high pressure system. 

406






NOTE: Some RTV is lose or missing from the harness exit point of solar panel 7. This 

should be re-glued, particularly underneath on of the solar cell connectors. There are a few 

other examples and the whole set of solar panels should be checked. 

The external accelerometers are removed from the spacecraft. 

The lateral panels are removed and laid down next to the spacecraft. A visual and tactile 

inspection proceeds 

NOTE: The final missing piece of original T-Pod cable tie is found straddling the –x-y 

lateral panel underneath the –y coil. 

PROBLEM 259: The upper bolt on the mid-height bracket on the +x side of the 0x+y 

compartment is very loose (about to fall out). 

PROBLEM 260: The bolt from the +x+y top bracket to the +x+y shear panel is very loose. 

NOTE: Problems 251, 254, 255, 259 and 260 combine to nicely explain problems 257 and 

258. With all of these bolts loose the +x+y shear panel will have been freely oscillating on 

407






the y axis from a pivot of the lower bolt of the mid-height bracket. This can be seen also by 

superficial damage to the +y lateral panel, where the internal surface paint was scratched off 

by the vibrations. Extra care should be taken to ensure that all of these bolts are torqued and 

glued securely for the launch. 

TEST: The spacecraft is powered up again to see if the OBC boot attempts increments. It 

does (twice), implying that the flash ram is fine. The flash CRC is regenerated and an 

integrity check performed, the report is positive. 

The internal accelerometers are removed from the spacecraft. 

ESA_Neil and STRU_Antonio replace the lateral panels (using only a handful of bolts), add 

the side protectors, lift the spacecraft onto a trolley, pack up all the equipment, and move 

everything back to the SSETI Express cleanroom. 

MILESTONE 29: The vibration tests are completed. 


PROP_Sascha tidies cleanroom. 

ESA_Neil charges the +x T-Pod and prepares for the functional checkout. 

NOTE: EPS functional checkout doc missing, OBC functional checkout document missing, 

CAM functional checkout document missing, 

408






PROBLEM 261: The +y T-Pod current consumption wavers around during charging and the 

battery will not rise about 10.55V. 

ESA_Neil charges the –x T-Pod and makes the T-Pod arm connector. 

PROBLEM 262: The test connector for the pyro does not match the current harness. After a 

phonecall to MAGIC_Renato it is confirmed that the MAGIC pinout does not match the 

PROP pinout. Use of a DVM to buzz through the connector confirms this. 

The spacecraft is booted up, a new version of term is being used with the new formulas for the 

EPS telemetry. 

- safe mode beacon detected 

- nominal mode beacon detected 

- EPS temp = 16,73 deg 

- OBC temp = 17.1 deg 

- EPS voltage = 24.2 V (oscillating +/- 0.1) 

- OBC boot attempts = 4 

ACDS checkout (in italics, answers in normal) 

1. Boot up the spacecraft 

Done 

2. Upon initialisation ACDS puts a message in the alarm stack. Download the alarms and 

write down the message. 

Get alarms: 

AX.25 sender: SSETI1 

Subsystem: ACDS | MID: 0x88 | Length: 21 | Time: 15-04-2005 18:18:05 

0x00000000: 41 43 44 53 20 4E 6F 72 6D 61 6C 20 4F 70 65 72 |ACDS Normal Oper| 

0x00000010: 61 74 69 6F 6E |ation...........| 

409






3. Keep an eye on the beacons, within a few minutes: 

- The magnetometer reading settles at a non-zero level 

- The ACDS "field rates" settles at zero. 

After a few minutes the following ACDS data is stable in the beacon: 


Subsystem: BEACON: 

Callsign: www.sseti.net * SSETI-Express 

On board time: 16-04-2005 14:44:14 

EPS battery voltage: 24.418 V 

EPS Temperature: 22.36 °C 

OBC temperature: 20.6 °C 

OBC boot attempts: 4 

ACDS Magnetometer: [-2642 1947 7521] nT 

ACDS field rate: [0 0 0] nT/s 

The above verifies that: 

- The ACDS driver board is powered 

- ACDS driver board generates 12V for the magnetometer 

- Communication between magnetometer and OBC is ok. 

- ACDS code on OBC executes nominally 

4. Issue the telecommand: ADS_MODE_SELECT with param1=1 and param2=10 

Live stream down UHF is turned on. 

ACDS TM data is every 18 seconds. 

Issue command. 

5. Wait 30s and downlink the telemetry stream. Verify: 

- The interval between ACDS data falls from 20 seconds before the telecommands to 6 

seconds after telecommand has been issued 

ACDS data is now every 6 seconds 


- The ACDS software react to telecommands (thus also a test of the 

complete UHF->OBC->ACDS chain) 

6. Find something magnetic and place it close to the spacecraft (alternatively then rotate the 

spacecraft around Z). Observe from the beacons that: 

- The magnetometer reading changes 

- The "field derivatives" increases for a time and then resettles at zero 

410






Spacecraft turned approximately 90 degrees during approximately 20 seconds. The following 

data is from successive beacons (18 seconds between each) 






ACDS field rate: [2 116 -2] nT/s 


ACDS field rate: [127 127 -6] nT/s 















The above further verified that the magnetometer is alive. 

7. Shade for the X+ sunsensor and downlink the telemetry stream. Copy one ACDS 

houskeeping packet (from when the sensor was shaded) to the open file and designate the 

packet X+1 

Um - the sunsensors are on the Y sides. Here is a shaded (by cover and tissue) +Y packet: 


Subsystem: ACDS | MID: 0x0A | Length: 32 | Time: 16-04-2005 15:03:41 

0x00000000: 26 FE C7 10 2A 1D F1 05 FF 05 10 00 18 00 02 00 |&...*...........| 

0x00000010: 02 00 0A 01 14 01 2C 02 00 00 00 00 00 00 00 00 |......,.........| 

411






8. Direct a spotlight at the X+ sensor (lamp ~3cm from the sensor). and downlink the 

telemetry stream. Copy one ACDS houskeeping packet (from when the sensor was shaded) to 

the open file and designate the packet X+2 

I assume you mean "when the sensor was lit", here is a lit (by halogen lamp) +Y packet: 



0x00000000: 29 FE CC 10 25 1D F7 05 02 06 13 00 23 00 13 00 |)...%.......#...| 

0x00000010: 00 00 0B 01 14 01 2C 02 00 00 00 00 00 00 00 00 |......,.........| 

+Y sensor light is removed and cover is replaced 

9. Repeat the above for the X- sensor, designate packets: X-1 and X-2 

Here is a shaded (by cover and tissue) -Y packet: 



0x00000000: 28 FE C8 10 27 1D F7 05 FC 05 1E 00 29 00 13 00 |(...'.......)...| 

0x00000010: 00 00 0B 01 14 01 26 02 00 00 00 00 00 00 00 00 |......&.........| 

Here is a lit (by halogen lamp) +y packet: 



0x00000000: 1A FE D1 10 3F 1D CC 06 28 06 16 00 29 00 7A 05 |....?...(...).z.| 

0x00000010: 2F 00 0B 01 14 01 3C 02 00 00 00 00 00 00 00 00 |/.....






Turn on CAM, ping pong is ok. 

Turn on magic, CAN data is received. 

Check pressure, about 0.5 bar (2C 00). Check temp, 17.8 deg, (5E). 

S-band checkout is not possible without the necessary equipment. 

PROP_Sascha and ESA_Neil try to identify which thrusters are which - is not possible 

without gas. 

PROBLEM 263: ACDS_Lars reports that +y sun-sensor is not working. Suspect harness 

problem. 

MODIFICATION 135: PROP_Sascha modifies harness on magic box pyro connector to 

correct problem 262. Moves pin 1 to 1, 2 to 2, 3 to 9, 4 to 10, 5 to 5, 6 to 6, 7 to 13, 8 to 12. 

LESSON LEARNED 21: Never close a box without checking that it is properly 

documented. 

LESSON LEARNED 22: Explicitly check all interfaces before committing hardware. 

The modification buzzes through ok. 

PROP_Sascha and ESA_Neil perform PROP functional checkout. PROP_Sascha fills the 

low pressure system while ESA_Neil mans the groundstation. 

Low-power MAGIC is turned on and CAN data received. 

High-power MAGIC is turned on. 

Thrusters are fired in various configurations to ascertain which one is which. 

Thruster 1 is -x-y 

Thruster 2 is -x+y 

Thruster 3 is +x+y 

Thruster 4 is +x-y 

NOTE: This is the inverse of the plan, therefore all positive manoeuvres become negative 

and vice versa. 

PROBLEM 264: Sampling of pressures and temperatures do not seem to be working 

properly during thruster firing, but this could just be operator error. 

413






The pyro test connector is re-wired to match the FPP, then a current probe is hooked up to an 

oscilloscope to record the pyro firing pulse. 

Initially no pulse can be recorded, but we eventually realise that the squibs are labelled 

incorrectly compared to the MAGIC pinout: the primary squib is known as squib number 2 in 

the MAGIC TC lists, and the secondary as squib number 1. 

Both test squibs (resistors) “fire” fine. 

PROP_Sascha vents the gas and moves the gas bottle into the airlock while ESA_Neil shuts 

down the spacecraft. 


ESA_Neil and Aage_NCube take out NCube-2 from the –x pod for analysis. 

ESA_Neil commences the CAM checkout: 

Spacecraft is under-charged so it takes 12 minutes to power up. 

The CAM is turned on and pinged ok. 

A housekeeping request gives 0x08 0x04 0x24 0x33 0x06 0x81 

PROBLEM 265: A faster sampling of the acceleration data is requested, and then the results 

are queried. Instead of getting housekeeping data an item is added into the alarm stack from 

OBC, MID 0x88, 17 bytes long “CAM: Chunk Error”. After this it is necessary to power 

cycle both the OBC and the CAM before the CAM responds again. 

414






The specified parameters are uploaded, a picture is taken (using auto-exposure), the 

thumbnails are transferred to the OBC, and then downlinked via s-band. This is repeated 

several times. 

PROBLEM 266: All downloaded pictures are black. 

PROBLEM 267: OBC_Karl reports that the fast sampling command responds with an MID 

that looks like a picture, which is why OBC gets confused. This could only be changed with 

an update of the OBC software. 

PROBLEM 268: It is impossible to upload parameters to the camera - the OBC software, 

rather stupidly, re-writes them all to zero. This means that it is impossible to take pictures. 

The only sensible way to fix this is to change the OBC software. 


ESA_Marie tightens the screw-locks on the MAGIC pyro connector (small hands are a big 

advantage here). 

ESA_Neil and ESA_Marie take off the –y lateral panel. 

The spacecraft is powered up and the voltages on the +y sun-sensor tested as per the request 

of ACDS_Lars. They all read zero. NOTE: Kapton is removed from the surface of the sunsensor 

as agreed with ACDS_Lars. 

415






We disconnect the coil driver 25-pin connector (adding savers, of course), and then buzz it 

through to the +y sun sensor. The following connections are identified (implying that 

ACDS_Lars numbered the pads upside-down). In the image above the pads on the right-hand 

side are numbered sequentially downwards starting with “1” at the top. 

Pin Pad 

12 6 

20 5 

22 3 

23 1 

24 4 

25 2 

This is reported to ACDS_Lars for analysis and a further procedure requested. 

ESA_Neil applied the s-band antenna caps, installs the latest S-Band TM decoding software 

and updates the parameters. 

ESA_Neil proceeds with the S-band checkout. Procedures are in italics, and the results in 

normal font. 

The spacecraft is powered up using the EM battery. 

1+ Command S band transmitter (TX) on into "24/7" mode(7) - OK Result is TX comes on 

and stays on – failure is no signal being transmitted – possible causes failed power supply to 

S band , failed S band transmitter chain 

Carrier comes up and stays up. 

Measure and record TX frequency- OK Result is 2410.835MHz +/- 10kHz but acceptable 

tolerance cannot be defined until accuracy of frequency counter in use can be determined 

416






Frequency measured is 2401.8292 MHz. After 10 minutes it drops to 2401.8280 MHz. 

Frequency counter is fresh from the calibration lab, so it should be “very accurate”. 

Command DTMF telemetry on (mode 6) - OK Result is after waiting no more than 1 minute a 

DTMF burst is transmitted – failure is no DTMF tones – possible cause – damaged encoder 

board inside S band box 

DTMF telemetry transmitted. 

Receive and record first DTMF telemetry message - OK Result is valid data received and 

decoded/displayed on software – failure is no satisfactory decode – possible causes – 

incorrect software version on laptop, incorrect connections to soundcard, incorrect levels, S 

band RX off frequency. 

DTMF telemetry is received and decoded. 

Command TX off (mode 0)- OK Result is TX stops 

Earlier “TX” is the name given to the whole unit, but it is assumed that, since “mode 0” is 

mentioned, this step is intended to simply command the unit back to the default state rather 

than actually power it off. In this case the result is ok. 

Validate received DTMF telemetry as satisfactory - OK Result is Power within range of 2.25 

and 2.75 watts, PA temperature within the range of ambient temp +/- 5 degree C –failure is 

values outside these ranges – possible causes -, defective power sensor, failed temperature 

sensor, failed encoder 

Temperature is initially reported at 24.6 degrees and the PA is at 1.636 Watts. NOTE: This 

power output is not within the acceptable range, however, this is almost certainly due to the 

calibration of the ground software. (It has always read about the same.) 

2+ Command DTMF telemetry on (mode 1) - OK Result is after waiting no more than 1 

minute a DTMF burst is transmitted 

DTMF telemetry transmitted. 

Receive and record first DTMF telemetry message- OK Result is valid data received and 

decoded/displayed on software 

DTMF telemetry is received and decoded. 

Command TX off (mode 0)- OK Result is no more telemetry is received 

Again, it is assumed that this simply means to command the unit back to the default state, in 

which case the result is ok. 

417







and 2.75 watts, PA temperature within the range of ambient temp +10 degree C 

Temperature is initially reported at 24.29 degrees and the PA is at 1.636 Watts. NOTE: This 

power output is not within the acceptable range, however, this is almost certainly due to the 

calibration of the ground software. (It has always read about the same.) 

3+ Command TX into data configuration (mode F)- OK Result is transmitter does not come 

up until data telemetry burst is transmitted 

Transmitter comes up only when data telemetry is transmitted. 

Command a transmission of a telemetry data burst- OK Result is data telemetry burst is 

transmitted 

Data telemetry commanded and transmitted. 

Command TX off (mode 0)- OK Result is no more data telemetry is received 



Validate received data is correct - OK Result is………………….. 

Data is received and correct. 

4+ Command TX into transponder configuration (mode 3) - OK Result is TX does not come 

on 

The carrier did initially come up, but this was because the commanding unit was already 

using CTCSS tone encoding. Repeating this step without a PL tone works correctly. 

Test voice transponder mode with and without CTCSS tone encoding. - OK Result is with PL 

tone voice transponder is operational but with no PL tone transponder does not operate and 

TX does not come on – failure is either voice transponding not possible or always possible – 

possible cause failed controller 

Transponding is possible with, and only with, a PL tone. 

Check squelch "tails" correctly OK Result is noise is transmitted for approx 3 seconds after 

input carrier drop before TX carrier drops- 

It is more like 5 seconds, but this is probably fine. 

418






Command TX off (mode 0) - OK Result is TX stays off 



5+ Command TX on into "24/7" mode (mode 7)- OK Result is TX does come on at once 

The carrier comes up. 

Command TX into transponder configuration (mode 4) - OK Result is no change 

There is no change. 

Test voice transponder mode with and without CTCSS tone encoding. - OK Result is is with 

PL tone voice transponder is operational but with no PL tone transponder does not operate 

and TX does not come on – failure is either voice transponding not possible or always 

possible – possible cause failed controller 

Transponding is possible with, and only with, a PL tone. 

Check squelch "tails" correctly OK Result is noise is transmitted for approx 3 seconds after 

input carrier drop before squelch closes and hissing noise is replaced by silent carrier 

It is more like 5 seconds, but this is probably fine. 

Command TX off (mode 0) - OK Result is TX drops 



6+ Command TX on into "24/7" mode (mode 7)- OK Result is TX does come on at once 


Command TX into data configuration (mode 8) - OK Result is no change 

The carrier does come up every 36 seconds to transmit nominal mode beacons, but this is ok. 

NOTE: Must remember to switch to a wide filter for data reception. 

Command a transmission of a telemetry data burst- OK Result is data telemetry burst is 

transmitted 

Data telemetry commanded and transmitted. 

419






Command TX off (mode 0)- OK Result is TX carrier drops 



Validate received data is correct- OK Result is ……………………………… 

Received data is fine. 

The UHF downlink frequency is measured as 437.2502 MHz (which is better than expected). 

To test the uplink ESA_Neil performs a very simple adjustment of the handheld frequency 

and listens to the difference when transponding. A "clear" silence is heard from XXX.X40 to 

XXX.X65 MHz, but one step outside of this in either direction and noise can be heard in the 

background. This puts the receive centre at XXX.X525, which is a little surprising as 

AMS_Howard measured it at 5kHz down. I can also only command the S-Band unit in the 

same range, which adds to the evidence. 

Note 1: It is tough to test the strength of the transponder signal alone, as it's hard to 

distinguish one’s voice itself from the copy of it streaming from the laptop speakers. 

Note 2: ESA_Neil could only get the handheld to vary by increments of 5 kHz. 

7+ Wait 10 minutes before performing this test 

Command TX on into "24/7" mode(mode 7)- OK Result is TX does come on at once 


Command DTMF telemetry on (mode 8) - OK Result is no change 

Only change is the reception of DTMF telemetry. 


tolerance cannot be defined until accuracy of Frequency counter in use can be determined 

The frequency is measured as 2401.8277 MHz. 

420






Receive and record ten DTMF telemetry bursts- OK Result is ten full DTMF telemetry bursts 

received 

Done. 


tolerance cannot be defined until accuracy of Frequency counter in use can be determined. 

The frequency is measured as 2401.8255 MHz. 

Command TX off (mode 0) - OK Result is TX carrier drops 




and 2.75 watts, PA temperature within the range of ambient temp +35 degree C 

The ten DTMF telemetry bursts are recorded as: 

09F388 

D8F367 

B8E3F7 

98E398 

88E348 

78D300 

68D3C8 

58D318 

48D328 

38C308 

421






These are translated as a steady increase in temperature (ending at around 32 degrees) and a 

very slight drop in power. 

PROBLEM 269: The second incarnation of the groundstation laptop dies. It appears to be a 

RAM malfunction. It refuses to boot back up and the relevant CSV file from the DTMF 

decoding is probably lost. (Lucky I wrote them down.) 

8+ Simulate expected working levels of S/N ratios say 

?30dB, ?20 dB and ?15 dB for each S band downlink and UHF uplink/downlinks and repeat 

tests 1- 6 above at each S/N level 

It is not possible to repeat the tests without fetching and setting up another laptop for the 

groundstation. Also, ESA_Neil is not clear on what is meant by this step – putting attenuators 

between the ground antennas and radios? 

Assuming all ok then proceed to pub 

Given the hour of the day and the lack of hardware and expertise with which to continue, it is 

assumed (for now at least) that everything is ok and ESA_Neil dutifully proceeds to the pub. 

ESA_Neil sets the spacecraft charging. 

21 st April 2005 

MODIFICATION 136: Pin 12 in the 25-pin ACDS connector is moved to pin 21, and the 

old pin 21 is removed 

ESA_Neil removes pin 21 from the harness side of the ACDS coil driver 25-pin connector. 

The end is taped over with kapton for safety. Pin 12 is then moved to become pin 21. The 

422






spacecraft is powered up and ACDS telemetry received with the sun sensors lit and shaded in 

turn. 

-y shaded packets: 



0x00000000: 74 05 86 0E 22 1E F1 05 EB 05 F7 05 F1 05 00 00 |t..."...........| 

0x00000010: 00 00 70 00 8E 00 0D 02 22 01 00 00 00 00 00 00 |..p.....".......| 



0x00000000: 6C 05 73 0E 0A 1E EB 05 E0 05 F1 05 EE 05 0B 00 |l.s.............| 

0x00000010: 00 00 70 00 8E 00 08 02 22 01 00 00 00 00 00 00 |..p.....".......| 



0x00000000: 66 05 68 0E 0E 1E F9 05 EE 05 EE 05 F4 05 00 00 |f.h.............| 

0x00000010: 00 00 70 00 8E 00 08 02 25 01 00 FE 01 00 00 00 |..p.....%.......| 

-y lit packets: 



0x00000000: 72 05 7F 0E 12 1E DE 05 04 06 F7 05 F4 05 84 04 |r...............| 

0x00000010: 00 00 70 00 8E 00 02 02 3B 01 00 00 00 00 00 00 |..p.....;.......| 



0x00000000: 70 05 7E 0E 10 1E DB 05 04 06 E9 05 DB 05 39 04 |p.~...........9.| 

0x00000010: 00 00 70 00 8E 00 05 02 28 01 00 00 00 00 00 00 |..p.....(.......| 



0x00000000: 70 05 7D 0E 0F 1E DB 05 0D 06 F1 05 F4 05 4D 04 |p.}...........M.| 

0x00000010: 16 00 70 00 8E 00 FF 01 1F 01 00 00 00 00 00 00 |..p.............| 

+y shaded packets: 



0x00000000: 6B 05 6D 0E 0B 1E F1 05 F7 05 F4 05 F4 05 00 00 |k.m.............| 

0x00000010: 00 00 70 00 8E 00 10 02 14 01 00 02 00 00 00 00 |..p.............| 



0x00000000: 6B 05 6D 0E 0B 1E F1 05 F7 05 F4 05 F4 05 00 00 |k.m.............| 

0x00000010: 00 00 70 00 8E 00 10 02 14 01 00 02 00 00 00 00 |..p.............| 



0x00000000: 6C 05 68 0E 0A 1E E0 05 F1 05 07 06 EE 05 08 00 |l.h.............| 

0x00000010: 00 00 70 00 8E 00 26 02 11 01 00 00 00 00 00 00 |..p...&.........| 

+y lit packets: 

423








0x00000000: 69 05 64 0E 10 1E EE 05 07 06 73 06 A4 05 00 00 |i.d.......s.....| 

0x00000010: 7D 03 70 00 8E 00 0A 02 33 01 00 00 00 00 00 00 |}.p.....3.......| 



0x00000000: 6C 05 65 0E 0E 1E EE 05 EE 05 73 06 BC 05 00 00 |l.e.......s.....| 

0x00000010: 8B 03 70 00 8E 00 10 02 30 01 00 00 00 00 00 00 |..p.....0.......| 



0x00000000: 6C 05 65 0E 0E 1E EB 05 E9 05 60 06 B1 05 00 00 |l.e.......`.....| 

0x00000010: 64 03 70 00 8E 00 0A 02 2D 01 00 00 00 00 00 00 |d.p.....-.......| 

ESA_Neil tests the timer reset and skip functionality with the EM e-box. 

ESA_Neil applies the T-Pod arm connector and the EPS arm connector at 17:06:30. 

The +x pod is reloaded with no cube in it. 

ACDS_Lars reports that the +y sun-sensor problem is fixed. The ACDS checkout is complete 

and successful. 

The loose wire from pin 21 is cut at the nearest fixation point and the cable is recommitted 

NOTE: The following connectors need to be reglued: MAGIC pyro, all OBC (after prom 

change), ACDS 25-pin. 

After only 66 minutes the x pods fire but the +y one does not 

PROBLEM 270: The electronics of the +y T-Pod have been damaged to the extent that the 

relay and heater block do not fire. This is almost certainly related to problem 261. 

The +x pod is secured with safety bolt and UWE-1 inside. 

ESA_Neil powers down the spacecraft and leaves it to charge. 

22 nd April 2005 

ESA_Neil buzzes through and confirms that the +y pod is located as pod number two on the 

FPP T-Pod safe/arm connector. 

ESA_Neil removes the other pod connections from the arm connector and runs the +y pod 

connection through an ammeter to see the pulse. 

424






When the pulse is sent it registers on the ammeter, and click is heard as the relay switches, but 

the light does not come on. This implies that the fault is in the e-box itself. 

NOTE: The following OBC work is not strictly allowed after the vibration test. However, 

the requirements from the launch authority are simply for the safety of the other spacecraft, 

which this can have no effect on. Project management decides that, in the light of problems 

267 and 268, it is worthwhile to take the very small risk of changing something that has 

passed vibration testing in order to retain the camera payload. 

MODIFICATION 137: The OBC is disintegrated and the PROMS replaced. 

ESA_Neil labels all the connectors and then disintegrates OBC. 

PROBLEM 271: One of the screw locks will not unscrew and simply rotates the mating 

half. The only way to get it off is to cut the head of the bolt. 

425






The OBC is carefully opened and the PCBs separated from the box sides. 

The old proms are removed and the new proms inserted. 

NOTE 1: The glue holding the old proms in caused some problems to remove them. Most 

importantly one corner of each prom was held more tightly than the other and the proms each 

came half-out at an angle that stressed the sides of the socket slightly. They had to be 

carefully levered all the way out with a screwdriver. 

NOTE 2: One of the four parts of the circular standoff in socket two broke off. Hopefully 

this is not serious. 

The OBC is powered up using a long pair of wires from its power connector in the spacecraft, 

while the debug port is plugged into the linux box. It comes up fine, including the ACDS and 

TCS threads. 

ESA_Neil reintegrates the OBC box, including replacing the damaged screw lock (both 

halves). The proms and new screw lock are glued in as before, then all the external bolts are 

retightened and re-glued. 

426






Once the OBC is integrated back into the structure all the connectors are replaced, tightened 

and glued. 

PROBLEM 272: One of the heads of one of the screw locks on the OBC snapped off during 

torquing. It is not possible to remove the threaded shaft from the socket. Instead the 

connector is simply glued in place (also bolted at the other side). 

The spacecraft is booted up for the functional checkout of the CAM. 

Pinging the camera looks ok: 


Subsystem: CAM | MID: 0x03 | Length: 3 | Time: 22-04-2005 16:36:26 

0x00000000: 43 41 4D |CAM.............| 

Asking for housekeeping data looks ok: 


427







0x00000000: 25 35 06 81 

Asking for some accelerometer data returns some surprising results the first time (not two 

times 64 bytes): 



0x00000000: 06 02 06 02 06 02 06 02 08 02 07 02 07 02 09 02 |................| 

0x00000010: 05 02 07 02 05 02 08 02 07 02 07 02 06 02 06 02 |................| 

0x00000020: 07 02 05 02 07 02 06 02 08 02 07 02 07 02 07 02 |................| 

0x00000030: 05 02 05 02 06 02 07 02 06 02 07 02 08 02 07 02 |................| 



0x00000000: 02 08 02 07 02 07 02 07 |................| 



0x00000000: 02 07 02 06 02 07 |................| 



0x00000000: 06 02 |................| 



0x00000000: 05 02 |................| 



0x00000000: 07 02 |................| 



0x00000000: 07 02 |................| 

The parameters are uploaded and three pictures are taken and downlinked. The first one just 

seems to be noise, the second seems to be entirely white, and the third seems to be noise 

again: 

CAM_Morten advises a power cycling of the camera, suggesting that noise is a known 

problem. The white picture looks like the camera is just pointed at something too bright (a 

mirror is being used to point it outside). 

The camera is power cycled. The ping is then ok, as is the accelerometer data: 



0x00000000: 07 02 06 02 07 02 07 02 07 02 07 02 0A 02 06 02 |................| 

428






0x00000010: 07 02 07 02 07 02 07 02 07 02 07 02 06 02 07 02 |................| 

0x00000020: 07 02 06 02 06 02 07 02 07 02 06 02 07 02 07 02 |................| 

0x00000030: 06 02 07 02 07 02 06 02 07 02 07 02 06 02 06 02 |................| 



0x00000000: 08 02 08 02 06 02 07 02 07 02 08 02 06 02 07 02 |................| 

0x00000010: 08 02 07 02 07 02 07 02 07 02 07 02 07 02 07 02 |................| 

0x00000020: 06 02 07 02 07 02 07 02 07 02 08 02 07 02 06 02 |................| 

0x00000030: 09 02 07 02 07 02 07 02 06 02 08 02 07 02 07 02 |................| 

Two new pictures are taken, this time pointing indoors. They seem fine, although there is 

some kind of block error which is offsetting the lower half of each. This is strange but not 

insurmountable. 

NOTE: This image is an unprocessed thumbnail taken with rough parameters of an indistinct 

object (the blinds) which is far too close. The important thing is that it is clearly of something 

and is not just homogenous or noise. 

The rest of the camera checkout (getting good pictures) is postponed until Monday and will be 

done remotely via MCC in Aalborg by CAM_Morten. 


ESA_Neil configures the spacecraft and test groundstation so that it can be remote controlled 

from Aalborg by MCC. 

CAM_Morten performs a remote checkout of the camera. Several pictures are taken and 

thumbnails downloaded and decoded successfully. 

PROBLEM 273: During the checkout the CAM occasionally stops responding. This later 

transpires to be because the OBC was being rebooted by EPS – overnight it booted a total of 

forty-seven times. 

NOTE: The above problem is either a ping-pong problem between OBC and EPS during 

heavy payload operations, or a power problem with EPS since, due to lacking documentation, 

the spacecraft was not powered properly (using the battery AND external power). It is 

strange as the ping-pong was stable apart from the original EPS overflow when we tested it, 

429






and then the removal of the 'ticket' should have cured that (and was tested for a while and 

seemed fine). 


PROBLEM 274: Last night the OBC randomly reset the boot counter again. This could be 

due a momentary power drop during boot-up due to the connection of the external power 

suppy while the battery was still connected as well (due to lacking EPS documentation). 

PROBLEM 275: Due to the accidental (lacking documentation!) dual-powering of the 

spacecraft the batteries have run down low overnight and are found locked (7V) in the 

morning. Unlocking them by charging reveals them at about 18.5V, which is dangerously 

low. However, they seem to charge back up without issue. 

The reboot problem (273) is a serious issue as we would end up losing huge amounts of data 

on every orbit and this would greatly restrict the possibilities for payload operations. Perhaps 

part of the problem is that all the troubleshooting lately has been done by teams remotely, 

through ESA_Neil in both directions. This is not only time-consuming it is also quite 

certainly a somewhat inefficient way of passing information. 

ESA_Jason solders wires to the TCS thermistors. 


ESA_Neil charges the spacecraft, routes the new thermistor harness and disintegrates the old. 

Before each new thermistor is routed into the spacecraft it is tested with a power supply and 

voltmeter to make sure that they are wired up correctly this time. 

430






Several cable ties across the –x side of the spacecraft are replaced during the new routing. 

The shunt resistor thermistor measurement cable is connected to pin 8 of the ACDS MUX 

connector on the OBC and the titanium ring thermistor measurement cable to pin 9. 

431






On TCS connecter on the OBC pin 1 is the thermistor measurement from the –y lateral panel 

between the BATT and PCUEPS, pin 2 is from the –x+y top-bracket and pin 3 is from the 

Camera. 

ESA_Jason solders all the TCS power wires together and all the TCS ground wires together. 

ESA_Neil performs a functional checkout on the +y T-Pod, following instructions from 

CANX_Fred. 

Relay Test 

1. Disconnect a heater cable from T-POD. 

OK 

2. Measure the battery pack voltage. 

Fluctuating randomly at around 4-5 volts (this is strange, it was at 10.55V last time when I 

finished charging before attempting to fire it) 

3. Push the momentary pushbutton switch and keep pushing it. 

OK 

4. Provide power (9 V, max. 300 mA) to T-POD through DB15 connector. Be careful with 

polarity. 

OK 

5. Make sure the relay is powered on. Audible click sound shall be heard. 

Click heard. 

6. Check the current of the power supply and it shall be between 150 mA and 300 mA. 

Incorrect, it is 130 mA. 

7. Read the battery pack voltage using a voltmeter. The voltage shall be lower than the 

previous reading. 

Fluctuating randomly at around 4-5 volts 

8. If everything mentioned above fits, the relay itself is operating fine. 

? 

9. When the relay is powered on, the green LED should be lit. 

It is NOT lit. 

10. If the LED is not lit, make sure that the relay is powered on by the power supply. 

Is it, I can hear it. 

11. Measure the battery pack voltage. 

Again? Fluctuating randomly at around 4-5 volts 

432






12. Release the momentary push button. 

OK 

13. Record the voltage and current of the power supply plus the battery pack voltage. 

9V, 130mA, Fluctuating randomly at around 4-5 volts 

Battery 

1. Disconnect a heater cable from T-POD. 

OK 

2. Charge the battery pack with the power supply. (11 V, 500 mA) 

OK 

3. Leave the power supply connected to T-POD. 

OK 

4. Send an activation signal to T-POD. 

OK 

5. Is the relay still powered on? What's the current reading from the power supply? 

Relay clicks, power supply reads 136mA. 

6. Is the green LED lit? 

Yes 

7. Push the momentary pushbutton switch. 

OK 

8. Everything should be turned off. 

Yes (but only if activation signal is off) 

So, It looks as if there is a problem with the battery pack, as suggested. 

ESA_Neil tests all 5 thermistors powered together using a power supply and reads out the 

measurement with a voltmeter, they work fine. 

The spacecraft is booted and nominal mode is achieved. The TCS cable is plugged in the 

spacecraft stays in nominal mode. 

By heating the TCS thermistors with a finger it can be seen that first in TM is the one on the – 

y shear panel between the battery and the PCU, the second is on the –x+y top bracket and the 

third is on the CAM. 

Spacecraft is shut-down and put on charge. 

433






ESA_Neil glues the new thermistors in place and tightens and glue screw locks on the TCS 

connector. 


ESA_Neil powers up the spacecraft for long-duration testing and monitors the telemetry from 

remote (s-band data is enabled). This is in an attempt to replicate problem 273 and determine 

if it was a software problem or an external power problem. 

The spacecraft remains much more stable than problem 273, with only two unexpected 

reboots in 5 hours (and at least one of these could have been operator error). This implies that 

problem 273 was due to incorrect external powering, and is nothing to worry about. 

At a battery voltage of around 22.3V the spacecraft drops to safe-mode. This successful as 

part of the EPS functional checkout. 

The spacecraft is shutdown and left to charge overnight. 


The spacecraft is fully charged and then powered up. (1040) 

ESA_Neil remote controls the spacecraft, via the groundstation, from the office, testing 

payload operations with the camera. 

434






Bottom portion is ‘missing’ (white) off of the first two pictures. (May not have got timing 

right). The power to the camera was cycled in between. 

The third is better, but too dark. (No power cycle.) This is a large thumbnail. 

Power cycle, 4 th is missing the bottom again. Do not power cycle. 5 th is ok, but dark again. 

435






Taking 6 th after power cycle: too dark. Full picture downloaded though. 

SOLUTION: Someone has turned the cleanroom lights off. 

Taking 7 th and 8 th with integration time high = 4D. Still too dark, no real difference. Power 

cycle and take 9 th with integration time high = 8f. This time it is too bright. 

Take 10 th with integration time high = 6c, this is just noise and the file size is wrong (20672 

bytes instead of 20480.) Take 11 th slower with integration time high = 6c. Again, just noise. 

(File size is 20654 instead of 20480.) 

436






Going slower again makes sure the pictures are not noise. 12 is still too bright, but better. 13 

uses integration time high =59, and is, strangely, too bright. (Maybe the lights were turned 

back on.) 

Picture 14 is taken, using auto exposure, get half picture missing again… 

SOLUTION: The partial missing image only seems to happen when using auto-exposure. 

Taking single pictures is fine. Asking CAM_Morten to check his code. 

The spacecraft battery runs down low enough to enter safe-mode. However, the OBC has not 

been re-booted all day, despite several hours (5) of uptime with heavy payload operations. 

This is very good and heavily implies that problem 273 was due to incorrect powering. 


Spacecraft is not fully charged so it is booted up in “battery failure mode” (after 12 minutes). 

The plan was for CAM_Morten to perform some camera calibration work remotely, but the 

OBC / EPS was clearly unstable and rebooting often, a recurrence of problem 273, which 

suggests that it could be a software error in the PCU code that only manifests itself during a 

“battery failure”. This is annoying, but not too serious, as this is a failure mode anyway, and 

the spacecraft will only be powered in the sunlit phases, so the OBC would be not be on for 

long to start with. 

The spacecraft is left on charge and ESA_Neil closes the sides. 

437






1 st May 2005 

ESA_Neil removes and packs up the side protectors, attenuation caps and related bolts and 

clips in preparation for the fit check. 

The spacecraft is fully charged so that external harness is removed and the charging station 

switched off. 

The area is made safe with security tape so that no-one goes close to the spacecraft without 

the side protectors on it. 

2 nd May 2005 

ESA_Neil and SYS_Jörg travel to Omsk for the fit-check, meeting SSTL_Andy, SSTL_Ed 

and CST_Nina in Moscow. 

3 rd May 2005 

SSTL, Polyot, CST and SSETI delegations discuss the fit-check plan and schedule in the 

Polyot Design Office in Omsk. 

4th May 2005 

The Fit-Check team go to the Polyot assembly facility. 

ESA_Neil and SYS_Jörg open the transport container and the tool-box, add the +z attenuation 

cap and the lifting frame, and then remove the dimension dummy from the container and use 

the lid of the container as a temporary integration table. They carry out a visual inspection. 

PROBLEM 276: Two washers are found in the +y0x compartment, and 2 washers and a 

long M5 bolt are found under the spacecraft. 

438






Concerned about the unexplained bolts, SYS_Jörg and ESA_Neil remove the lateral panels 

and discover the following: 

- On the –x-y compartment top bracket the top –x bolt, and both –y bolts are missing 

- On the +x-y compartment top bracket both –y bolts are loose and both +x bolts are 

missing 

- Two bolts and a nut (all M5) are laying in the activation switch well. 

- One bolt, two washers and three nuts are found underneath the foam in the bottom of 

the transport container 

- On the +x+y compartment top bracket the +x bolts are loose 

- On the 0x+y mid height bracket the bolts are loose 

PROBLEM 277: It seems form the above that several of the top-bracket bolts have loosened 

and fallen during the transport vibrations. This is very surprising and unexplained – since the 

bolts were torqued at integration. However, this was a long time ago and some settling would 

have taken place since then. 

439






All bolts are replaced and tightened, apart from one bolt in the top –x slot of the –x-y top 

bracket – since the bolt canon be located (but it not in the spacecraft). 

The T-Pod protrusions are checked against the latest drawings. The upper T-Pod dummy is 

relocated slightly. 

The lateral panels are replaced and all bolts tightened. Then the lateral patch antenna caps are 

placed. 

PROBLEM 278 The +x cap does not fit since the foam antenna was made before the 

attenuation foam was added and is not quite accurate. 

ESA_Neil removes the outer surface of the antenna dummy with a scalpel. 

SYS_Jörg and ESA_Neil place the side protectors and red tags, an d then confirm that all the 

following RBF items are present: 

1) lifting frame 

2) –Y side protector 

3) +X side protector 

4) +Y side protector 

5) –X side protector 

6) Camera lens cap 

7) +X thruster caps 

8) –X thruster caps 

9) +Y sun sensor cover 

10) –Y sun sensor cover 

11) –X attenuation cap 

440






12) +X attenuation cap 

13) +Z attenuation cap 

14) +Y T-Pod safe bolt 

15) –X T-Pod safe bolt 

16) +X T-Pod safe bolt 

17) EPS safe plug 

18) T-Pod safe plug 

19) Pyro safe plug 

In preparation for integration to the launch adapter the +y side protector, +y T-Pod bolt and 

the +y sun sensor cover are removed. 

The spacecraft is lifted to a transport trolley and wheeled into the integration hall. 

Inspection of the interface on the launch adapter yields no concerns: the asap ring is in the 

correct orientation, the activation pillar mounting appears to be accurately positioned, the 

Polyot deployment switch is not obtrusive, and there is adequate clearance with the 

neighbouring spacecraft (China-DMC to the left and Moxaets to the right). 

441






The dummy is then lifted onto the launch adapter. 

NOTE: There is a large patch antenna on the side of the Topsat interface pillar, directly 

“behind” SSETI Express. Polyot reassures us that this will not transmit until after SSETI 

Express has separated from the launcher. 

NOTE: The –x side protector is very close to the China-DMC quadrafila antenna, and to the 

nearest corner of China-DMC. It appears to be dangerously close as the spacecraft is 

swinging slightly as it is lowered. This side protector probably has to be removed before 

integration. 

NOTE: The +y sun sensor cover is in line with one of the side protector bolts of the Moxaets 

spacecraft. We discuss and agree that this protector will not still be present when the flight 

model of Express is integrated to the adapter, therefore it is safe. 

SSTL_Ed routes the trickle charge harness, the separation detonator harness, the telemetry 

switch harness and the grounding strap for SSETI Express. 

442






SYS_Jörg and ESA_Neil measure the location of the Polyot deployment switch, it will 

interface to the baseplate 50mm from the –x lateral panel and 140mm from the -y lateral 

panel. 

PENDING MODIFICATION: The skin of the baseplate should be strengthened at the point 

where the Polyot deployment switch interfaces to SSETI Express. 

Polyot perform a pyro firing test with two command current pulses on each of the two 

detonators on each of the spacecraft. The pulses to the primary SSETI Express detonator are 

5.46A, and the pulses to the secondary detonator are –5.52A. This is fine. 

In discussion it is discovered that the other end of the battery trickle charge harness runs to a 

d-sub connector for each spacecraft on the side of the launch adapter. These are accessible 

through a hatch in the launcher fairing in case of a launch delay. 

During the horizontal phase of the integration of the launch adapter SSETI Express and 

China-DMC will be each 45 degrees from the top of the adapter. This means that the 

requirements on safe RBF and ABF work are stricter, since any dropped items would impact 

on other spacecraft. 

The launch adapter is moved back in the assembly room and lifted onto the fairing fit-check 

adapter. It is then carefully rotated to ensure that all spacecraft extremities fit within the 

dynamic envelope of the launcher fairing. 

443






NOTE: For SSETI Express the test is very successful, as there are clearances of several 

centimetres around each of the external corners. For some of the other spacecraft it is a lot 

closer, but there are no serious problems. 

ESA_Neil integrates the activation pillar to the launch adapter. It fits fine and appears to 

depress the activation switches properly. SSTL_Andy confirms that the SSETI Express 

activation switches are within about 0.5mm of the China-DMC switches in the vertical 

direction. 

NOTE: It is slightly difficult to position the pillar as the battery charge stud is protruding 

slightly too low and is therefore a tight fit with the spring contact on the pillar. However, if 

memory serves then this is corrected in the flight model already. 

ESA_Neil, SYS_Jörg and SSTL_Andy discuss the RBF / ABF sequence for SSETI Express. 

The following procedure is identified: 

444






Phase I: Pre-integration 

1) The spacecraft shall be lifted, the engineering separation ring removed and the flight 

separation ring integrated. Then the spacecraft shall be returned to the table. 

2) The +y side protector is removed (since it is behind the s/c after integration). 

3) The +y lateral panel corner bolts are applied 

4) The +y T-Pod bolt is removed (since it is behind the s/c after integration). 

5) The +y sun sensor cover is removed (since it is behind the s/c after integration). 

6) The –x protector is removed (since it is too close to China-DMC during the 

integration) 

7) The –x lateral panel corner bolts are applied. 

8) The +y and –x lateral panel corner bolts are torqued and glued. 

9) The –x antenna cap is removed (since the quadrafila antenna on China-DMC would 

block its removal later) 

10) The –x T-Pod safe bolt is removed (since it still has two electrical barriers) 

11) The –x thurster caps are removed (since they are not necessary at this time) 

12) The +x side protector is temporarily removed (for the next three points) 

13) The +x antenna cap is removed (for consistency, and since it has small droppable 

objects) 

14) The +x T-Pod safe bolt is removed (since it still has two electrical barriers) 

15) The +x thrusters caps are removed (since they are not necessary at this time) 

16) The +x side protector is replaced (to protect solar cells during integration) 

17) The +z antenna cap is removed (for consistency and since it has small droppable 

objects) 

Phase II: Post-integration, pre-turn 

1) The lifting frame is removed (since it is no longer needed) 

2) The grounding strap is applied 

3) The activation pillar is applied 

4) The trickle charge harness is applied 

5) The telemetry switch harness is applied 

6) The deployment switch harness is applied 

7) The EPS SAFE plug is removed, with the saver 

8) The voltage across the battery pins on the EPS RBF/ABF socket is checked to make 

sure that the activation switches are depressed and the battery is disconnected 

9) The EPS ARM plug is applied 

10) The T-Pod SAFE plug is removed, with the saver 

11) The voltage across the T-Pod RBF/ABF socket is checked to make sure that the T- 

Pods are not being fired by EPS. 

12) The T-Pod ARM plug is applied 

13) The Pyro SAFE plug is removed, with the saver 

14) The Pyro ARM plug is applied 

445






15) The timer capacitor discharge cable is temporarily applied between the external power 

supply socket and the OBC / S-Band socket to reset the timers 

16) The saver is removed from the OBC / S-Band socket 

17) The S-Band ENABLE plug is applied 

18) The +x side protector is removed (since it would be difficult to remove later without 

damaging Moxaets, and the corresponding protector on Moxaets is already removed 

so we don’t need to worry about damage from them) 

19) The EPS ARM plug, the T-Pod ARM plug, the Pyro ARM plug and the S-Band 

ENABLE plug all have their screw locks tightened and glued. 

20) The –y thermal foil is kapton taped into place. 

21) The –y side protector is removed (see note below) 

22) The –y lateral corner bolts are applied (see note below) 

23) The +x lateral panel corner bolts are applied 

24) The –y (see note below) and +x corner bolts are torqued and glued. 

NOTE: The –y side protector removal may be moved into the next phase in order to ensure 

that the side of the spacecraft outermost on the adapter is protected from passers-by and the 

tilting apparatus. This would move steps 21, 22 and part of 23 to phase III. However, this 

will only be possible if a way is found to encapsulate the corner bolts as they are applied so 

that there is no possibility of dropping them. 

Phase III: Post-turn (to horizontal position for integration to the launcher upper stage) 

1) The camera lens cap is removed 

2) The –y sun sensor cover is removed 

The SSETI Express dummy is disintegrated from the launch adapter and placed on top of the 

transport container again. 

May 5 th 2005 

SSTL, Polyot, CST, SYS_Jörg and ESA_Neil discuss launch site activities in the Polyot 

design office all morning. 

It is envisaged that the integration route and RBF sequence will be rehearsed in the afternoon 

once Polyot are ready, but the schedule slips and we are left to our own devices for the rest of 

the day. 

446






May 6 th 2005 

Polyot perform a firing test of the China-DMC mass dummy. 

ESA_Neil and SYS_Jörg prepare the dummy model for integration to the launch adapter by 

performing all Phase I activities as defined on the 4 th May. 

STRU_Antonio reports the latest mass properties from the CATIA model, and ESA_Neil 

calculates the final launch and flight mass of the spacecraft: 

Launch configuration: 

Measured mass of Express 58 kg 

Fuel 2.4 kg 

EM separation ring 

-2.35 kg 

M8 bolts x12 0.28 kg 

FM separation ring 3.0 kg 

Separation springs x6 0.72 kg 

Total 

61.95 kg 

(This does not include activation pillar and harness.) 

Flight configuration: 

Measured mass of Express 58 kg 

Fuel 2.4 kg 

EM separation ring 

-2.35 kg 

Top half of FM separation ring 1.0 kg 

Three Cubesats -3.0 kg 

Total 

55.95 kg 

The two Russian spacecraft are integrated back to the adapter, Topsat is already present, then 

China-DMC is lowered onto it also. 

The SSETI Express dummy in Phase I configuration is lifted into the assembly hall and 

integrated the launch adapter. To do this it is positioned off in the –y direction (s/c axes) of 

the launch adapter, aligned on the x axis with the bolt holes, at such a height that the 

separation ring just clears the Polyot deployment switch. It is then manoeuvred in the +y 

direction until it is over the bolt holes, and then lowered into place. The bolts should be 

present in the ASAP ring during this whole operation. 

447






SYS_Jörg and ESA_Neil perform the phase II activities of launch preparation. However, the 

–y side protector is left ON to test the possibilities of removing it in phase III instead. 

NOTE: We need to get a stumpy screwdriver for removing the +x side protector at the 

launch site. The one we are using here is too long and dangerously close to the Moxaets solar 

panel. 

PENDING MODIFICATION: We need to add handles to the +x and –y side protectors for 

ease of manoeuvring. 

The launch adapter is turned by 90 degrees into the orientation it will be in when integrated to 

the launcher. 

The lifting struts are still attached to the sides of the launch adapter when it is rotated, which 

is unexpected. After discussion with Polyot they agree to remove the top and bottom struts 

before turning, but leave the others for handling reasons. This is fine, since it is the top one 

that is problematic for us. 

448






Sys_Jörg and ESA_Neil perform the phase III activities of the launch preparation, with the 

addition of removing the –y side protector as a test of feasibility. 

It turns out to be best to remove the ‘upper’ (-x) –y protector bolts first, so that the lower edge 

is being controlled once the protector is free to move. 

Access is not too hard – it would be possible, and probably best, to remove this protector after 

the turn. Therefore: 

PENDING MODIFICATION: A way has to be found of encapsulating the bolts as they are 

applied to the –y panel, so that there is no way of dropping them. 

PENDING MODIFICATION: A way should be found of attaching the screw-driver to a 

wrist strap so that it cannot be dropped. 

PROBLEM 279: The Moxaets team requests that their solar panel protector is replaced 

while we perform our RBF work, since it is a large horizontal plane right underneath us. This 

presents us with a problem though, since one of the bolts on their protector puts their tooling 

perilously close to the +y sun sensor. 

449






SOLUTION: We negotiate and agree on a compromise: the Moxaets team will remove from 

their side protector the four bolts which are closest to SSETI Express before SSETI Express is 

integrated to the launch adapter. This resolves the issue. 

The SSETI Express RBF sequence is therefore defined and completed. The spacecraft is 

removed from the adapter. 

LESSON LEARNED 23: Side protectors should have handles 

LESSON LEARNED 24: All RBF and ABF should be BIG and preferably there should be 

no ABF items 

LESSON LEARNED 25: Side protectors should not need replacement bolts 

LESSON LEARNED 26: Submit drawings to launch authority after each update 

LESSON LEARNED 27: Request early-on the orientation for latest RBF activity 

LESSON LEARNED 28: RBF tooling should be small and minimised 

LESSON LEARNED 29: Set up document control system with launch authority 

7 th May 2005 

We all attend the design office in the morning to discuss and agree the protocol. 

SYS_Jörg and ESA_Neil pack up the dummy model and the toolbox into their original 

containers. The Polyot guys do not have a metal strap crimper, but come up with an 

ingenious way of re-tightening and securing them. 

450






The shipment is ready and handed over to CST_Nina. 

Polyot return from the vibration facility with the launch adapter. They have good news: the 

test was passed with excellent results. 

They install the adapter back on the frame, remove Topsat, tilt the adapter to the horizontal 

and perform test firings of the tilt platforms and separations of China-DMC and S1. 

We all pack up and go back to the hotel to catch some sleep before the early rise and journey. 

MILESTONE 30: The fit-check is successfully completed 

451






8 th May 2005 

SYS_Jörg and ESA_Neil travel back from the fit-check in Omsk, parting company with the 

CST and SSTL delegations in Moscow on the way. 

9 th May 2005 

ARRIVAL 81: The thermal foil arrives at ESTEC. 

ARRIVAL 82: The thermal-vacuum chamber interface is manufactured. 

SYS_Jörg makes new thermal foils 

ESA_Neil takes of all but +y sides and disconnects antenna, routing the coax to the outside of 

the spacecraft. 

ESA_Neil torques bolts behind the foils - they almost all move and will need gluing 

SYS_Jörg applies the foils around the t-pods 

452






ESA_Neil makes harness for the chamber (9-pin, 15-pin ext, 25-pin EPS) 

ESA_Neil applies antenna caps and kapton shields over them. 

ESA_Ceaser and ESA_Bernd make external harness. 

ESA_Jörg makes foil covers around thrusters. 

453






We place thermistors in the following locations and numberings, routing the harness with 

kapton tape. 

Thermocouple Location Tnop Max Tnop Min Top Max Top Min 

1 S-Band +65 -30 +50 -20 

2 +x+y shear 

3 PIN 

4 UHF +85 -35 +60 -20 

5 +x0y shear 

6 Battery +60 -10 +50 0 

7 PCU +125 -10 +70 0 

8 -y shear -10 (PMS) 

9 MGT Driver +125 -40 +85 -40 

10 Top plate 

11 Camera +150 -55 +40 0 

12 +x sun sensor +100 -40 +65 -40 

13 -y lateral 

14 +x lateral -20 (ant) 

15 -x lateral -20 (ant) 

16 ASAP 

17 Adapter 

18 Baseplate 

* all unspecified items are Tmax = +100 and Tmin = -50 

We place all sides on apart from –y. SYS_Jörg applies foil over the gaps under the pods and 

across the +y access hole. 

454






ESA_Neil prepares antenna for outside chamber and puts laterals on and torques the external 

bolts. 

The spacecraft is moved to the TV chamber facility, lifted on a crane, the lower foil is applied 

and the chamber adapter is bolted into place. 

455






PROBLEM 280: The spacecraft is very difficult to turn and it is supported manually by 

SSY_Jörg and ESA_Neil on one top corner during most of the process. This is dangerous and 

a better system should be found for the return to the vertical. 

PROBLEM 281: The fork lift impacted on the -x panel. It is not clear whether any damage 

has been done since viewing access is restricted. It is probably ok, but more by luck than 

judgement. 

The spacecraft is clamped to the fork lift, inserted into the chamber, the cable is hooked on 

and kept tight as we move forwards. The alignment is difficult and takes several attempts, but 

eventually it slides into place. 

Once the horizontal insertion is adequate the cable is secured to the ASAP bolts, and the 

clamps removed. The spacecraft does not drop at all, and the fork lift is removed. 

456






The external harness is buzzed through and connected to the outside of the chamber. 

A simple Integrated System Check is performed and passed without issue. 

MILESTONE 31: The spacecraft is ready for thermal-vacuum testing 

Then the chamber is closed and left to pump down overnight. 

10 th May 2005 

The chamber has pumped down overnight to 6.1x10-5 millibar. Initial spectrometer readings 

show normal water, nitrogen, oxygen, etc, but also a strange peak at approximately 100 AMU. 

This soon disappears though. 

An Integrated System Check is carried out via external power feeding after the timers. 

Everything seems fine, apart from an unexpected reboot of the OBC. 

457






We start raising the shroud temperature and watch the spacecraft responses. It seems that the 

laterals are quite decoupled, but everything is approximately as predicted in the model. 

PROBLEM 282: There are lots of strange mass spectrometer readings, many with 

inexplicably high AMUs. Possibly magnetorquers, tank coating or attenuation caps. The 

pressure rises to around 6.6x10-4 mill bar. 

We wait until UHF and S-Band are approaching 40 degrees, then do a Integrated Platform 

Check. All platform systems appear to work fine. In addition MAGIC powers up high and 

low sides without a problem. 

In an attempt to rectify problem 282 we decide to leave the chamber at a high temperature for 

a long time to “bake out” and reduce out gassing. 

The bake out is taking a long time, and mass spec cannot be used during higher pressure. So 

we decide to leave overnight at a safe temperature with the hope of completing bake out by 

tomorrow morning and then starting cycling again. 

The battery voltage reads at 24.56V, which is healthy. 

At 20:30 the pressure has fallen back to 6x10-5 mille bar, which is encouraging. 

11 th May 2005 

Pressure is low enough to use mass spectrometer. Still have strange peak at 100, and another 

at 38 AMU. 

Ramp temperature down to –40. Pressure fluctuates from 2.1x10-5 to about 3.5x10-5 – this is 

strange and seems to imply very small leakage. 

NOTE: From here onwards Integrated System Checks are referred to as ISCs and numbered 

sequentially. The detailed procedure is given at the start of this logbook as a copy of the form 

filled in hardcopy for each test. These are to be stored with the hardcopy of the TV test data. 

Perform ICS1. All fine except: 

- OBC rebooted occasionally, seems to be linked to high power consumptions. Perhaps this 

power supply is not good enough for the whole spacecraft! 

- Need to know the actual locations of thermistors in MAGIC. Reported 7E 7F 7D 2E, the 

last one seems very low. 

458






- S-Band PA power reported at 1.58 initially and then dropped to 0.7 for next two bursts. 

Need to check this again next time. Temperatures were perfect. 

- Could not measure frequency as did not have counter with us -> next time. 

Otherwise the functional check seems fine. 

We receive confirmation from EPS_Fulvio that the battery is totally disconnected and cannot 

charge or discharge without he ABF plug. Also, while the battery is diconencted no 

temperature or voltage readings can be taken from it, so this is why the TM is empty. This 

means that the PCU is assuming that the battery is broken during external powering. 

NOTE: There are slight temperature peaks on almost all thermistors during the UHF 

transmissions. Some of these are physically impossibly and must be due to EMI, probably 

because many of the thermocouples are not isolated from the structure. 

NOTE: Perhaps the OBC reboot problems could be due to current flow across the OBC 

ground plane, since we are powering the whole spacecraft across it. In hindsight this is not 

such a good idea. The reboot problems are NOT due to EPS (for a change) since they are too 

fast for the software flow in EPS. 

PENDING MODIFICATION: Add another ground wire to the FPP. 

We continue allowing the spacecraft to cool down. The shroud is at –35 degrees. 

Once the lateral panels were at –10 degrees we perform an ICS and start to ramp the 

temperature back up again. 

ICS2 is all fine except: 

- OBC rebooted on power up of ACDS. Perhaps this is current limiting from power supply, 

maybe because lacquer has out-gassed and coils have shorted. 

- Need to know the actual locations of thermistors in MAGIC. Reported 53 53 56 30, the last 

one seems very low. 

- S-Band PA power reported at about 0.8W consistently. 

- Could not measure frequencies as did not have counter with us -> next time. 

- Current consumptions of UHF and S-Band were a little higher than usual. 

Temperature begins to stabilise in the second hot case. It is too hot to perform a full system 

check, but a simple platform check, ICS3 is performed and passed, apart from: 

459






PROBLEM 283: Trying command S-Band up results in a jump in signal but no carrier, 

however, the current consumption continues as if it is up. Temperature in S-Band unit jumps 

up by a few degrees. 

ACTION: Turn off spacecraft ASAP. The temperatures drop straight away, implying some 

EMI as before. 

Frequencies are measured as: 437.25003 MHz and 2401.817 MHz. 

Ramp down for cold case. Perform ISC as we pass 25 degrees on top of S-Band and 18 

degrees on the panel. New ICS, 1645. 

Everything fine apart from S-Band being at power 0.8W via DTMF. The S-Band carrier 

seems to work properly again – demonstrating that problem 283 is certainly temperature 

dependant. 

We set the shroud to ramp up to 55 degrees, wait for the system to settle, and then leave it 

overnight to bake out further. 

12 th May 2005 

The whole spacecraft has reached around 55 degrees. We raise the shroud to 65 degrees for a 

while to satisfy the SSTL requirements, and then ramp it down to –80 degrees and wait for the 

spacecraft to respond. 

We do ISC5 when S-Band has dropped to 45 degrees. The ISC works fine, apart from: 

- S-Band cannot be commanded up at 43.8 deg, but it can at 39.6 deg. Ten DTMF 

bursts are then recorded, along with lid temperatures and passed to the AMSAT team 

for analysis. 

Thermistor data recorded and compared to the thermocouples for calibration purposes. 

S-Band transmits an entire picture without problems. 

We continue to ramp down to the cold case and let the spacecraft settle for a while. 

ICS6 is performed and has no problems (save the low PA power report), half a dozen DTMF 

bursts are recorded. 

The chamber is set ramping up for next hot case. 

460






As it ramps up we carry out a normal ISC (7) with no problems, this test included four DTMF 

bursts with the carrier not permanently up. 

Extended S-Band tests were performed in order to properly characterise problem 283. The 

details can be found in DTMFFFT20050512 ISC7.csv. 

We then left s-band with the carrier up and DTMF TM turned on and recorded the DTMF 

data, the lid temp, and occasionally the current consumption and the frequency. We also 

toggled the carrier down and back up again after each DTMF burst, the success of this is 

shown in the "toggle" column of DTMFFFT20050512 ISC7.csv 

The PA power telemetry, the current consumption and the frequency all dropped together as 

the temperature rose. At 44.9 degrees the carrier failed to come up, and problem 283 was 

reproduced - the current consumption was "normal". Leaving it to cool for a minute allowed 

the carrier to come up on the next attempt, but then it dropped spontaneously, the box was at 

45.3 degrees. Higher than that it didn't come up at all. 

The problem was therefore fully characterised and all data passed on to the S-Band team for 

analysis and recommendations. 

We allow the chamber to reach the required 65 degrees and to settle, and then start to ramp 

down to the fourth, and last, cold case. We perform ISC8 on the way down, experiencing no 

problems. 

The chamber is set to hold the shroud at –80 while the spacecraft responds, and then ramp to – 

40 for an hour to let it settle. ISC9 is carried out, and then the spacecraft and groundstation 

are left on as an endurance test, but with a timer switch to power down the spacecraft at 3am. 

Afterwards the chamber will ramp up to +42 and remain there for the rest of the night. 

Therefore, once the endurance tests ends at 3am, we should have at least 5 hours of dwell time 

for the whole spacecraft to reach 42 degrees as a good reference point for dissipation tests 

tomorrow morning. 

13 th May 2005 

Assess endurance test: the boot attempt has incremented from 121 to 161. This is about the 

same as last time it was done on external power. 

We perform ISC 10, it works fine apart from a couple of reboots. Take TCS readings early 

for calibration purposes. 

ESA_Neil removes panel loops from EPS ABF, but leaves battery line in (pins 1-3). The 

Keithly is set up as panel simulator, 750mA and 30V. Connected across pin 4 of EPS ABF to 

461






simulate panel 1, and the ground of OBC. Ammeter and voltmeter used to verify Keithly 

output. 

We power up internally and perform ISC 11. This seems much more stable - no reboots, even 

when toggling ACDS power. 

So, we start dissipation tests: 

- On internal power 

- The s-band carrier is brought up 

- UHF live stream 

- ACDS TM every 6 seconds 

- S-band carrier drops once, but can be brought back up 

NOTE: S-band so near the temperature threshold that data transmission causes it to drop. But 

DTMF tones are transmitted fine. 

Eventually the S-band carrier drops and fails to come up. Last temperature readout on lid = 

42.2, from DTMF = 47.3. 

We repeatedly command large thumbnails down UHF to heat it 

PROBLEM 284: The spacecraft stops responding, ESA_Neil debugs down the OBC port, it 

seems fine and responds to internal commands to transmit. This implies that the UHF TNC 

has locked up in the receive direction. 

Using the debug port we cycle UHF power, it then comes up fine and transmits pictures as 

asked. 

S-band comes up but is a bit shaky and drops again. After a couple of tries it can't come up. 

ESA_Neil turns on the Keithly to simulate a solar panel. The battery is initially at 23.304V. 

Consumption should be around 360mA @ 28V, so simulating a panel at 750mA should give 

around 390mA of charging current, equivalent to 11.2W. 

Reading of battery voltage is not accurate when charging, but we can see that it is rising so the 

panel simulation is working. 

If simulate the panel over 825mA then can see the supply switch to a voltage source instead of 

a current source. This is because it is more than 500mA above the consumption current, 

which is the maximum current that the BCR can draw 

Transmit full picture, start at 6100 packets -> 26200 packets. About half an hour. 

462






PROBLEM 285: UHF does not seem to be transmitting constantly... it is incrementing by 

about 10 packets at a time, and only up about 50%. 

Debug: getting a lot of fetch-queue timeouts from UHF. Re-setting TX delays only confuses 

things. Power cycling fixes it. But UHF locks up too easily when asked to transmit 

constantly. 

Dissipations tests are complete at this temperature. 

Meanwhile, in the UK, AMS_Howard performs some thermal tests with the “flat sat” 

engineering model of the S-Band unit. He finds that this model suffers from the same 

problem (283), but at the higher temperature of 63 degrees Celsius. This is probably due to 

the fact that this model has not been tuned for good modulation (it is a trade-off). 

We ramp down the shroud to -50 to cool the spacecraft, and power it off. 

We do not have time to wait for a steady state, but choose a moment when we believe that the 

positive offset of the s-band unit will be compensated for by the fact the rest of the spacecraft 

is still cooling down. 

ISC12 is carried out, and goes fine. NOTE: Carrier up the whole time. 

After twenty minutes of carrier up the S-band unit temperature levels out at about 41 degrees. 

Manage to get UHF to lock up again - this implies that it is not a thermal problem (was at 

about 30 degrees). However, we cannot unlock it, as we are grounding through debug port. 

We just leave it for now and concentrate on s-band tests. 

In fact, term had locked up. UHF is fine. 

463






S-band seems to settle at about 8 degrees above the +x+y shear panel. 

We pressurise the chamber over about half an hour and remove the spacecraft carefully to 

carry out a visual inspection, clamping it securely to the fork-lift. 

- The damage inflicted on the -x lateral panel during loading is superficial 

- The solar cells are all intact 

- The Kapton coating looks fine 

- The camera looks fine 

- No immediate issues 

ESA_Ceaser, ESA_Neil and SYS_Jörg attach the lifting frame to a crane to support the 

spacecraft, and then gradually raise it back to the vertical by lowering the forks, raising the 

crane and sliding the spacecraft. This works fine. 

464






SYS_Jörg applies some extra kapton tape to the baseplate inside the ASAP ring, then 

ESA_Neil and SYS_Jörg pack up and move the spacecraft and ground equipment back to the 

cleanroom. 

We then cover the top-plate of the spacecraft in kapton tape, lift it to the integration table and 

apply some kapton tape to the baseplate also. 

16 th May 2005 

ARRIVAL 83: The flight battery arrives at ESTEC. 

ARRIVAL 84: The replacement e-box arrives at ESTEC. 

465






ESA_Neil removes the +y lateral panel and the +y T-Pod. 

MODIFICATION 138: ESA_Neil follows the instructions from CANX on how to replace 

the damaged E-box: 

Remove the Teflon heater cable from its connector on the E-box. 

Done 

Remove the pusher plate from the T-POD by cutting the zip-tie that attaches it to the main 

spring. 

Done 

Remove the old zip-tie that holds the main spring and Teflon block to the bottom plate of the 

T-POD. 

Cannot until the ebox is removed 

If not already done, separate the old E-box from the T-POD by removing the four screws 

holding it to the bottom plate of the T-POD. 

466






Done 

With the old E-box that was attached to the T-POD, remove the four #4-40 screws holding 

down the E-box cover and the two hex head female screw locks for a DB15 connector. 

Done 

Take off the E-box cover. 

Done 

Fold out the small side of the E-box base at bottom. (If you have a small vice that can grip the 

small side of the E-box base, it will be easier to fold it out.) 

Fold it out? This is not a good design. Done 

Take out the old electronic board from the E-box base. 

Done 

467






Repeat steps 5 through 8 with the new E-box. 

Done 

Replace the base on the new E-box with the base that has a slot cut out for the zip-tie 

connecting the Teflon block inside the T-POD to the bottom panel of the T-POD. 

Done 

Put the electronics board in the base that has the zip-tie cutout. 

Done 

Put the cover on the new E-box assembly and perform a fit check. The two covers from both 

old and new E-boxes should be interchangeable. However, decide which cover fits better than 

other by performing fit checks. 

Done – the new one fits fine 

Screw down the four #4-40 screws to attach the cover to the electronics board. Make sure 

that there is enough clearance between mounting holes and standoffs on the bottom side of 

the E-box. 

Done 

Remove the four screws and the cover. 

Done 

468






Fold in the small side of the base in a similar fashion as in step 7. (If you have a problem of 

folding the small side in, you can leave it as it is or cut it out, since we are not using this side 

for mounting the cover.) 

PROBLEM 286: The e-box tab snaps off due to work hardening. 

CANX_Fred advises that it is not necessary anyway. 

Put the cover on the electronics board and attach it using four #4-40 screws and two female 

screw locks. 

Done 

Apply Kapton tape to the mounting holes on the top and side of the E-box. 

Done 

At this point, the casing from the old E-box should contain the new electronics board. The 

swap was necessary so that the zip-tie used to hold down the Teflon block inside the T-POD 

does not interfere with the E-box casing. Use four #4-40 screws and lock washers to close the 

new electronic board within the old E-box casing. 

Done 

Perform the functional test. 

The box is only at 10.5V, so it must be charged first. 

469






Ensure that the E-box has functioned properly before continuing with the assembly. 

Cannot really until it is attached to the heater block again (as functional test details). 

Place the Teflon block and main spring back into the T-POD. Make sure that the holes bored 

out of the side of the Teflon block line up with the screw holes for the E-box to minimize 

interference between the Teflon block and the screw head. 

The Teflon block is securely glued in already. Done. 

Re-attach the Teflon block and main spring to the T-POD by looping the new zip-tie down 

through one hole in the Teflon block, down through the first hole of the bottom plate of the T- 

POD, up through the second hole in the bottom plate, and finally up through the second hole 

in the Teflon block. The locking end of the zip-tie should be inside the T-POD on top of the 

Teflon block as to not cause interference with any other parts. Do not tighten the zip-tie at 

this point. 

Done 

Ensure that main spring is sitting on the Teflon block such that the cross-member end of the 

spring can be looped through the zip-tie once it is tightened down. 

Done 

Start to feed the end of the zip-tie through the locking end while leaving some slack to allow 

some freedom when threading the screws into the E-box. 

Done 

Hold the E-box on the bottom of the T-POD and align the holes in the outer casing to the 

holes in the bottom panel of the T-POD. Ensure that the serial and connector ports are 

facing the front (the side with the port hole) of the T-POD. 

470






Done 

With the E-box held in place, thread the four screws into the holes from inside the T-POD and 

tighten them down as firmly as possible. The screw sizes are #4-40 and ensure to use a lock 

washer between the screw head and bottom panel of the T-POD. 

Done 

Once the screws are firmly in place and holding the E-box to the bottom of the T-POD, 

tighten the zip-tie running through the Teflon block so that the Teflon block and main spring 

are held firmly in place. (Note: there will be some freedom in the movement of the main 

spring, but this should not be a concern, as the pusher plate will hold it in place.) 

Done 

Cut off the end of the zip-tie to remove unnecessary material. 

Done 

Stake the screws holding the E-box to the T-POD with either an RTV or epoxy to ensure that 

the screws are not loosened during vibration testing or launch. 

“Stake”? I assume you mean “stick”. Done. (There will be no more vibration testing, you 

are too late.) 

Re-attach the pusher plate to the main spring via a zip-tie by looping it around cross members 

of the pusher plate and the cross member end of the main spring. 

Done 

Cut off the end of the zip-tie to remove unnecessary material. 

Done. 

After the E-box is attached to T-POD, re-tightened the four #4-40 screws located on the top of 

the E-box casing and two female screw locks for DB15 connector. Then, stake them with an 

RTV or epoxy. 

Done 

The T-Pod is re-integrated to the structure for testing. 

NOTE: The heater block cable fits rather unconvincingly into the socket on the new pod, 

although no debris is visibly in the way. 

471






A functional checkout on the re-integrated T-Pod is performed. 

PROBLEM 287: The T-Pod firing fails. 

MODIFICATION 139: ESA_Neil replaces the heater block with one of the spare ones, 

making sure to cover the top of it in kapton tape to insulate the heater elements from the pod 

itself. 

A functional checkout on the re-integrated T-Pod is performed. This time it succeeds – it is a 

bit slow, but that it probably due to low battery charge after the last firing (no time to charge 

it, the Xi-V guys arrive tomorrow). 

ESA_Neil glues the cable in place with Scotch-Weld, re-glues the rubber standoffs that are 

not secure on the pod lid, and then re-integrates the +y lateral panel. 

472






17 th May 2005 

ARRIVAL 85: The Xi-V team arrive with the flight model and all associated ground support 

equipment. 

ESA_Neil charges the +y T-Pod and the spacecraft batteries. 

The Xi-V team perform a functional checkout of Xi-V. 

18 th May 2005 

ESA_Neil loads the +y T-Pod and performs a second firing test with the new E-Box. The test 

is successful, but it takes longer than expected: about 12 seconds instead of 2 seconds. 

The XI-V_Ryu trains ESA_Neil on the Xi-V ground support equipment, battery charging, and 

apply before flight item. The documentation governing these procedures was amended during 

the training and will be uploaded to the FTP in the Cubesat/Xi-V folder when the soft-copy 

has been updated. 

473






The Xi-V team insert Xi-V into the +y T-Pod and perform a fit check on the lid. 

PROBLEM 287: The lid is not quite a tight enough fit to securely depress the activation 

switches during the launch vibrations. 

MODIFICATION 140: We add several layers of Kapton tape on top of the rubber standoffs 

to tighten the fit of Xi-V in the T-Pod. 

474






ESA_Neil secures and loads the +y T-Pod with Xi-V inside. 

PROBLEM 288: A lose bolt is found in the NCube-II pod. It is from the e-box mounting on 

the back left. 

475






The Xi-V team pack up their ground support equipment and supporting hardware. 

ESA_Neil replaces the +y thermal foil and the side protector. 

20 th May 2005 

ESA_Neil, UHF_Holger and AMS_Graham power the satellite on external power, connecting 

the UHF coax line into a ‘Stabilock’ communications test unit. (The line has not been 

reconnected to the flight antenna since before the thermal vacuum test to avoid excessive 

cycling of the connector.) 

AMS_Graham measures the field strength near the patch antennas when the S-Band carrier is 

up. The results are not conclusive and the test should be repeated with the caps off. 

We repeatedly request a small picture downlink while measuring characteristics of the UHF 

transmission: 

- The centre frequency is only 10-20 Hz too high 

- The deviation is 3.2 kHz RMS 

- Data received seems fine 

- The output is measured at 2.45W, which is a bit low, but once the attenuation in the cable is 

accounted for it should be fine 

By transponding via s-band we can measire the UHF sensitivity = -118 dBm. 

476






We try to test audio level of control tone required to bring transponder up - it is VERY 

sensitive, around 125dBm. This is certainly lower than the level at which an audio signal is 

discernable anyway, so it is very good. 

The receiver pass-band looks symmetrical, about 8kHz either side of centre. 

UHF_Holger measures a DC offset 300mV, modulation voltage 1V peak-to-peak from the 

TNC3 on the groundstation. This could be why the radio has to be off-freq, but isn’t a full 

explanation. 

We repeatedly transmit time synchs and vary the transmit frequency of the groundstation. 

The spacecraft responds up until 10kHz high and 12kHz low (exclusive), at -100dBm. 

We then vary the transmission power on the centre frequency. Data reception of the 

spacecraft is good until -115dBm. 

We attempt to reproduce TNC locking problem but fail, three large thumbnails are 

downlinked no problem. UHF_Holger advises that he cannot envisage how such a lock up 

(one direction only) could occur anyway. 

ESA_Neil and AMS_Graham remove the relevant side protectors, remove the antenna caps, 

and test the field strength on s-band. The results are inconclusive, so we must measure on end 

of coax, this means disconnecting one of the antennas. 

NOTE: The emulsion paint on the inside of the antenna caps now appears to be dry. Perhaps 

this a culprit for the out-gassing during the thermal-vacuum tests. 

AMS_Graham and ESA_Neil take off the -y and +x lateral panels, then disconnect the coax 

from the +x antenna and run it into a spectrum analyser. 

ESA_Neil measures the resistance of the magnetorquer coils at 97 and 98 ohms (coil 1 and 

coil 2 respectively), this should be fine and implies that they survived the thermal vacuum 

tests without issue. 

477






PROBLEM 289: We only measure 240mW being delivered to the +x S-Band patch antenna, 

it should be 750mW. 

21 st May 2005 

AMS_Howard and AMS_Graham disintegrates the coax cable from the –3dB port on the S- 

Band box. Then measures the test cables to find their attenuation, test cable = 1.8dB. 

We then measure the power being delivered to the +x antenna (-6dB port) again, this time 

taking into account the loss in the test cable: 478mW. (This is not including the flight coax 

loss.) 

We then disintegrate and measure the top (-3dB) port at: 1.12W. This measurement is done 

directly on the port and does not include flight coax. The test coax is taken into account. 

When previously measured this was 1.45W. But maybe the difference can be put down to 

experimental error. 

Using a heat-gun we start to heat the S-Band unit and watch the power drop. Several power 

readings on DTMF are compared to actual measurements, we now have several points to 

calibrate the readings. 

Deg C P (Measured) Old ADC New 

19 0.120 -0.4909 25 -0.00217 

21 0.191 -0.109 32 0.19055 

22.6 0.275 0.1091 36 0.30067 

24.3 0.407 0.4363 42 0.46585 

25.1 0.490 0.6545 46 0.57597 

25.8 0.575 0.8182 49 0.65856 

26.25 0.638 0.9818 52 0.74115 

26.6 0.692 1.036 53 0.76868 

26.8 0.724 1.145 55 0.82375 

478






27 0.759 1.145 55 0.82375 

27.1 0.776 1.2 56 0.85128 

27.2 0.794 1.255 57 0.87881 

27.25 0.804 1.309 58 0.90634 

27.34 0.820 1.309 58 0.90634 

27.41 0.834 1.309 58 0.90634 

28.5 1.072 1.636 64 1.07152 

AMS_Howard calibrates the DTMF telemetry readings and defines the following parameters 

as correct: 

Offset = -0.690 

Multiplier = 0.02753 

At just over 45 degrees on DTMF readings the S-Band carrier fails, the external temp is also 

about 45 degrees. 

SUMMARY: The slightly low power is not a big enough problem to take the box apart, it is 

just annoying. The temperature problem might be more serious, the AMSAT team will 

internally discuss the possibilities to rectify it and discuss with SYS and Management later 

one. The final deadline for fixing, acceptance testing, reintegrating and functional testing the 

S-Band unit would be the 20 th June. 

We reconnect the UHF flight antenna and do field strength tests. The field strength meter 

locks on frequency at something like 30cm from the antenna (range switch on 3GHz position, 

and dummy load used instead of antenna). 

The power levels looks fine - everything the antenna is getting, it is radiating. The second 

harmonic is -60dB down from carrier, the third is -76dB, none of the others are present. 

AMS_Holger declares UHF ready to fly. 

479






24 th May 2005 

ARRIVAL 86: The UWE-1 flight model arrives with UWE_Yohko and UWE_Radu. 

The UWE-1 team unpack the spacecraft and perform a visual inspection. The kill-switch 

improvements are significant. 

ESA_Neil prepares the +x T-Pod, removing the engineering model of UWE-1. 

The UWE-1 team train ESA_Neil on the charging and flight preparation procedures. They 

are simple and unproblematic. 

UWE-1 weighs in at 977g: 

480






We perform a fit-check with the spacecraft into the T-Pod. The lateral fit is good and the 

access hole is adequate for charging and pre-flight preparation procedures. 

PROBLEM 290: One rubber standoff has come loose from the +x T-Pod. 

PROBLEM 291: The standoffs do not quite hold the spacecraft tight enough (the push plate 

can still be moved by hand through the access port). 

MODIFICATION 141: ESA_Neil glues the loose rubber standoff back into position on the 

inside of the lid of the +x T-Pod. 

We wait for the glue to dry. 

25 th May 2005 

We perform a fit-check with the UWE-1 spacecraft into the T-Pod. The longitudinal fit is 

now better, but still requires some tuning. 

481






MODIFICATION 142: ESA_Neil uses small squares of 5 layers of kapton tape to thicken 

the rubber standoffs in the +x T-Pod. 

The fit is now good. UWE-1 is loaded into the pod, the pod is closed and armed. 

UWE_Yohko and UWE_Radu depart. 

26 th May 2005 

The post-thermal-vacuum-test ACDS functional checkout is performed following instructions 

from ACDS_Lars: 

1. Boot up the spacecraft 

The spacecraft is booted up 

- safe beacon detected 

- nominal beacon detected 

- EPS temp = 17,23 deg 

- OBC temp = 19.3 deg 

- EPS voltage = 24.116 V (oscillating +/- 0.1) 

- OBC boot attempts = 29 

2. Upon initialisation ACDS puts a message in the alarm stack. Download the alarms and 

write down the message. 

Get alarms: 


482








0x00000010: 61 74 69 6F 6E |ation...........| 

3. Keep an eye on the beacons, within a few mintues: 

- The magnetometer reading settles at a non-zero level 

- The ACDS "field rates" settles at zero. 

after a few minutes the following ACDS data is stable in the beacon: 


Subsystem: BEACON: 

Callsign: www.sseti.net * SSETI-Express 

On board time: 26-05-2005 14:48:04 

EPS battery voltage: 24.073 V 

EPS Temperature: 16.38 °C 

OBC temperature: 19.7 °C 





- The ACDS driver board is powered 

- ACDS driverboard generates 12V for the magnetometer 

- Communication between magnetometer and OBC is ok. 

- ACDS code on OBC executes nominally 

We live stream down UHF 

4. Issue the telecommand: ADS_MODE_SELECT with param1=1 and param2=10 

(0x0a 0x01 0x01 0x0a) 

Command issued 

5. Wait 30s and downlink the telemetry stream. Verify: 

- The interval between ACDS data falls from 20 seconds before the 

telecommands to 6 seconds after telecommand has been issued 

Data is every 6 seconds 


- The ACDS software react to telecommands (thus also a test of the 

complete UHF->OBC->ACDS chain) 

6. Find something magnetic and place it close to the spacecraft (alternatively then rotate the 

spacecraft around Z). Observe from the beacons that: 

- The magnetometer reading changes 

483






- The "field derivatives" increases for a time and then resettles at zero 

Spacecraft turned slowly approximately 180 degrees in +z direction 

On board time: 26-05-2005 14:56:52 



On board time: 26-05-2005 14:57:10 



On board time: 26-05-2005 14:57:28 



On board time: 26-05-2005 14:57:46 

ACDS Magnetometer: [1327 3866 7915] nT 

ACDS field rate: [127 -32 14] nT/s 

On board time: 26-05-2005 14:58:04 


ACDS field rate: [127 -127 5] nT/s 

Completed 180 degrees and then reverse direction of turn to –z 

On board time: 26-05-2005 14:58:22 



On board time: 26-05-2005 14:58:40 



On board time: 26-05-2005 14:58:58 


ACDS field rate: [-127 127 -18] nT/s 

Completed turn, spacecraft is back in its original position. 

On board time: 26-05-2005 14:59:17 



On board time: 26-05-2005 14:59:35 



On board time: 26-05-2005 14:59:53 


ACDS field rate: [-29 7 0] nT/s 

On board time: 26-05-2005 15:00:11 



On board time: 26-05-2005 15:00:29 



The above further verified that the magnetometer is alive. 

7. Shade for the X+ sunsensor and downlink the telemetry stream. Copy 

one ACDS houskeeping packet (from when the sensor was shaded) to 

the open file and designate the packet X+1 

484






The sunsensors are on the Y sides. Here is a shaded (by cover and tissue) +Y packet: 



0x00000000: 25 FA 51 0E 94 1E F9 05 F1 05 E6 05 EB 05 00 00 |%.Q.............| 

0x00000010: 05 00 3D 02 F4 01 20 02 D2 00 00 00 00 00 00 00 |..=... .........| 



0x00000000: 25 FA 51 0E 94 1E 00 00 00 00 00 00 |%.Q.............| 

8. Direct a spotlight at the X+ sensor (lamp ~3cm from the 

sensor). and downlink the telemetry stream. Copy one ACDS 

houskeeping packet (from when the sensor was shaded) to the open 

file and designate the packet X+2 

Here is a lit (by halogen lamp) +Y packet: 



0x00000000: 24 FA 50 0E 98 1E EB 05 F9 05 47 06 16 05 00 00 |$.P.......G.....| 

0x00000010: 02 06 3D 02 F4 01 26 02 E5 00 00 00 00 00 00 00 |..=...&.........| 



0x00000000: 24 FA 50 0E 98 1E 00 00 00 00 00 00 |$.P.............| 

+Y sensor light is removed and cover is replaced 

9. Repeat the above for the X- sensor, designate packets: X-1 and X-2 

Here is a shaded (by cover and tissue) -Y packet: 



0x00000000: 3E 07 9E 00 F0 1E F7 05 F7 05 F7 05 E0 05 00 00 |>...............| 

0x00000010: 00 00 C9 02 80 02 1E 02 D2 00 0B F2 00 00 00 00 |................| 



0x00000000: 3E 07 9E 00 F0 1E 0B F2 00 00 00 00 |>...............| 

Here is a lit (by halogen lamp) -y packet: 



0x00000000: 40 07 9E 00 F0 1E 12 06 18 06 E3 05 F1 05 D8 05 |@...............| 

0x00000010: 05 00 C9 02 87 02 26 02 F0 00 00 00 00 00 00 00 |......&.........| 



0x00000000: 40 07 9E 00 F0 1E 00 00 00 00 00 00 |@...............| 

When analysed by Lars the above will verify: 

- That the intensity measurements from each sensor works. 

10. Send the datafile to Lars for further analysis. 

485






Done. 

ACDS_Lars responds declaring that ACDS is ready for flight. 

ESA_Neil turns on CAM and pings it, a pong is received. 

ESA_Neil turns on magic, CAN is received. Ask for pressures from MAGIC: 


Subsystem: MAGIC | MID: 0x01 | Length: 6 | Time: 26-05-2005 15:21:42 

0x00000000: 2B 00 8F 00 A3 00 

This is equivalent to around 2 bar on PT1, 1.37 bar on PT2 and 1.05 bar on PT3. These are 

all within acceptable limits and seem sensible, however, the PT3 reading during thermal 

vacuum was showing around 0 bar. 

PROBLEM 292: It appears that the low-pressure tubing leaks, since the reading of pressure 

transducer three was 0 bar when the spacecraft was in the thermal vacuum chamber, and 1 bar 

when back at ambient pressure. It is not clear how serious this problem is. 

Ask for temperatures from MAGIC: 


Subsystem: MAGIC | MID: 0x11 | Length: 4 | Time: 26-05-2005 15:22:17 

0x00000000: 61 5F 63 30 |a_c0............| 

These correspond to 19.43 degrees, 18.34 degrees, 20.52 degrees and an unknown (but low) 

temperature. However this 4 th temperature corresponds to the missing thermistor after the 

MAGIC processor replacement, so it is just reading noise. The others are within acceptable 

limits. 

ESA_Neil turns off CAM and MAGIC, puts the S-Band carrier up, and logs beacons until the 

spacecraft descends into safe mode. 

The spacecraft enters safe mode without issue after a few hours. 

ESA_Neil leaves the spacecraft charging over night. 

30 th May 2005 

ESA_Neil powers up s/c at 1514, and waits for timers while setting up spare laptop as 

groundstation. 

PROBLEM 293: The spacecraft starts up at 1619, only 65 minutes instead of 74. 

486






ESA_Neil sets up an overhead projector to shine on the lateral panels. Solar panels reads at 

about 32.4V when OHP is shining on them 

When connecting panel to spacecraft (in FPP via ammeter), voltage drops to 24.5V, it is 

sourcing 27mA only (+y panel) in nominal mode. This is about 0.5V above the battery. 

The panel input to the PCU is varied by moving the connection down the FPP. No difference 

is encountered – therefore implying that all four panels are wired identically inside the PCU. 

The voltage on the +x panel is 23.9V in ambient light, then 2.9V when connected and 

sourcing 0mA. 

ESA_Neil turns the spacecraft and watches the voltage and current from +x increase. When 

fully lit it is at 24.3V and sourcing 13mA. When disconnected it shows 32.13V (through side 

protector). At 45 degrees the voltage is 30.2V, at 0 degrees is 24.3V. 

When switching OHP to "dim" can see current consumption drop by 7mA. It doesn't take 

much shading to make it drop right off to zero as the voltage goes below the battery voltage. 

Clearly the OHP gives nowhere near enough light intensity to power the spacecraft. Upon 

reflection this is quite sensible, and the only reason that the AAUCubesat team was able to 

use one for this task was because they were using the whole “light cone” of an OHP to light 

the side of a Cubesat, with only 2 cells. We are providing almost the same power (since cone 

is fixed) but over 45 cells. 

ESA_Neil performs part of the OBC checkout, following instructions given by OBC_Karl in 

the OBC functional checkout procedure. 

Lines starting with "DO >" indicate actions to be performed by the 

test personal. Lines starting with "RES>" indicate the desired result 

of the previously stated action. If one or more "RES>" are not 

487






fulfilled the OBC functional test has failed. It is up to management 

to decide if failing certain tests will result in total rejection of 

OBC or if partial test success is allowed. 

Parameters to telecommands will be in the following form: OBC_TM_MODE_3(param1, 

param2). 

Please, read entire document before initiating any tests. If in doubt 

about any part of the procedures, do not hesitate to contact the OBC 

team coordinator Karl Kaas Laursen by phone or e-mail. 

Be sure to download the latest OBC ICD from the SSETI ftp server before 

undertaking any test activity!! 

Preparing the test setup: 

--------------------------- 

The functional test should be as independent of other subsystems as 

possible. Therefore, the test procedure uses the debugging interface 

for initial confirmation of OBC hardware integrety and only EPS is functional. 

DO > Power up the OBC. 

DO > Connect a PC to the debugging interface at the FPP using null-modem. 

DO > Open a serial terminal program and set baud to 57k6 (8N1). 

Done. 

OBC-only checkout tests: 

---------------------------- 

*** Test 1: Debugging screen *** 

DO > Type in the terminal program. 

RES> The debugging screen should appear in the terminal program. 

Test 1: Passed 

*** Test 2: Flash integrity *** 

DO > type "ferase" to erase FLASH. 

DO > Type "fgeni". 

RES> A new FLASH integrity field is generated. 

DO > Type "fcheck". 

RES> Debugger replies: "flash crc check ok, flash is 'good'". 

Test 2: Problematic! The flash erase does work (files are wiped and counter reset), but the 

computer is powered down and back up again (not gracefully). Also note that procedure 

488






should say "ferase!" instead of just "ferase". "fcheck" was succesful. Please 

advise if this is ok or not. 

*** Test 3: EPS ping *** 

DO > Wait for 2 minutes. 

RES> If the OBC hasn't been rebooted then the EPS ping works. The ping should also be 

displayed in the debugging interface. 


*** Test 4: Kill WD *** 

DO > Type "wdtog". 

RES> OBC should reboot within 120 seconds. 


Context dependent tests: 

--------------------------- 

For the context dependent tests the UHF must be operational and a 

ground station must be set up. Use Term.exe to issue commands and 

review replies for the following tests. The debugger must still be 

hooked up to the debugging interface during these tests. 

DO > Set up Term to use the proper serial port. 

DO > Hook up Term to the ground station TNC. 

Done. 

*** Test 5: Beacon *** 

DO > Power up spacecraft and allow it to go to nominal mode. 

DO > Wait for nominal mode beacons to appear in Term. 

RES> Correct-looking nominal mode beacons should appear every 18 sec. 


*** Test 6: Sync time *** 

DO > Press sync time in Term 

DO > Wait for next beacon 

RES> Next beacon should contain the correct time. 


*** Test 7: OBC uptime *** 

DO > Send OBC_GET_UPTIME to OBC 

489






RES> The OBC should reply with the current uptime in Term. 


*** Test 8: TCS sampling *** 

DO > nothing 

DO > Wait for two minutes then send OBC_GET_HK to OBC. 

RES> If the subsystem worked the temperatures in the spacecraft could be seen in the 

telemetry. 


*** Test 9: Test shutdown through UHF *** 

DO > send OBC_DO_SHUTDOWN to obc 

RES> The OBC should shut down 


*** Test 10: Get houskeeping *** 

DO > Send OBC_GET_HK(0x0A,0x04) to OBC 

RES> 10 housekeeping packages stored on the OBC, if there is 10, should appear in Term. 


*** Test 11: Get alarms *** 

DO > Send OBC_GET_AL(0x0A,0x04) to OBC 

RES> 10 alarms stored on the OBC, if there is 10, should appear in Term. 


*** Test 12: Non-public telecommands *** 

DO > Turn off encryption in Term. 

DO > Send the non-public telecommand, OBC_SYNC_TIME. 

RES> TM packet from OBC with MID OBC_TC_CRYPT_ERROR (0x40) is recived by Term 

within 5 sec. 

Test 12: Passed, however, please note that the OBC ICD has OBC_TIME_SYNCH 

erroneously marked as public telecommand 

*** Test 13: Public telecommands with public commands disabled *** 

DO > Do this test with all OBC telecommands marked *PUBLIC* in the OBC ICD. 

DO > Send public telecommand without encrytion using Term. The first parameter must not 

be 0xFF in this test. 


within 5 sec. 

490






Test 13.1: Passed, however please note that the live stream command SHOULD take FF as 

the first parameter (hams will not be interested in turning it OFF, only turning it ON). Also, 

please note that the OBC ICD does not have the full picture downlink as a public 

telecommand, but it should have. 

*** Test 13: Public telecommands *** 


DO > Turn on public access: Send OBC_PUBLIC_ACCESS(0xFF,0x00) and wait 1 second. 

DO > Send public telecommand without encrytion using Term. The first parameter must not 


RES> A normal acknowledge is received by term. 

Test 13.2: Passed, however please note that the live stream command SHOULD take FF as 

the first parameter (hams will not be interested in turning it OFF, only turning it ON). Also, it 

would be prudent at this point to check that ONLY the public telecommands can be used 

without encryption, so please change the document to reflect this. 

Note: OBC_Karl to please update the document so that there are not TWO "test 13"s. 

*** Test 14: Public telecommands with bad parameter *** 

DO > Do test 13 with the first paramter set to 0xFF in all commands. 


within 5 sec. 

PROBLEM 294: Test 14: FAILED!! All public telecommands work even if the first 

parameter FF is used. This is potentially very dangerous as a command to downlink a full 

picture on UHF would effectively lock us out of the spacecraft for 30 minutes. OBC_Karl 

advise on reasons and possible solutions. 

Test 15: Not conducted. Please clarify how we can tell if the magnetorquers are used (if this 

should come from ACDS then please ask him to answer here). Should we assume that the 

ACDS operational mode is given in the alarm stack? 

*** Test 16: Commands with verify bit *** 

DO > Send OBC_GET_HK(0x01,0x04) with verify bit (press the "V" button in Term in stead 

of "Send") 

DO > Wait for 1 second. 

DO > Get alarms: Send OBC_GET_AL(0xFF,0x04); 

RES> The alarm data is received by Term and one entry has the previously sent TC 

embedded in the data section. The MID of the alarm entry is 0x95. 


*** Test 17: Flushing TM *** 

491






DO > Send OBC_FLUSH_HK(0x00,0x00). 

RES> A TM packet with MID OBC_HK_FLUSHED is received with the number of TM 

packets in the queue in the data field. 

DO > Send OBC_FLUSH_HK(0xA0,0x00). 


packets in the queue in the data field. This number should be appr. 10 smaller than before. 

DO > Repeat test with OBC_FLUSH_AL. 


*** Test 18: Flush flight plan *** 

DO > Send 10 commands to the OBC to be executed in 500 sec. OBC_GET_HK(0x00,0x04). 

DO > In the debugger, type "plist". 

RES> 10 commands are in the list appearing on the debugging screen. 

DO > Flush 5 commands: Send OBC_FLUSH_PF(0x05,0x00). 

RES> TM package from OBC with MID OBC_FP_FLUSHED is received with number of 

remaining FP items in the data section (0x05). 

DO > Type "plist" again. 

RES> The five commands with the EARLIEST timestamps are removed. 

DO > Clean up after the test by flushing the rest: OBC_FLUSH_FP(0xFF,0x00) 





DO > Send OBC_GET_UPTIME(0x00,0x00). 

RES> TM packet from OBC with MID OBC_OBC_UPTIME is received. In the data section is 

the number of seconds since last reboot in 32 bit little endian. 


ESA_Neil sets the spacecraft charging overnight. 

1 st June 2005 

Re-configure ESA_Jason's repaired laptop for groundstation. 

ESA_Neil powers up spacecraft and continues OBC checkout. 

*** Test 20: File allocation table test *** 

DO > Test 2 

DO > Select nice view in Term 

492







DO > Press file list in Term 

RES> A file list containing 31 lines (files 0 to 30) should appear in Term. No files should be 

marked open (no crosses in the 'O' column). No files should be in use. 

Passed 

*** Test 21: Opening a file *** 

DO > Test 2 (if the flash is not already cleared) 


DO > Send a telecommand requesting a file to be opened. OBC_OPEN_FILE(0x09,0x00) will 

open file 0x09 


RES> A file list containing 31 lines (files 0 to 30) should appear in Term. No file 0x09 should 

be marked open (a cross in the 'O' column). No files should be in use 

Passed 

*** Test 22: Closing a file *** 

DO > Test 21 

DO > Send a telecommand closing a file. OBC_CLOSE_FILE(0x09,0x00) will close file 0x09 



marked open (no crosses in the 'O' column). No files should be in use 

Passed 

Skip tests 23-28 for now because of time required. 

Camera dependent tests: 

-------------------------- 

Camera must be turned on from OBC for these tests to work. The specific telecommands for 

the camera can hopefully be found in the 

camera ICD or the camera functional checkout. 

*** Test 29: Turn on Camera *** 

DO > Send the appropriate telecommand to turn on camera to EPS through OBC. Check EPS 

ICD to find the command. 

DO > Enable live streaming for the next minute by sending OBC_TM_MODE_1(0xFF,0x04). 

DO > Send ping command to camera (command 0x50) 

RES> A telemetry package from camera should appear in Term within 10 seconds with the 

word "CAM" in the data part of the package. 

DO > Disable live streaming by sending OBC_TM_MODE_1(0x00,0x04). 

Passed 

493






*** Test 30: Get small thumb from CAM *** 

DO > Test 29 

DO > Send the appropriate telecommands to setup camera. 

DO > Send the appropriate telecommand to camera to take a picture. 

DO > Send the appropriate telecommand to camera to transfer small thumbnail to OBC. 

DO > Set picture filename in Term to Sex.raw and disable nice view in Term. 

DO > Switch to Messages pane in Term. 

DO > Send OBC_GET_THUMB1 to OBC. 

DO > Wait for a little while while the picture is transfering until theres a message in Term 

that the file has been written. 

RES> Open picture in appropriate picture viewer. Should look really nice. 

Passed 

*** Test 31: Get large thumb from CAM *** 

DO > Test 29 

DO > Send the appropriate telecommand to camera to transfer large thumbnail to OBC. 

DO > Set picture filename in Term to Sex.raw and disable niceview in Term. 






Passed 

*** Test 32: Get picture from CAM *** 

DO > Test 29 

DO > Send the appropriate telecommand to camera to transfer picture to OBC. 



DO > Send OBC_GET_PIC to OBC. 



RES> Open picture in appropriate picture viewer. Should look really really nice. 

PROBLEM 295: FAIL! S-band carrier only up about 50% of the time, with a period of 4 or 

5 seconds... (Or, if put carrier up full-time then modulation is only about 50% of the time -> 

OBC problem) 

So, reboot OBC, restart CAM, debug on the OBC so that can see what is going on 

Add solar panel simulation with Keithley at 750mA (takes 28.8V), can see battery level 

increase. 

494







DO > Test 29 


Watching on debug port - several chunks have to be re-requested due to: 

"Invalid SID: " followed by 5 bytes. The picture lakes longer to transfer than was waited last 

time this implies that the test was too fast and that was probably the problem. 







Passed 

S-BAND dependent tests: 

-------------------------- 

DO > Switch debugging port back to S-Band mode by typing "sb" in the debugging terminal. 

DO > Reconnect S-Band to the debugging connector on the FPP. 

DO > Make sure the ground station is set up to receive from S-Band and UHF. 

DO > Send command to EPS to power up S-Band and switch S-Band into data mode. 

Passed 

*** Test 33: Switch acknowledge target *** 

DO > Send OBC_SET_ACK_TGR(0x0C,0x00). 

DO > Send time sync via Term. 

RES> The Acknowledge comes down S-Band. 

DO > Wait 11 minutes without sending any commands to the OBC. 

Passed 

We wait for the acknowledge target to switch back to UHF. 

Note that these tests were already effectively performed during cam checkout: 

*** Test 37: Download thumb 1 *** 

DO > Turn off nice view for this test. 

DO > Send OBC_GET_THUMB1(0xFF,0x0C). 

RES> 320 packets of picture data come down S-Band (plus beacons every 36 sec). 

495






Passed 





Passed 

*** Test 39: Download picture *** 


DO > Send OBC_GET_PIC(0xFF,0x0C). 


Passed 

(from test 33) 


RES> The Acknowledge has switched back due to timeout and ack comes down UHF. 

PROBLEM 296: After 11 minutes the ACK is still coming down S-BAND, when it should 

have switched back to UHF by now. 

We try again, waiting for 25 minutes, but problem 296 occurs a second time. 

*** Test 34: Live streaming for 1 minute on S-Band *** 

DO > Send OBC_TM_MODE_1(0xFF, 0x0C). 

DO > Observe incoming telemtry 

RES> Telemetry comes crashing down S-Band and stops after 1 minute. 

Passed 

*** Test 35: Live streaming for 5 minutes on S-Band *** 



RES> Telemetry comes crashing down S-Band and stops after 5 minutes. 

Passed 





496






Passed 

Now that time has passed and the file list has started to be populated we can return to the file 

tests: 

*** Test 23: Writing a file *** 


DO > Increase telemetry rate eg. set acs in propulsion support mode (maximum telemetry 

rate) 

DO > In the debug interface type "tmrate" 

DO > Based on the tmrate wait until at least 3000 telemetry items have been generated (if 

time is not of the essence, wait until 31000 items have been generated) 


RES> A file list containing 31 lines (files 0 to 30) should appear in Term. Atleast 23files 

should have been created, depending on the number of tm items generated. 

Passed. 

NOTE: Not much time - only two files are written, but it is enough to continue. 

*** Test 24: Reading a file *** 

DO > Test 21, with file 0x00 instead of 0x09 

DO > Send a telecommand getting 25 tm items. OBC_GET_FILE(0x19,0x04) get 25 tm items 

on UHF 

RES> 25 tm items recieved in term 

DO > In the debug interface write "hexdump" "2010000" "1800" 

RES> Hexdump the first 1800 bytes~25 tm items from file 0x00 

DO > Compare The output in term with the output from the hexdump. The data format for the 

hexdump follows OBC ICD.txt Data formats: TM - Telemetry format (line 149). 

RES> The output matches. 

Passed. 

*** Test 25: Deleting a file *** 

Do > Test 24 

DO > Send a telecommand deleting a file. OBC_DELETE_FILE(0x01,0x00) will delete file 

0x01 


0x02 



marked open (no crosses in the 'O' column). File 0x01 and 0x02 should be marked deleted (x 

in the D column) 

497






Passed (used files 00 and 01 since they are the only ones we have. Prove that can also delete 

a file while it is open, and close it while it is deleted.) 

*** Test 26: Undeleting a file *** 

Do > Test 25 

DO > Send a telecommand undeleting a file. OBC_UNDEL_FILE(0x01,0x00) will undelete 

file 0x01 



marked open (no crosses in the 'O' column). Only file 0x02 should be marked deleted 

Passed (but using files 00 and 01, undeleted 01) 

*** Test 27: Protecting a file *** 

DO > Test 26 

DO > Send a telecommand protecting a file. OBC_PROTECT_FILE(0x00,0x00) will protect 

file 0x00 


marked open (no crosses in the 'O' column). All files but file 0x00 should be marked deleted (x 

in the D coloumn). File 0x00 should be marked protected (x in the P coloumn) 

Passed 

*** Test 28: Testing protection and deletion *** 

DO > Test 27 


rate) 


DO > Based on the tmrate wait until at least 3100 telemetry items have been generated. 



marked open (no crosses in the 'O' column). Files 0x00 should have the oldest date, file 0x01 

should have a newer date than 0x02. 

Leave s/c powered overnight for test 28. 

2 nd June 2005 

ESA_Neil boots up the spacecraft. OBC test 28 is passed, since the protected file remains, 

but the deleted one has been replaced. Only test 15 remains of the OBC checkout. 

498






ESA_Neil performs the PROP checkout according to the functional checkout procedure 

provided. (See sseti/DOMAIN/ftp/Express/AIV/Functional_Testing/2005-06- 

02_PROP_Testing.sex for details.) 

Controller power 

Test 4.1 is successful: the controller powers up fine, and the first pair of telemetry packets is 

received. 

Sensors 

Test 4.2.1 is successful: the pressures are out as 2C 00, 90 00, A2 00. (Corresponding to 

approximately 1 bar in each system, within error margins.) 

Test 4.2.2 is successful: the first set of temperatures are 5F, 5C, 61 and 30 – which is about 

normal. 

Test 4.2.3 is successful: the second set of temperatures are 62, 61, 30 and 5F. NOTE: This 

implies that both MAGIC thermistors are missing, not just one. 

Test 4.2.4 is successful: the third set of temperatures are 63, 6A, 62 and 60. 

High power 

Test 4.3 seems to be successful: the high power side of MAGIC is powered up. 

Pyro firing 

Test 4.4.1 is successful: the correct TMs are received and audible clicks heard as the PYRO 

ARM relay turns on and off with the correct delay. 

Test 4.4.2 is successful: the correct TMs are received and audible clicks heard as the PYRO 

ARM relay turns on and off to command. 

Test 4.4.3 is eventually successful after ESA_Neil struggles with a remarkably stubborn 

oscilloscope for an hour. The discharge curve across the test PYRO is textbook perfect 

(except for when the probe is overloaded). 

NOTE: PROP should fix the axes labels on the graphs in the checkout procedure: they show 

time against voltage, when they should show current against time. 

NOTE: The squibs are wired up the wrong way around. This command to fire the primary 

squib actually would activate the one without redundancy. 

Test 4.4.4 is successful: the primary pyro cannot be fired without arming it. 

499






Test 4.4.5 is successful: The discharge curve across the test PYRO is textbook perfect (except 

for when the probe is overloaded). 

NOTE: The squibs are wired up the wrong way around. This command to fire the secondary 

squib actually would activate the one with redundancy. 

Test 4.4.6 is successful: the secondary pyro cannot be fired without arming it. 

Branch valve 

Since it is important to know the exact status of the branch valve in order to deal with 

problem 292 (leak in low or mid pressure tubing), and it is important not to cycle the branch 

valve more often than necessary, tests 4.5.1 and 4.5.2 are not carried out. 

Test 4.5.3 is carried out three times, attempting to open the branch valve, with the 

acknowledge turned on: 

The first time there is no audible effect and MAGIC responds with 30 01 55 

The second time there is an audible click and MAGIC responds with 30 01 AA 

The third time there is no audible effect and MAGIC responds with 30 01 AA 

It is assumed that this is fine, since the valve did move on the second attempt, therefore 

implying that it was closed beforehand and the leak must be in the low-pressure tubing (this is 

very good, as we would have a serious problem if it was in the mid-pressure tubing). 

However, it is not clear why the valve did not move on the first attempt, or what the meaning 

of the initial erroneous TM packet is. 

NOTE: The functional checkout procedure should state an assumption about the initial state 

of the valve before the tests. It should also give a failure criteria here just to be clear. Also, 

there appear to be typographical errors swapping between “open” and “close” labels in tests 

4.5.3 and 4.5.4. The MAGIC user manual is consulted and considered as the authority on the 

matter. 

Test 4.5.4 is carried out twice, attempting to close the branch valve with the acknowledge 

turned on: 

The first time an audible click is heard and MAGIC reports 70 01 55 

The second time no click is heard and MAGIC reports 70 01 55 

Tests 4.5.3 and 4.5.4 are considered a success, but will be discussed with PROP to make sure, 

and to help shed light on problem 292. 

Sampling configuration 

500






Test 4.6 is successful and the sampling rate is set. 

Thruster operation 

Tests 4.7.1 to 4.7.6 are all successful (note one reboot during 4.7.6, probably unconnected). 

Each time a thruster is fired the valve can be heard opening and closing with approximately 

the correct timing, and each time the burns are associated with the correct TM. 

Manoeuvres 

Tests 4.8 – 4.15.5 are all successful. Each time a manoeuvre is performed the valves can be 

heard opening and closing with approximately the correct timing, and each time the burns are 

associated with the correct TM. 

NOTE: Without pressurising the system the only manoeuvre on which it is possible to verify 

the direction is those about the y-axis, since the order in which the clusters fire can be 

discerned audibly. However, the rotations on this axis are mirrored. This implies that the 

functional checkout document has not been updated since PROP_Sascha discovered the 

mirroring of the clusters on the 16 th April. This should be updated ASAP. 

Unless told otherwise by the PROP team upon review of this work, ESA_Neil declared PROP 

ready for flight. 

Simulate a solar panel at 600mA, and leave spacecraft in nominal mode overnight for 

endurance testing. 

3 rd June 2005 

Spacecraft has rebooted three times overnight: 

9 to 10 2005-06-02, 18:44 

10 to 11 2005-06-02. 20:47 

11 to 12 2005-06-02. 22:26 

12 to 13 2005-06-03. 05:52 

(See sseti/DOMAIN/ftp/Express/AIV/Functional_Testing/2005-06- 

03_Endurance_Testing.sex for details.) 

It does not seem periodic, which implies a COMM problem rather than a buffer overflow or 

something similar. 

The next reboot is around 15:00, after the OBC has been up 10 hours and used extensively. 

This is not good, as the 24-hour reset counter would probably never activate (and therefore 

501






leave us vulnerable to PCU / UHF lockups). However, it is not too serious either, and only 

proper “quiet” tests will characterise the problem fully. 

ESA_Neil sets the system up for CAM checkout and lets CAM_Morten play. 

Although the “auto exposure” function is somewhat temperamental, many good pictures are 

taken and CAM_Morten declares the camera ready-to-fly. 

ESA_Neil resets the groundstation and leaves the spacecraft on over the weekend for 

endurance testing. 

5 th June 2005 

ESA_Neil checks the endurance testing. The spacecraft has rebooted a total of 8 times in 

around 42 hours, this is not great, but not critical as long as it does occasionally last over 24 

hours so that the UHF and PCU would get power cycled if necessary. 


0x00000000: 00 15 00 00 FF 00 00 00 00 |................| 

0x00000000: 02 70 6F 6F FF 15 B5 A0 42 |.poo....B.......| 

On board time: 03-06-2005 19:53:01 


0x00000000: 00 15 00 00 FF 00 00 00 00 |................| 

0x00000000: 02 70 6F 6F FF A3 DB A0 42 |.poo....B.......| 

On board time: 03-06-2005 22:37:31 


0x00000000: 00 15 00 00 FF 00 00 00 00 |................| 

0x00000000: 02 70 6F 6F FF 5D 52 A2 42 |.poo.]R.B.......| 

On board time: 05-06-2005 01:16:21 


0x00000000: 00 15 00 00 FF 00 00 00 00 |................| 

0x00000000: 02 70 6F 6F FF F3 55 A2 42 |.poo..U.B.......| 

On board time: 05-06-2005 01:31:39 


0x00000000: 00 15 00 00 FF 00 00 00 00 |................| 

0x00000000: 02 70 6F 6F FF E3 7D A2 42 |.poo..}.B.......| 

On board time: 05-06-2005 04:22:03 


0x00000000: 02 70 6F 6F FF E3 7D A2 42 |.poo..}.B.......| 

On board time: 05-06-2005 04:22:03 

502







0x00000000: 00 15 00 00 FF 00 00 00 00 |................| 

0x00000000: 02 70 6F 6F FF 56 9D A2 42 |.poo.V..B.......| 

On board time: 05-06-2005 06:36:14 


0x00000000: 00 15 00 00 FF 00 00 00 00 |................| 

0x00000000: 02 70 6F 6F FF 38 C2 A2 42 |.poo.8..B.......| 

On board time: 05-06-2005 09:13:36 


NOTE: One boot was a “double” boot, as seen a few times during thermal vacuum testing. 

PROBLEM 297: Between boots 18 and 19 there are more that 24 hours, but no 

telecommands were sent, so it should have rebooted after 24 hours. This implies that the 

OBC failed the end of test 15. 

(See sseti/DOMAIN/ftp/Express/AIV/Functional_Testing/2005-06- 

03_Endurance_Testing_weekend.sex for details.) 

ESA_Neil turns off the solar panel simulator and powers up everything on the spacecraft, 

including CAM, MAGIC, ACDS high sampling, S-Band carrier and UHF live streaming in 

order to run down the battery as quickly as possible. 

After a few hours the battery voltage has dropped below the minimum threshold and the 

spacecraft enters safe-mode. 

ESA_Neil turns on the solar panel simulator and leaves the spacecraft on, with the 

groundstation configured to catch beacons again once the spacecraft re-enters nominal mode. 

A summary of the current testing anomalies: 

EPS 

When on external powering 

(battery “broken”) the OBC gets 

shut down far too often 

EPS Timers appear to be only 65 

minutes instead of 74. (293) 

OBC Friendly telecommands do have 

the “full” downlink parameter 

System Anomaly Consequences 

CAM Auto-exposure function is prone 

to memory writing errors 

Might make it harder to take good pictures, 

as will have to tweak values manually if this 

manifests itself in orbit. It was only a “niceto-have” 

in the first place though 

If the battery is having problems then the 

OBC will lose data periodically. This is not 

too serious though, since in this case can 

only be operational in sunlight anyway 

Affects the details of the collision avoidance 

analysis, but it not critical (hopefully) 

Radio amateurs could potentially “lock us 

out” the spacecraft by performing lengthy 

503






available (they should not) 

(294) 

OBC The acknowledge target does not 

automatically reset to UHF after 

10 minutes (296) 

OBC OBC was up for longer than 24 

hours without receiving a 

telecommand, but it did not 

reboot (297) 

PROP Apparent leak detected by lowpressure 

regulator, apparently it 

is before the branch valve (292) 

S-Band 

Smaller upper end of functional 

thermal range than expected 

S-Band Power output is down almost 1/3 

on expected 

T-Pod Takes 12 seconds instead of 2 

seconds to fire the +y pod 

downlinks on UHF. Details of ground ham 

s/w could help prevent this 

We should be very careful using this, 

otherwise we could end up without an 

acknowledge at all and not realise why 

This means that there is no real protection 

against the UHF or PCU entering a fault 

mode – there is no mechanism to power 

cycle them 

If this is the case then we would lose 

pressure as soon as the pyro valve was 

blown (but perhaps the second fill and drain 

valve was not done up properly). Waiting 

for answers from PROP on this 

Will not be able to have S-Band carrier up 

for very long if spacecraft is above 35 

degrees ambient 

Signal strength slightly weaker – this is not 

serious (it easier to change ground hardware 

than space hardware if necessary) 

Hopefully none. 


The spacecraft has returned to nominal mode, this completes the EPS functional checkout 

apart from detailed solar panel checks. Here it in its entirety for completeness: 

1- Solar Panels Checkout 

Test #1 

To test if the solar panels are working 

-Remove the ABF item 

-Release the AS 

-Illuminate a side of the satellite with the Solar Simulator 

-Measure the voltage with a Voltmeter drop on pins 17 to 24 of the ABF connector on the 

satellite. The GND is connected to the structure or to any other GND pin. 

-Test #1 is passed if on the eight pins (each couple of pins is connected to the panels of one 

side of the satellite) there is a voltage drop of around 33V 

-Repeat the test for each side of the satellite 

504






(Please note that if the voltage drop is lower than 33V, it means that the light source is not 

good) 

Passed 30 th May. NOTE: Do not have a solar simulator, and OHP is not quite bright enough. 

Test #2 (to be done only if test #1 has been successful) 

To test if the solar panels are giving enough power 



-Connect a Current Meter on pins 17 to 24 of the ABF connector. In series with the Current 

Meter connect the 22 OHM Shunt Resistor. The other side of the Shunt Resistor must be 

connected to GND. 

-Illuminate a side of the satellite with a Solar Simulator 

-Measure the voltage drop across the Shunt Resistor 

-Test #2 is passed if the output power of the panels (measured by multiplying the voltage by 

the current)is at least 10 Watts 

This test is not possible since we do not have a solar simulator. 

Test #3 (to be done only if test #2 has been successful) 

To test if the VR is working correctly 

-Charge the battery of the satellite (see Note #2) 

-Apply the ABF item 


-Measure the voltage on pins 17 to 24 of the ABF 

-Check that the other subsystems are being powered correctly 

-Test#3 is passed if the voltage on the checked pins is about 29V. This means that the panels 

are feeding the satellite 

-If the voltage drop on pins 17 to 24 decreases to around 24-25 V it means that the panels are 

not giving enough power to let the satellite work without the battery 

This test is not possible since we do not have a solar simulator. 

2- Solar Panels Simulation 

Test #4 

To power the satellite up with an external Current Source 

-Remove the ABF 

-Apply the MABF (see Note #1) 


-Power the Current Source up 

-Check the voltage on the Current Source 

505






-Test #4 is passed if the voltage is about 29V 

Passed 

3- Timers Checkout 

Test #5 

To test if the Timers are working correctly 

-Charge the battery (see Note #2) 

-Push the AS 

-Apply the ABF 

-Connect the Reset wire of the timers to GND for a few seconds 

-Release the AS and reset the Chronometer 

-Wait until the T-Pods are fired. When it happens, stop the Chronometer and check the time 

passed 

-Test #5 is passed if the delayed time is around 74 min and if the satellite keeps working 

Passed, but was only 65 minutes. (Problem 293.) 

Test #6 

To test if the “One-Shot” system for the Timers is working correctly 

-Start after test #5 

-Push the AS 


-Check that T-Pods are fired immediately 

-Test #6 is passed if the T-Pods are fired and the satellite keeps working correctly 

Passed 

4- PDU Checkout (*) 

Test #7 

To test if the OBC is pinged and it’s answering correctly 


-Trick the Timers (see Note #3) 


-Get everything ready to check if the OBC stays on 


-Wait a few minutes, and check if the OBC is switched on 

-Wait for some hours, checking if the pings are sent regularly to the OBC 

-Test #7 is passed if the OBC stays on for hours without ever being rebooted 

506






Failed. OBC is rebooted occasionally and randomly. (3 rd June, 5 th June…) 

Test #8 

To test if all the “Switch on/Switch off” IC’s are working 




-Get everything ready to check if the other subsystems are on or off 



-Switch the S-Band on by sending the IC: 70 0f 0f 

-Wait a few minutes 

-Switch the S-Band off by sending the IC: 70 1f 1f 

-Switch the Camera on by sending the IC: 70 4f 4f 


-Switch the Camera off by sending the IC: 70 5f 5f 

-Switch the ADCS on by sending the IC: 70 6f 6f 


-Switch the ADCS off by sending the IC: 70 7f 7f 

-Switch the Magic Box on by sending the IC: 70 8f 8f 


-Fire the Pyros by sending the IC: 70 af af 

-Wait a few seconds and check if the Pyros have actually been fired 

-Switch the power to the Pyros off by sending the IC: 70 bf bf 

-Switch the Magic Box off by sending the IC: 70 9f 9f 

-Test #8 is passed if every time an IC is sent, the respective subsystem is really switched on or 

off 

Passed 

Test #9 

To test if the power of the UHF is cycled correctly 




-Get everything ready to check if the OBC and UHF stay on 



-Send the Switch_UHF_Cycle IC: 70 1f 1f 

-Test #9 is passed if the power to UHF is cycled (off for 5 seconds and then on again) 

Passed 

507






Test #10 

To test if the reset IC for the PDU works 







-Send the Reset_PDU IC: 71 cf cf 

-Test #10 is passed if the PDU is reset; that implies that all the subsystems are switched off 

for a few seconds and the satellite goes back to Safe Mode 

Passed 

Test #11 

To test if the battery sensing is working 







-Wait until the OBC receives a few Nominal Beacon Telemetry packs 

-Check the voltage and the temperature values on the battery sent with the Nominal Beacon 

Telemetry 

-Measure with a Voltmeter the voltage on pin 1 of the ABF item 

-Test #11 is passed if the Battery Temperature value received is around the external one 

(maybe a few degrees more), and if the Battery Voltage is the same of the one checked on 

the ABF item 

Passed 

Test #12 

To test if the “Safe Beacon” on the PTT is sent correctly 




-Get everything ready to check if the OBC stays on and if the UHF sends the Safe Beacon 


-Measure with a Voltmeter the voltage on pin 1 of the ABF item 

-Make the OBC to not answer to the pings (in this way the spacecraft does not go to Nominal 

Mode) 

508






-Wait a few minutes until the UHF has sent the Safe Beacon a few times 

-Test #12 is passed if the last twelve bits sent by the UHF represent the Battery Voltage 

checked on the ABF item 

Passed 

NOTE: We need to make a program that can decode these beacons. 

Test #13 

To test if the battery is considered broken when absent 


-Connect Pins 4 and 5 of the ABF connector to the positive output of the Current Source 

-Connect the negative output of t he Current Source to GND 


-Switch the Current Source on 

-Test #13 is passed if after a few minute the satellite goes to Recovery Mode (the battery is 

considered broken because it is not connected) 

Passed. However, extra reboots are occurring in this configuration. 

(*) All the IC’s are sequences of three hexadecimal numbers 

5- Battery Checkout 

Test #14 

To check if the battery is working 

-Charge the Battery (see Note #2); do not disconnect the power source 

-Check the Battery Voltage on the positive output pin of the External BCR 

-Wait some hours (depending on the actual charge of the battery) 

-Test #14 is passed if the current taken by the External BCR constantly decreases until it gets 

constant to a few mA when the Battery Voltage has increased to about 24.6 V 

Passed 

Test #15 

To check if the BDR is working 



-Connect a Power Source (24 V) to pin 4,5 or 6 of the ABF connector 


-Switch the Power Source on 

-Test #15 is passed if the satellite works normally (the BDR is working) 

509






Passed 

This is a total rewrite of OBC test #15 from ACDS_Lars and performed by ESA_Neil. First 

the current mode of ACDS operation is checked: 




0x00000010: 61 74 69 6F 6E |ation...........| 

DO > Send OBC_ACDS_ENABLE(0x01, 0x00). 

DO > Reboot OBC. 

RES> ACDS places string: "ACDS normal Operation" in the ALARM stack. 



0x00000000: 41 43 44 53 20 4E 6F 20 4F 70 65 72 61 74 69 6F |ACDS No Operatio| 

0x00000010: 6E |n...............| 

PROBLEM 298: ACDS operation mode reports incorrectly. 



RES> ACDS places string: "ACDS inverted Operation" in the ALARM stack. 



0x00000000: 41 43 44 53 20 49 6E 76 65 72 74 65 64 20 4F 70 |ACDS Inverted Op| 

0x00000010: 65 72 61 74 69 6F 6E |eration.........| 



RES> ACDS places string: "ACDS No Operation" in the ALARM stack 




0x00000010: 6E |n...............| 

DO > Wait for 24 hours without transmitting any commands to the OBC and 

without power cycling or rebooting it. 

This is not possible, since the OBC fails on this point (problem 297). 

RES> OBC gets shut down after 24 hours and when it wakes ACDS places 

string: "ACDS inverted Operation" in the ALARM stack. 

Result not obtainable. 

510












0x00000010: 6E |n...............| 

Failure, but could be due to problem 297. 

So, ESA_Neil sets OBC_ACDS_ENABLE to (0x00 0x00) and reboots the OBC. 




0x00000010: 61 74 69 6F 6E |ation...........| 

Results are transmitted to ACDS_Lars for analysis. 

ACDS_Lars corrects the test procedure and ESA_Neil performs it again: 




Passed 




Passed 




Passed 





511






Not possible due to OBC failure. 




Passed 

Aside from ACDS mode iteration upon OBC 24-hour timeout, the OBC side of ACDS is 

ready to fly. 

The spacecraft is left on overnight with both the groundstation software and a terminal on the 

debugging port logging to try and gain some insight on the OBC reboot problem. 


The spacecraft has rebooted once overnight. 

EPS_Fulvio confirms that the EPS code 

(/sseti/DOMAIN/ftp/Express/EPS/Sseti_EPS_Electronics/PDU/Sseti_Express_PDU_Softwar 

e/Sseti_Express_PDU_Flash_Software_ver_7.5/) should do the following after the last 

successful pong from OBC: 

1) Last successful pong from OBC (sets range and counter to 0) 

2) 20 second wait 

3) Ping (range switches from 0 to 1) 


5) Ping (range switches from 1 to 2) 


7) OBC is shutdown and power cycled 

So, there is about ONE MINUTE (ignoring interrupt times) between the last successful pong 

from OBC and the initiation of shutdown. 

This is an extract from the debugger: 

>Received ping 06-06-2005 20:25:30 

>Received ping 06-06-2005 20:25:52 

>Received ping 06-06-2005 20:26:14 

>Received ping 06-06-2005 20:26:36 

>Received ping 06-06-2005 20:26:58 

>fetch queue timeout, sid (2) 

Hex dump: 

0x03017B98: 

Suhtdown from EPS 

shutdown initiated at 

512






06-06-2005 20:27:40 

>shutdown complete 

Note: that the initiation of shutdown is only 42 seconds after the last successful ping. This 

implies that the pong at this time was NOT successful, and that the last successful pong was 

actually after the 20:26:36 ping. Therefore either the OBC TX or the EPS RX is not working 

correctly. 

This information is passed to the EPS and OBC teams for discussion and further tests are 

carried out. 

ESA_Neil, OPER and MCC set up their local infrastructures so that OPER can control the 

MCC from Poland at the same time as the MCC is controlling the groundstation (cleanroom) 

from Aalborg at the same time as the groundstation is controlling the spacecraft. 

ESA_Neil and OPER have an operations training session, in which 

- two-way communication is established 

- basic housekeeping and alarm retrieval commands are tested 

- s-band downlink is tested 

- picture downlink is tested 

- MAGIC is powered up and all four thrusters fired 

Issues arising from the session were: 

i) The MCC should use the local machine time for the time-synchs 

ii) It should be made possible to see the data for undefined items (just in case 

we get something we are not expecting), and it should be made obvious 

that an undefined item has arrived 

iii) It should either be possible to define particular instances of commands with 

parameters included, or be possible to “import” lists of TCs already 

prepared with parameters 

The spacecraft is again left turned on overnight for troubleshooting purposes, logging on both 

the groundstation and the debugging port. 


Another reboot occurred overnight, and another in the morning. The debug logs are analysed: 

This seems to happen occasionally: 

>Received ping 07-06-2005 18:18:26 

>Received ping 07-06-2005 18:18:48 

>Invalid SID received: 0x99 0x06 0x00 0xCC 0x00 0xC4 0x00 

>Received ping 07-06-2005 18:19:10 

>Received ping 07-06-2005 18:19:32 

513






Which is a failure of OBC to send itself a nominal mode beacon packet (it appears that the 

start bit has been missed, and therefore the SID (00) also). 

This happens a little less often: 

>Received ping 07-06-2005 20:41:40 


Hex dump: 

0x03017B98: 

Received ping 07-06-2005 20:42:22 

Which is an EPS TX or an OBC RX failure. However, these two are just examples of what 

sometimes happens - they were NOT reboots. 

Here are the extracts from the debugger over the latest two captured reboots: 

>Received ping 07-06-2005 22:47:24 

>Received ping 07-06-2005 22:47:46 

>Received ping 07-06-2005 22:48:06 

>Suhtdown from EPS 


07-06-2005 22:48:27 


and 

>Received ping 08-06-2005 10:18:41 


Hex dump: 

0x03017B98: 

Received ping 08-06-2005 10:19:23 

>Received ping 08-06-2005 10:19:45 

>Received ping 08-06-2005 10:20:07 

>Received ping 08-06-2005 10:20:27 

>Suhtdown from EPS 


08-06-2005 10:20:49 


In both of these cases it seems that the ping was fine but the pong failed twice in a row. That 

is now THREE captured reboots with pong failures. 

The separation between ping and pong failures here lends credence to OBC_Karl's hypothesis 

that they are not directly causally related. 

More tests are being run and we would like to see the same thing demonstrated TWICE more 

before taking a concrete decision on how to respond. 

The following provisional to-do list is prepared for OBC: 

514






1) Fix the endian problem so that we disallow the FF parameter for radio hams, apart 

from on command 16 (live stream), where they will need it 

2) After command 31 make the acknowledge target switch back to UHF after 10 mins 

(REGARDLESS of whether or not TCs are received) 

3) Fix the 24-hour timeout problem by making the timeout counter ignore all 

"automated" processes on the spacecraft and only be reset by received and valid TCs 

4) Perform the following list of actions, in order, upon a 24-hour timeout: 

a) Write the housekeeping and alarm stacks to flash 

b) Iterate the ACDS mode if (and only if) the counter has not been "frozen" 

manually 

c) Send the "UHF power cycle" command to the PCU 

d) Perform whichever graceful shutdown procedures are appropriate and do 

NOT restrict the ability of OBC to command the PCU (see (f) and (h)) 

e) Wait for 10 seconds 

f) Send the "UHF power cycle" command to the PCU (just to be sure) 

g) Wait for 10 seconds 

h) Send the "Reset PCU" command to the PCU (after which the OBC will be 

immediately powered off) 

5) Fix the ACDS counter bug 

6) Have the OBC save data to flash during any graceful shutdown 

7) Triple the ping responses so that after a "ping" is received from the PCU the OBC 

sends THREE "pong"s with a gap of 2 seconds between each. THIS ITEM TO BE 

CONFIRMED TOMORROW 

8) Burn PROMS and ship them to ESTEC so that they arrive Monday at the latest 

9) Update the ICD with all changes noted during checkout 

10) Update the functional checkout document to reflect all changes 

A prom change after this should fix problems 220, 294, 296 and 297. 

The spacecraft is left on overnight again for more data collection. 


Data from the debugger surrounding one of the overnight reboots: 

>Received ping 09-06-2005 04:30:16 

>Received ping 09-06-2005 04:30:38 

>Received ping 09-06-2005 04:31:00 

>Received ping 09-06-2005 04:31:22 

>de-kissification error (-9) 

fetch queue timeout, sid (2) 

Hex dump: 

0x03017B98: 



515






09-06-2005 04:32:04 


ESA_Neil sets up “constant” live-streaming (by putting 10 minute live streaming commands 

into the flight plan with timestamps incrementing by 10 minutes each) so that it is possible to 

watch the “errors” the EPS registers from OBC (first byte in EPS TM). 

The latest two reboots are given below, both from the debugger and from term.exe. 

First reboot from debugger: 

Received ping 09-06-2005 10:16:10 

Received ping 09-06-2005 10:16:31 

Received ping 09-06-2005 10:16:52 



09-06-2005 10:17:13 

shutdown complete 

First reboot from term: 


Subsystem: EPS | MID: 0xF0 | Length: 16 | Time: 09-06-2005 10:16:03 

0x00000000: 00 00 AC 00 AF 00 D7 00 17 00 06 00 18 00 BE B8 |................| 



0x00000000: 01 00 B1 00 AF 00 DB 00 1D 00 18 00 F7 00 CA B8 |................| 



0x00000000: 04 00 B4 00 B0 00 E0 00 46 00 17 00 F0 00 C3 B8 |........F.......| 


Subsystem: OBC | MID: 0xF8 | Length: 9 | Time: 09-06-2005 10:17:12 

0x00000000: 00 15 00 00 FF 00 00 00 00 |................| 

Second reboot from debugger: 

Received ping 09-06-2005 11:48:02 

Received ping 09-06-2005 11:48:24 

Received ping 09-06-2005 11:48:44 



09-06-2005 11:49:06 

work: flash TM 

Attempting to open file (-1) 

Opening file (28) 

Erasing block number (36) 

...ACDS shutdown 

flight planner shutdown 

done 

Programming from block (36) 

516






...done 

Wrote 65536 bytes to file 28 

Closing file (28) 

shutdown complete 

Second reboot from term: 



0x00000000: 00 00 AF 00 B3 00 DB 00 18 00 18 00 FF 00 BF B8 |................| 



0x00000000: 04 00 B8 00 C0 00 E0 00 40 00 4F 00 E1 00 D3 B8 |........_at_.O.....| 


Subsystem: OBC | MID: 0xF8 | Length: 9 | Time: 09-06-2005 11:49:04 

0x00000000: 00 15 00 00 FF 00 00 00 00 |................| 

There were 4 lots of "rubbish" received by EPS from OBC in each case, this certainly implies 

pong failure. EPS_Fulvio and OBC_Karl can conceive of how a failed pong could be split 

into two fragments by the EPS linear buffer read vulnerability, but not how more pieces could 

be generated. This seems to be reasonable, implying that the 4 errors encountered were from 

two split pongs. This is very important, since, if the EPS is actually rebooting because of 

excess “rubbish” from OBC then tripling the pong response (favourite course of action) will 

only make it worse. 

From looking at the EPS code: timeout in the buffer is 2 seconds. Therefore OBC_Karl, 

EPS_Fulvio, SYS_Jörg and ESA_Neil agree to making the time between triple-pongs set to 

THREE seconds, to make sure the timeout has occurred (otherwise it'll just make more 

rubbish). 

ESA_Neil and OPER go through a second training session. Items generated for MCC are: 

- The database does not store the parameters on sent telecommands 

- The acknowledge data is not represented properly 

The spacecraft is left on overnight again. 

10 th June 2005 

The overnight reboot is similar to the others. 

ESA_Neil analyses the log files from yesterday and discovers several cases where the EPS 

error byte registers two errors from OBC, but then gets reset to zero. This is consistent with 

517






the current theory that two errors are caused by a split pong. (Note that the EPS telemetry 

will only register this if no successful ping is received before the telemetry is sent, so there 

must have been more cases than recorded.) 

The above two points lend further credence to the plan. 

OBC_Karl performs all actions on the to-do list and ships the proms to ESTEC. 

12 th June 2005 

ESA_Neil powers up the spacecraft and allows remote control from MCC and OPER. 

PROBLEM 299: OPER discover that the marker on the GET_FILE command increments 

one packet too far on each downlink. This means that single packets will be missed between 

increments of GET_FILE, unless a FILE_SEEK of –1 (FFFF) is used in between. 

The following to-do list is defined for MCC: 

1) Time synch should get time from local machine and convert to UTC 

2) MCC should store paramters that were sent 

3) All data should be visible in all recieved TM packets 

4) User should be notified if an undefined TM packet is received 

5) "Verify TC" functionality should be added (toggle upper bit of CID) 

6) Definition of TCs with parameters should be allowed, or, import of scripts (txt / csv / 

similar) 

7) Define all SIDs properly (TCS was missing) 

8) Decode the beacon data properly into human-readable format 

ESA_Neil leaves the spacecraft powered up for OPER to continue checkout. 

13 th June 2005 

ARRIVAL 87: The new OBC flight proms arrive. 

MODIFICATION 143: ESA_Neil replaces the OBC flight proms, which necessitates 

disintegrating the box and cycling the prom sockets. 

518






ESA_Neil removes the –x and +y lateral panels, then disintegrates the OBC. 

The OBC is opened, and the first old prom removed. NOTE: It is VERY difficult to remove 

the prom, since it just pivots on the glue. The amount of stress put on the socket is worrying, 

especially since it does not come out “parallel”. 

519






ESA_Neil cleans the socket and the new pros, then inserts the new one. It fits fine and seems 

secure. 

The second prom is removed. It comes out a little easier than the first, but not much. 

ESA_Neil cleans the socket and the new prom, and then inserts it. 

520






ESA_Neil connects the power line of the OBC from the spacecraft, and the debug port to a 

laptop running a terminal. 

PROBLEM 300: There is no response from OBC on the debugging line when it is powered 

up with the new proms in place. 

ESA_Neil checks the voltage being supplied to the OBC, it is correct. He checks the current 

consumption, and it is about right, perhaps a little high (22.5mA). 

After consultation with OBC_Karl, ESA_Neil puts the old proms back in and attempts to 

bring the OBC up again. 

PROBLEM 301: There is no response from OBC on the debugging line when it is powered 

up with the old proms in place. 

ESA_Neil removes the chips and carries out a visual inspection of the sockets, there is no 

apparent damage. 

A microscope is used to inspect the sockets. One pin appears to be slightly misaligned and 

discoloured – perhaps with glue. 

MODIFICAITION 144: A scalpel is used to scrape the discolouration off of the potentially 

problematic pin and to gently bend it back into alignment with the rest of the socket. 

The new proms are placed again and the OBC booted up successfully. The EPS RS232 line is 

connected with savers on the OBC and the cable. It is clear that the new ping-pong structure 

is working, since the triple pongs separated by 3 seconds cause an inter-ping time of 28 

seconds instead of 22. 

ESA_Neil conformal coats the top of the flight battery and uses the small vacuum oven to 

remove the air. 

521






The OBC is left up overnight to ascertain its stability. 

14 th June 2005 

The OBC is still up after 14 hours. There were several instances like the following, but the 

OBC was not rebooted: 

>Received ping 13-06-2005 18:22:36 

>Invalid SID received: 0x99 ¹x06 0x00 0xCD 0x00 0xB6 0x00 

>Received ping 13-06-2005 18:23:04 

>Received ping 13-06-2005 18:23:32 

>Received ping 13-06-2005 18:24:00 

>Received ping 13-06-2005 18:24:28 

>de-kissification error (-9) 

fetch queue timeout, sid (2) 

Hex dump: 

0x03017BD8: 

Received ping 13-06-2005 18:25:16 

>Received ping 13-06-2005 18:25:44 

An extra serial cable is added, using savers, to connect the UHF system. The spacecraft is 

then left up to continue endurance testing. 

15 th June 2005 

The spacecraft has not rebooted overnight, staying up for about 14 hours. This is a good sign 

for the stability tests – although proper context testing must wait for re-integration. 

NCube-II finishes its vibration tests. The antennas did not deploy, implying that the 

modification was successful and problem 242 is solved. 

522






PROP_Nils and PROP_Sascha arrive to test the PROP system (regarding problem 292) and 

the propulsion ground support hardware. 

ESA_Neil glues the new OBC proms in place just as before, tightens and glues the PCB bolts, 

closes, torques and glues the box, reintegrates it to the structure and connected all the 

connectors. The spacecraft is then booted up successfully. 

PROP_Sascha and PROP_Nils conduct a functional checkout of the propulsion system. 

ESA_Neil powers up s/c and MAGIC. PROP_Sascha and PROP_Nils starts with functional 

checkout of the PROP system. All fine. 

PROP_Nils and PROP_Sascha starts to pressurise the high pressure part using the test port up 

to 300 bar. Pressures read out, all fine. The FD equipment still was mounted on the s/c. 

Performed an checkout of the activation commands for turning the s/c around specified axis. 

FAILED, 

PROBLEM 302: Some thrusters are activated the wrong way, see the following table 

Command sent to should be activated was fired first 

MAGIC 

first 

+x 1,4 1,3 

-x 2,3 2,4 

+y 3,4 1,2 

-y 1,2 3,4 

+z 4,2 2,3 

-z 1,3 1,4 

The base for the thruster labelling : Thruster 1 was the one, next to PMS and then 

rounding the spacecraft in –z direction to 4. 

Then the Leakage test started. First pressures were read out, all fine. Then the BV was closed, 

the pressures read out again. Still all fine. To check, if either the low pressure, the mid 

pressure or the high pressure part would leak, the BV and FD valve were closed and the FD 

equipment detached. The pressures of the middle and low pressure part now were monitored, 

to identify a leak and to clarify Problem 292. (See 

SSETI_EXPRESS_PROPPAYLOAD_Leakage-Report.xls for detailed figures, 

SSETI_EXPRESS_PROPPAYLOAD_Leakage-Test-Procedure_050616.DOC for the test 

report). 

16 th June 2005 

The spacecraft has not been rebooted overnight, it has been up for around 20 hours. This is a 

good indication of the stability of the triple-pong-response system. 

523






Concerning PROPLEM 292 there have been the following results: 

• There is a leakage in the low pressure part 

• The leakage-rate is about 0.3 bar / 100 mins 

• The leakage is after the Branch Valve, so its not a critical problem, the usage of 

PROPPAYLOAD is still possible 

• The BV should be closed, as soon as possible after finishing PROPPAYLOAD tasks 

PROBLEM 303: Due to some HPR effects the mid pressure is rising over time. 

Concerning the slightly pressure increase of the HPR, there probably won’t be a problem, due 

to the fact, that the LPR is proofed up to 210 bar and the increasing rate is decreasing over 

time. (See SSETI_EXPRESS_PROPPAYLOAD_Leakage-Report.xls). So the mid pressure 

will reach about 35 bar and then become stable. If the thrusters will be activated, the pressure 

will fall again to nominal 20 bar. 

So Problem 303 isn’t critical for PROP, too. 

NCUBE_Åge opens NCube-2 and performs a visual inspection of the interior. No serious 

problems are identified and NCube-2 is sealed again ready for the flight. 

ESA_Neil replaces the loose bolt in the –x T-Pod, therefore solving problem 288. 

PROBLEM 304: The OBC reboots after 1 day and 56 minutes. However, this appears to be 

an internal problem and was not a shutdown from EPS. Also, since this problem is certainly 

rare (first time we have seen it), and happened after 24 hours, it poses no significant danger to 

the mission. 

Received ping 16-06-2005 11:16:11 

>Live streaming! Timing out in 301 sec 

Live streaming! Timing out in 294 sec 




524










Received ping 16-06-2005 11:16:39 

>$T0athread:00000005;0f:c8850101;0d:fc040403;#29nüÀª’@@@@`¦¦¨’bað*8www.sseti.net * 

SSETI-Express 

ESA_Neil carries out an OBC function checkout, following instructions from OBC_Karl: 

OBC-only checkout tests: 

---------------------------- 

*** Test 1: Debugging screen *** 

DO > Type in the terminal program. 

RES> The debugging screen should appear in the terminal program. 

Passed. 

*** Test 2: Flash integrity *** 

DO > type "ferase" to erase FLASH. 

NOTE: The OBC reboots during “erasing chip…”. Perhaps it is latch-up protection. 

Regardless, the memory wipe does work, since after reboot the boot counter is incremented 

and all the flash files are missing. 

DO > Type "fgeni". 

RES> A new FLASH integrity field is generated. 

DO > Type "fcheck". 

RES> Debugger replies: "flash crc check ok, flash is 'good'". 

Passed. 

*** Test 3: EPS ping *** 

DO > Wait for 2 minutes. 

RES> If the OBC hasn't been rebooted then the EPS ping works. The ping should also be 

displayed in the debugging interface. 

Passed. 

*** Test 4: Kill WD *** 

DO > Type "wdtog". 

RES> OBC should reboot within 120 seconds. 

Additional from ESA_Neil: before it finishes shutting down the OBC should write the HK 

and AL stacks to flash. 

525






PROBLEM 305: Upon shutdown the OBC is powered off before it writes the HK file, and 

before it marked the AL file as “alarm”. 

However, the AL file was written and is downloadable upon reboot. Although it is not ideal 

that the HK file is not written, it was only a ‘nice to have’ and is not critical. 

This test is a partial failure, but an acceptable one. 

Context dependent tests: 

--------------------------- 

For the context dependent tests the UHF must be operational and a 

ground station must be set up. Use Term.exe to issue commands and 

review replies for the following tests. The debugger must still be 

hooked up to the debugging interface during these tests. 

DO > Set up Term to use the proper serial port. 

DO > Hook up Term to the ground station TNC. 

*** Test 5: Beacon *** 

DO > Power up spacecraft and allow it to go to nominal mode. 

DO > Wait for nominal mode beacons to appear in Term. 

RES> Correct-looking nominal mode beacons should appear every 18 sec. 

Passed. 

*** Test 6: Sync time *** 


DO > Wait for next beacon 

RES> Next beacon should contain the correct time. 

Passed. 


DO > Send OBC_GET_UPTIME to OBC 

RES> The OBC should reply with the current uptime in Term. 

Passed. 

*** Test 8: TCS sampling *** 

DO > nothing 

DO > Wait for two minutes then send OBC_GET_HK to OBC. 

RES> If the subsystem worked the temperatures in the spacecraft could be seen in the 

telemetry. 

526






Passed. 

*** Test 9: Test shutdown through UHF *** 

DO > send OBC_DO_SHUTDOWN to obc 

RES> The OBC should shut down 

Additional from ESA_Neil: before it finishes shutting down the OBC should write the HK 

and AL stacks to flash. 

Passed. Note that the file writes on shutdown do work when the shutdown is not from EPS. 

*** Test 10: Get housekeeping *** 

DO > Send OBC_GET_HK(0x0A,0x04) to OBC 

RES> 10 housekeeping packages stored on the OBC, if there is 10, should appear in Term. 

Passed. 

*** Test 11: Get alarms *** 

DO > Send OBC_GET_AL(0x0A,0x04) to OBC 

RES> 10 alarms stored on the OBC, if there is 10, should appear in Term. 

Passed. 

*** Test 12: Non-public telecommands *** 

DO > Turn off encryption in Term. 

DO > Send the non-public telecommand, OBC_SYNC_TIME. 

RES> TM packet from OBC with MID OBC_TC_CRYPT_ERROR (0x40) is received by Term 

within 5 sec. 

Passed. 

*** Test 13: Public telecommands with public commands disabled *** 


DO > Send public telecommand without encryption using Term. The first parameter must not 



within 5 sec. 

Passed. 

NOTE: The OBC ICD must be corrected: currently it lists OBC_SYNCH_TIME as a public 

command, which it is not, and does not list OBC_GET_PICTURE as a public command, 

although it is. 

*** Test 13: Public telecommands *** 

527







DO > Turn on public access: Send OBC_PUBLIC_ACCESS(0xFF,0x00) and wait 1 second. 

DO > Send public telecommand without encryption using Term. The first parameter must not 


Correction from ESA_Neil: in OBC_LIVE_STREAM fist parameter should be 0xFF. 

RES> A normal acknowledge is received by term. 

Passed. 

*** Test 14: Public telecommands with bad parameter *** 

DO > Do test 13 with the first parameter set to 0xFF in all commands. 


within 5 sec. 

Correction from ESA_Neil: with command OBC_LIVE_STREAM a normal acknowledge 

should be received. 

Passed. 

*** Test 16: Commands with verify bit *** 

DO > Send OBC_GET_HK(0x01,0x04) with verify bit (press the "V" button in Term in stead 

of "Send") 

DO > Wait for 1 second. 

DO > Get alarms: Send OBC_GET_AL(0xFF,0x04); 

RES> The alarm data is received by Term and one entry has the previously sent TC 

embedded in the data section. The MID of the alarm entry is 0x95. 

Passed. 

*** Test 17: Flushing TM *** 

DO > Send OBC_FLUSH_HK(0x00,0x00). 


packets in the queue in the data field. 

DO > Send OBC_FLUSH_HK(0xA0,0x00). 


packets in the queue in the data field. This number should be approx. 10 smaller than before. 

DO > Repeat test with OBC_FLUSH_AL. 

Passed. 

*** Test 18: Flush flight plan *** 

DO > Send 10 commands to the OBC to be executed in 500 sec. OBC_GET_HK(0x00,0x04). 

DO > In the debugger, type "plist". 

RES> 10 commands are in the list appearing on the debugging screen. 

DO > Flush 5 commands: Send OBC_FLUSH_PF(0x05,0x00). 

528








DO > Type "plist" again. 

RES> The five commands with the EARLIEST timestamps are removed. 

DO > Clean up after the test by flushing the rest: OBC_FLUSH_FP(0xFF,0x00) 



Passed. 

NOTE: be careful using this too much, as it also flushes pongs out of the flight plan, 

therefore reducing the EPS-OBC watchdog stability. 


DO > Send OBC_GET_UPTIME(0x00,0x00). 

RES> TM packet from OBC with MID OBC_OBC_UPTIME is received. In the data section is 

the number of seconds since last reboot in 32 bit little endian. 

Passed. 

File system checks: 

_________________________ 

*** Test 20: File allocation table test *** 

DO > Test 2 

DO > Select nice view in Term 




marked open (no crosses in the 'O' column). No files should be in use. 

Passed. 

*** Test 21: Opening a file *** 



DO > Send a telecommand requesting a file to be opened. OBC_OPEN_FILE(0x09,0x00) will 

open file 0x09 


RES> A file list containing 31 lines (files 0 to 30) should appear in Term. No file 0x09 should 

be marked open (a cross in the 'O' column). No files should be in use 

Passed. 

529






NOTE: The ICD should be updated, the second parameter must be 0x00, where as the ICD 

claims it doesn’t matter. 

*** Test 22: Closing a file *** 

DO > Test 21 

DO > Send a telecommand closing a file. OBC_CLOSE_FILE(0x09,0x00) will close file 0x09 



marked open (no crosses in the 'O' column). No files should be in use 

Passed. 



*** Test 23: Writing a file *** 



rate) 


DO > Based on the tmrate wait until at least 3000 telemetry items have been generated (if 

time is not of the essence, wait until 31000 items have been generated) 


RES> A file list containing 31 lines (files 0 to 30) should appear in Term. Atleast 23files 

should have been created, depending on the number of tm items generated. 

Passed. 

*** Test 24: Reading a file *** 

DO > Test 21, with file 0x00 instead of 0x09 

DO > Send a telecommand getting 25 tm items. OBC_GET_FILE(0x19,0x04) get 25 tm items 

on UHF 

RES> 25 tm items received in term 

DO > In the debug interface write "hexdump" "2010000" "1800" 

RES> Hexdump the first 1800 bytes~25 tm items from file 0x00 

DO > Compare The output in term with the output from the hexdump. The data format for the 

hexdump follows OBC ICD.txt Data formats: TM - Telemetry format (line 149). 

RES> The output matches. 

Passed, but note problem 299. 

*** Test 25: Deleting a file *** 

Do > Test 24 


0x01 

530







0x02 



marked open (no crosses in the 'O' column). File 0x01 and 0x02 should be marked deleted (x 

in the D column) 

Passed. 



*** Test 26: Undeleting a file *** 

Do > Test 25 

DO > Send a telecommand undeleting a file. OBC_UNDEL_FILE(0x01,0x00) will undelete 

file 0x01 



marked open (no crosses in the 'O' column). Only file 0x02 should be marked deleted 

Passed. 



*** Test 27: Protecting a file *** 

DO > Test 26 

DO > Send a telecommand protecting a file. OBC_PROTECT_FILE(0x00,0x00) will protect 

file 0x00 


marked open (no crosses in the 'O' column). All files but file 0x00 should be marked deleted (x 

in the D column). File 0x00 should be marked protected (x in the P column) 

Passed. 



*** Test 28: Testing protection and deletion *** 

DO > Test 27 


rate) 


DO > Based on the tmrate wait until at least 3100 telemetry items have been generated. 


531







marked open (no crosses in the 'O' column). Files 0x00 should have the oldest date, file 0x01 

should have a newer date than 0x02. 

Passed. 

Camera dependent tests: 

-------------------------- 

Camera must be turned on from OBC for these tests to work. The specific telecommands for 

the camera can hopefully be found in the 

camera ICD or the camera functional checkout. 

*** Test 29: Turn on Camera *** 

DO > Send the appropriate telecommand to turn on camera to EPS through OBC. Check EPS 

ICD to find the command. 

DO > Enable live streaming for the next minute by sending OBC_TM_MODE_1(0xFF,0x04). 

DO > Send ping command to camera (command 0x50) 

RES> A telemetry package from camera should appear in Term within 10 seconds with the 

word "CAM" in the data part of the package. 

DO > Disable live streaming by sending OBC_TM_MODE_1(0x00,0x04). 

Passed. 

*** Test 30: Get small thumb from CAM *** 

DO > Test 29 

DO > Send the appropriate telecommands to setup camera. 

DO > Send the appropriate telecommand to camera to take a picture. 

DO > Send the appropriate telecommand to camera to transfer small thumbnail to OBC. 







Passed. 

*** Test 31: Get large thumb from CAM *** 

DO > Test 29 

532






DO > Send the appropriate telecommand to camera to transfer large thumbnail to OBC. 







Passed. 


DO > Test 29 








533






Passed. 

NOTE: This picture is scaled down and no attempt was made to optimise the exposure time. 

S-BAND dependent tests: 

-------------------------- 

DO > Switch debugging port back to S-Band mode by typing "sb" in the debugging terminal. 

DO > Reconnect S-Band to the debugging connector on the FPP. 

DO > Make sure the ground station is set up to receive from S-Band and UHF. 

DO > Send command to EPS to power up S-Band and switch S-Band into data mode. 

*** Test 33: Switch acknowledge target *** 

DO > Send OBC_SET_ACK_TGR(0x0C,0x00). 


RES> The Acknowledge comes down S-Band. 

DO > Wait 11 minutes without sending any commands to the OBC. 

Correction from ESA_Neil: the rest should not be TC dependant, but only on a timer. 


RES> The Acknowledge has switched back due to timeout and ack comes down UHF. 

Passed. 

*** Test 34: Live streaming for 1 minute on S-Band *** 


DO > Observe incoming telemetry 

RES> Telemetry comes crashing down S-Band and stops after 1 minute. 

534






Passed. 





Passed. 





Passed. 





Passed during camera dependant tests. 






*** Test 39: Download picture *** 


DO > Send OBC_GET_PIC(0xFF,0x0C). 



*** Test 15: ACDS startup mode *** 




Passed 


535








Passed 




Passed 



Spacecraft is left on overnight for this purpose, with a terminal logging the debug line in order 

to capture internal work performed by OBC on timeout. The OBC is on boot number 7. 

17 th June 2005 



Additional from ESA_Neil: before it gets powered off the OBC should write the HK and AL 

stacks to flash, power-cycle the UHF unit twice, then reset the PCU. 

PROBLEM 306: Upon a 24 timeout the OBC did not write the HK and AL stacks to flash, 

did not power cycle the UHF at all, and did not reset the PCU. On the debugging line it 

simply reported “Incremented ACDS mode counter to (7)” and then reset. 

>Received ping 16-06-2005 18:57:20 

>planner: time sync inject! 

sync time to (1118948268) 

time diff (0) 

Received ping 16-06-2005 18:57:48 

-- 24-hour gap here -- 

>Received ping 17-06-2005 18:57:11 

>Received ping 17-06-2005 18:57:39 

>Incremented ACDS mode counter to (7) 

nüÀª’@@@@`¦¦¨’bað*8www.sseti.net * SSETI-Express 

However, the ACDS mode was incremented. 

This test is a partial failure and is under consideration by OBC, EPS and SYS. 


536








Passed. 

ESA_Neil tries to erase the flash memory by telecommand and the following happens: 

- The debugger "froze" as during test 2. 

- An attempt was made to time-synch, and it WAS reported "Time synch inject" in the 

debugger, but it did NOT respond (on debug or term) with the time difference. 

- Attempt to send a couple of other telecommands, and nothing happened 

- Eventually a shutdown was reported from EPS, presumably on a ping timeout 

18 th June 2005 

ESA_Neil has the idea to mostly solve problem 306 by adding the following three commands 

into the pre-flight plan (the list of TC that will be sent to the spacecraft at the start of each 

pass): 

1) Timestamp: 0 (now), Flush the flight plan entirely (this can be missed out for those 

rare cases when we have a flight plan longer than one orbit) 

2) Timestamp: 23.5 hours, power cycle UHF 

3) Timestamp: 23.6 hours, reset PCU 

This would then ensure that any loss of uplink would be covered by a power cycling of the 

TNC in UHF. The only vulnerabilities would be during LEOP before the commands were in 

place, and during the second or two between sending commands 1 and 3. 

NOTE: We should only do this from ONE groundstation, and then only use THAT 

groundstation to update the long-term flight plan (otherwise the next pass of the other 

groundstation would automatically wipe out the long-term plan). 

NOTE: There is a very small chance that the first command could cause a pong failure as it 

flushes the pongs out of the flight plan. This is not very likely though, and is even less likely 

to be consecutive with another pong failure, the combination of which would reset the OBC. 

This possibility will be discussed with the teams as a possible solution to problem 306. 

ESA_Neil prepares the launch campaign laptops for use with the test ground station. 

537






ESA_Neil checks out the spare ground station TNC, it works fine. 

ESA_Neil tests the new beta version of the binary pulse beacon decoder – it does not work 

fine and the errors are reported to the programmer. 

ESA_Neil performs one thermal cycle on the flight model battery, from RTP to vacuum 

(approximately 1m mbar) and 50 degrees Celsius for two hours, then back to room 

temperature. 

ESA_Neil prepares and integrates the new credits plate 

538






The S-Band antennas are reconnected ready for flight. 

20 th June 2005 

ESA_Neil performs a second thermal vacuum cycle on the flight battery, and removes the –y 

lateral panel for access. 

ESA_Jason prepares three grounding straps, which ESA_Neil applies to UHF, S-Band and the 

PCU. A functional test proves that these do not alter the spacecraft behaviour. 

539






ESA_Neil torques / tightens (where torque wrench will not fit) all bolts, inspects and cleans in 

all compartments. 

ESA_Neil torques all interior bolts on the lateral panels, then inspects and cleans them. 

ESA_Neil re-glues all re-torqued bolts. 

The second thermal cycle on the battery is finished. 

21 st June 2005 

ESA_Neil performs a third thermal cycle on the flight battery. 

ESA_Marie and ESA_Neil prepare the shipping check-list. 

ESA_Neil takes photos of the whole of the inside of the sat, films a tour with ESA_Marie, 

and then closed the –x, +y and +x laterals. 

540






A 1200W Newton 1200 Followspot is set up to try and simulate sunlight. When disconnected 

but lit the –x panel presents 33.8V. 

When the panel is connected to the spacecraft the maximum current sourced is 68mA, at 

25.15V. At 75% intensity, it is down to 30mA at 24.7 volts. 

22 nd June 2005 

ESA_Neil replaces the –y panel temporarily and conducts some solar panel tests with the 

Newton 1200 followspot (http://www.sgm.it/eng/fs_seguipersona.htm). 

The solar panel simulator is turned off. Panel –x is lit and power is drawn from pin 17 of the 

EPS FPP connector while the spacecraft is in nominal mode. The line runs across a voltmeter 

and through an ammeter before returning to pin 4 of the EPS FPP. It consumes about 69mA 

at 25.3V. 

541






In order to test the internal wiring of the PCU the return feed from the panel is fed 

consecutively into pins 4 to 11. All of them consume similar amounts of power, +/- 1mA. 

This implies that the internal solar panel wiring of the PCU is correct. 

In order to test the redundant wiring on the –x solar panels the output from the panels is taken 

from pin 18 instead of 17. The power consumption is almost the same, this implies that the 

redundancy is present and correct, despite problem 281. 

The output is measured from pin 19 and the spacecraft is rotated to light the +y panel. The 

current sourced from the panel is 104mA at 25.7V. Pin 20 is almost identical, this implies 

that the redundancy is present and correct. 

The output is measured from pin 21 and the spacecraft is rotated to light the +x panel. The 

current sourced from the panel is 71mA at 25.16V. Pin 22 is almost identical, this implies 

that the redundancy is present and correct. 

The output is measured from pin 23 and the spacecraft is rotated to light the -y panel. The 

current sourced from the panel is 99mA at 25.5V. Pin 24 is almost identical, this implies that 

the redundancy is present and correct. 

542






NOTE: All the panels have a very similar power rating when their products are scaled by the 

different number of strings (2 strings per x side and 3 strings per y side). 

At 45 degrees incident light the +y panel provides 75mA at 25.0V. At about 20 degrees it 

produces 31mA at 24.5V. 

With the following colour lights (using the standard filters in the Newton 1200), it produces 

the following current and voltage: 

Colour Current Voltage 

White 104mA 25.5V 

Red 26.6mA 24.4V 

Green 36.0mA 24.5V 

Magenta 49.9mA 24.7V 

Blue 41.3mA 24.6V 

Cyan 50.1mA 24.7V (doesn’t look like cyan to me) 

Yellow 88.7mA 25.2V 

543






The technical specifications for the Netwon 1200, found in the user manual 

(http://www.sgm.it/eng/fs_seguipersona.htm), specify that the lamp has a luminous flux of 

110,000 lm. One lumen is equivalent to 0.001464W (this value is somewhat frequency 

dependent, but is a good value for white light), so the light output of the lamp itself is about 

161W (7.45% efficiency). However, we must bear in mind: 

1) That this is the spherically homogenous output, not all of which will be 

directed out of the front of the spotlight (picture shows some light going the 

wrong way). Estimated maximum efficiency: 0.75. 

2) There will be some losses in the system of mirrors and lenses. Estimated 

efficiency: 0.9. 

3) The output spectrum is not the same as natural sunlight, and experts 

(ESA_Monica) advise that at least one junction of the cells will be quite 

limited as a result. Estimated efficiency: 0.75. 

The fraction of output power remaining after all of these losses will be collectively referred to 

as the “test_efficiency”. 

544






The beam width projected onto spacecraft is about 85cm diameter. The distance has been 

carefully adjusted so the lamp is as close as possible, with a fully open iris, while still lighting 

all of the cells on one side. The area of the projected beam is therefore 0.567m 2 . 

Light intensity on spacecraft = followspot output / projected area 

= (161 * test_ efficiency) / 0.567 

= test_ efficiency * 284Wm -2 

Power sourced from +y panel = 0.104 * 25.5 = 2.65W 

Photovoltaic area on +y = 0.064 * 0.039 * 45 = 0.112 m 2 

Incident power on photovoltaic area 

= light intensity * area = test_efficiency * 31.8W 

Measured efficiency of solar panels (after harness and protection diode losses) 

= power sourced / incident power = 0.083 / test_efficiency 

Supposed efficiency of solar panels (after diode and harness losses) = 0.2 

Measured test_efficiency 

= measured panel efficiency / supposed panel efficiency 

= 0.083 / 0.2 = 0.42 

Estimated test_efficiency (from assumptions (1), (2) and (3) above) 

= 0.75 * 0.9 * 0.75 = 0.51 

A much simpler way of looking at this is as follows: 

Solar intensity in low Earth orbit = 1400 Wm 2 

Fraction of solar intensity supplied during test (using measured test_efficiency) 

= 284 * 0.51 / 1400 = 0.103 

Expected peak power input in orbit = power sourced in test / fraction of solar intensity 

= 2.65W / 0.103 = 25.7W 

This seems well within the margins of experimental error. We can probably assume that the 

solar panels are performing as they should. This analysis is forwarded to EPS, ACDS and 

SYS for discussion. 

ESA_Neil turns off the followspot and powers the spacecraft with the solar panel simulator 

again while OPER practise operations and develop the MCC with the MCC team. 

ESA_Neil turns off the spacecraft, disintegrates the battery box, then powers the spacecraft 

externally so that OPER can continue to practise. 

545






The battery box is opened and ESA_Jason removes the engineering battery and attaches the 

flight battery in its place. 

NCUBE_Aage delivers NCube-II to the cleanroom. 

ESA_Neil and NCUBE_Aage prepare a strip of aluminium tape to help protect against the 

antennas deploying through the access hole of the T-Pod again. The Cubesat is then loaded 

into the pod – the fit is good. 

546






ESA_Neil secures the lid and primes the pod. 

ESA_Neil closes the battery box, torquing and gluing all bolts into place and taking care to 

make sure that the base of the box is flat, with all bolt heads sunk, so that it can have good 

thermal contact with the shear panel. 

547






The battery box is reintegrated, this time not using washers to space it off from the wall. The 

bolts are torqued and glued and the connectors re-connected. 

ESA_Neil closes, torques and glues all four lateral panels. The spacecraft is powered down 

and the battery charging procedure begins. The flight battery is initially at 22.16V, and 

consumes 556mA to start with. 

NOTE: EPS_Fulvio advises that the battery was half-charged once before to test it, so this 

charging is a continuation of that first cycle. 

23 rd June 2005 

The flight battery has charged to full capacity overnight and is now consuming only 1mA and 

has a voltage of 24.66V. 

548






ESA_Neil powers up the spacecraft and gives OPER control while the battery runs down. We 

wait for collapse to safe-mode. This takes about 3.5 hours. 

The solar panel simulator is connected to charge up the battery (14:20), audio recordings are 

made of occasional safe mode binary pulse beacons. 

ESA_Neil removes the small pieces of kapton tape that were used to secure the solar panel 

harness during gluing, and then touches up the harness support in various places with Scotchweld. 

The knots in the nylon thread on the T-Pods are also secured. 

Extra kapton tape is added to the baseplate in order to improve the percentage coverage for 

thermal reasons. 

549






A small rectangle of aluminium is glued onto the baseplate in the location where the Polyot 

deployment switch will interface to it. 

The spacecraft is still in safe mode at 20:30, the battery voltage is 23.90V. 

This is considered adequate proof that the flight battery is charging as it should. Logistics 

necessitate that the spacecraft is powered down and the remainder of the battery charge is 

done manually. (This is preferable to leaving it as it is overnight, since then the spacecraft 

would be powered up for long durations using the flight battery.) 

The spacecraft is left charging overnight. 

24 th June 2005 

The spacecraft has reached full charge. The charging station is disconnect and the spacecraft 

powered up externally for operations practise. 

OPER train with pre-flight-plan definitions. 

ESA_Neil independently tests the orientation of the passive magnet: 

- We need the top of the spacecraft to point at the north pole, therefore it should have 

the same polarity as the north-pointing end of a compass 

- If the top of the spacecraft has the same polarity as the north-pointing end of a 

compass then, when a compass is held over it, the north-pointing end should be 

repelled and the south-pointing end attracted. 

550






Test positive. 

ESA_Neil runs the RX line from the UHF transceiver in the groundstation to the line-in 

socket on a laptop and records, at the highest quality possible, one minute of nominal mode 

beacons, and then one minute of live streaming. 

By playing these recordings back into the TNC it is verified that the data can be decoded 

properly, since it is received by term.exe as normal. 

ESA_Neil cleans the solar cells, and the lateral panels, applies all safety items and declares 

the spacecraft ready to be shipped. 

ESA_Neil packs up all ground support equipment ready for Russia. 

27 th June 2005 

ESA_Tor and ESA_Neil unbolt the spacecraft from the integration table and lift it to a trolley. 

It is then wheeled down to the workshop where the crane is used to lift it into the transport 

container. 

551






A grounding strap is added, the side protectors are removed, a bubble-wrap bag is made to 

cover the spacecraft and all loose bolts are removed from the box. 

The lid of the box is placed and secured tightly. 

552






The ground support equipment is packed for shipment. 

ESA_Neil proceeds to the pub. 

For the next exciting instalment please see the long-awaited and much coveted SSETI Express 

Launch Campaign Logbook, coming soon to a good website near you. 

The End 

553






Solar panel preparations 

Test 

Flight 

Broken Unused Class 0a 0b 1a 1b 2a 2b 3a 3b 3c 4 5a 5b 6a 6b 7a 7b 8 9a 9b 10 

0 14 

3 13 15 4 11 11 4 15 

4 12 7 8 6 9 13 1 1 14 1 7 8 15 

19 11 

4 10 

0 8 2 

1 4 7 5 

34 Totals 7 0 7 8 6 9 13 1 1 15 4 11 14 1 7 8 15 11 4 15 

7 15 15 15 15 15 15 15 15 15 15 

Pending 

Soldered 

Laid down 

554






555






4 Harness 

Special issues: 

5) TCS harness from OBC needs to triple the “single wire” positive and return 

connections, and then have two redundant signal wires each. 

6) ACDS power harness is flipped 

7) ACDS magnetometer harness is split 

8) T-pod harness is joined at safe/arm RbF and then split again to pods 

9) A 25-pin d-sub connector will have to be added to the ACDS system. The pins 

for the harness connector will have to be inserted into the harness connector 

during integration of the side panels. 

10) EPS harness between PCU, ACT, FPP and BATT is split after discussion with 

EPS_Fulvio and ESA_Neil. Now the harness is split such that the battery line 

(four wires) comes from the battery, to the FPP, to the ACT, and back to the 

battery. The panels come from the PCU to the FPP and back, and from the 

PCU to the ACT and back. 

Cable_name Quality Pinout Length Backshell Integrated 

P-PCU-PIN-1 EM All 25 straight 500 No Test 

P-PIN-OBC-1 EM Straight on pins 500 No Test 

1,2,3,9,10,11 

P-PIN-UTR-1 EM Straight on pins 500 No Test 

1,2,3,9,10,11 

P-PIN-STX-1 EM Straight on pins 500 No Test 

1,2,3,9,10,11 

P-PIN-PIC-1 EM Straight on pins 500 No Test 

1,2,3,9,10,11 

P-PIC-MGT-1 EM Pins 1,2,3 to pins 500 No Test 

9,10,11 and vice 

versa 

P-PCU-PIC-1 EM Straight on pins 500 No Test 

1,2,3,9,10,11 

R-OBC-EPS-1 EM Straight on pins 500 No Test 

2,3,8,9 

R-OBC-UTR-1 EM Straight on pins 500 No Test 

2,3,8,9 

R-OBC-STX-1 EM Straight on pins 500 No Test 

2,3,8,9 

R-OBC-FPP-1 EM Straight on pins 750 No Test 

2,3,5,8,9 

CAN-OBC-PIC-1 EM Straight on pins 

2,7,4,9 

500 No Test 

556






CS-PCU-UTR-1 EM Straight on pins 500 No Test 

1,2,6,7 

CS-UTR-UAN-1 FM Single coax TBD N/A No 

CS-UTR-STX-1 EM Straight on pins XXX 500 No Test 

CS-STX-SA1-1 FM Single coax 750 N/A No 



P-PCU-ACT-1 

P-ACT-FPP-1 

P-FPP-PCU-1 

P-BAT-PCU-1 

P-ACT-BAT-1 

EM 

See 

ftp/Express/EPS/Sseti 

_EPS_Electronics/ 

Harness/ 

Many No Test 

P-PCU-PIN-1 EM All 25 straight 500 No Test 

P-PCU-EB1-1 ½ FM 1,2,3,9,10,11 1000 No No 

P-PCU-EB2-1 ½ FM 1,2,3,9,10,11 1500 No No 

P-PCU-EB3-1 ½ FM 1,2,3,9,10,11 1000 No No 

P-PCU-PIN-1 ½ FM All 600 No No 

P-PCU-PIC-1 ½ FM 1,2,3,9,10,11 750 No No 

P-PIN-UTR-1 ½ FM 1,2,3,9,10,11 500 No No 

P-PIN-CAM-1 ½ FM 1,2,3,9,10,11 1000 No No 

P-PIN-STX-1 ½ FM 1,2,3,9,10,11 1250 No No 

P-PIN-MGT-1 ½ FM 1,2,3,9,10,11 1250 No No 

P-PIN-OBC-1 ½ FM 1,2,3,9,10,11 1750 No No 

P-PIN-PIC-1 ½ FM 1,2,3,9,10,11 1000 No No 

P-PCU-FPP-1 ½ FM 1,2,3,9,10,11 1000 No No 

R-OBC-PCU-1 ½ FM 2,3,8,9 1500 No No 

R-OBC-UTR-1 ½ FM 2,3,8,9 1250 No No 

R-OBC-CAM-1 ½ FM 2,3,8,9 1000 No No 

R-OBC-STX-1 ½ FM 2,3,8,9 1000 No No 

R-OBC-MGM-1 ½ FM 2,3,8,9 500 No No 

R-OBC-FPP-1 ½ FM 2,3,8,9 1000 No No 

CAN-OBC-PIC-1 ½ FM 2,4,7,9 600 No No 

½ FM 1 to 2, 2 to 7, 6 to 1, 7 750 No No 

CS-PCU-UTR-1 

to 6 

CS-OBC-MGT-1 ½ FM 1,2,3,4 2000 No No 

CS-OBC-MGT-2 ½ FM 1,2,3,5,6,8,9 2000 No No 

557






5 Problems 

1 

In panel –x+y a mistake was made in marking and cutting of side 

pockets four of the pockets are 7mm misplaced 

03/10/2004 

2 In panel +x+y the strip-mounted side inserts do not fit properly 03/10/2004 

3 Paper stuck to honeycomb around side pockets! 06/10/2004 

4 Some holes not filled completely with glue 06/10/2004 

5 

6 

7 

8 

9 

One large drip of glue was missed and had run all the way down the side 

and stuck to the cardboard beneath (arg!) 

Accidentally left glue too long with ‘excess blobs’ on the top they 

hardened a lot Also the remains of the glue in the mixing pot hardened a 

lot So all of S9 inserts have a problem 

Just after gluing ESA_Neil notices that the inserts in the –x side of the +y 

shear panel are the wrong way around 

SYS_Joerg counted side inserts and there are not enough we need to 

order 4 more at least (or make in workshop here) 

Rest of centre insert potting cannot proceed until ordered inserts have 

been delivered (before Wednesday we hope) 

6/10/2004 

7/10/2004 

09/10/2004 

9/10/2004 

9/10/2004 

10 The integration manual specifies M4x16mm bolts but the holes are M5 9/10/2004 

11 

12 

13 

The integration manual specifies three M5x16 bolts only for the midheight 

bracket #12 however the insert in the –y panel (and the drawings 

that led us to put the insert there) is an M4 insert 

The bracket that shares a mounting point with the –x+y top-bracket is 

mis-aligned with its bolt hole in the –x0y shear panel by a few 

millimetres 

There is a slight mis-alignment of the ‘lower-right’ screw on each of the 

protectors such that the bolts could be fixed but it puts unnecessary 

torque on the mounting points (side inserts) 

9/10/2004 

9/10/2004 

9/10/2004 

14 Some glue in the central holes of the ASAP inserts 10/10/2004 

15 

16 

The thruster mounting plates have one bolt hole too large for the 

associated bolt (the one over the thruster insert should be M4 instead of 

M5 

Uppermost (+z) side pocket on the –x side of the –y panel is originally 

cut 7mm too high by mistake 

13/10/2004 

14/10/2004 

558






17 

For most of the FM we have been using thinner (width) kapton tape and 

two strips were needed to cover each insert Glue has run down the join in 

between each pair of kapton strips 

15/10/2004 

18 Several of the bolts are extremely tight and difficult to get into place 16/10/2004 

19 

20 

On the –x thruster insert cluster some glue has run onto the top of the 

panel from one of the inserts 

The mounting plate on the +x thruster insert cluster has become stuck to 

the base-plate and cannot be removed with bare hands 

17/10/2004 

17/10/2004 

21 Some glue is left on the surface of the baseplate 17/10/2004 

22 

23 

24 

The bolts used to secure the separation ring to the EM structure (same as 

will be used for FM) are actually 5mm longer than planned This could 

potentially interfere with the propulsion tubing 

Westend BV discovered that they used alloy 6082 in the manufacture of 

all recent inserts!(Instead of 7075)ESA_Neil asks STRU_Antonio for 

advice on this issue 

STRU_Melro advises that the inserts made with alu 6082 are not suitable 

for the flight model as the strength is almost 50% less than that of 

7075This renders “FM” –y, -x0y, +x0y and -x-y as useless 

17/10/2004 

19/10/2004 

20/10/2004 

25 TCS thermistors do not have a power supply defined! 22/10/2004 

26 

27 

28 

29 

30 

No external communications on the computer are working and the utility 

processor simply oscillates instead of exchanging proper CAN messages 

EPS_Fulvio does not bring the Battery Charge Regulator as it is still not 

finished Must chase this up tomorrow 

The holes on the ACDS magnetorquer driver do not match up with the 

holes on the mounting brackets 

The ‘reset’ wire that OBC_Karl attached to the FM OBC yesterday 

accidentally touched a positive power connection This fried the ARM 

processor and the flash chips on board 

EPS_Fulvio has lost all of his components and can’t start soldering his 

board We have to order them again 

23/10/2004 

24/10/2004 

24/10/2004 

24/10/2004 

24/10/2004 

31 The glue from the kapton is often left behind on the surface 25/10/2004 

32 

33 

The FM titanium ring is slightly to large for the exposed area of panel 

and instead the edges rest on the layer of thermal paint at the outside 

The M6 bolts between the separation system and the satellite do not run 

smoothly through the base brackets – some are extremely tight 

25/10/2004 

25/10/2004 

34 ACDS power connector is inverted 26/10/2004 

35 The dummy antenna is too wide beneath the mounting plate and does not 26/10/2004 

559






36 

37 

38 

fit into the insert Maybe the flight one is the same size! 

The mounting points on the –y lateral panel for ACDS coil driver do not 

match up to the size of the board These will have to be re-positioned 

The ACDS ‘internal’ connectors are not flight-worthy but hard-soldering 

is not feasible since we don’t want to have the thread fragile items like 

sun-sensors through harness holes 

The ACDS magnetometer has a cadmium connector that is not suitable to 

fly in space as it will off-gas severely 

27/10/2004 

27/10/2004 

27/10/2004 

39 Cannot establish proper communication between OBC and CAM 27/10/2004 

40 

Communication cannot be established with the EM MAGIC because both 

ends have a CAN termination We phone MAGIC_Renato to ask if 

OBC_Karl can de-solder the CAN termination on the EM 

MAGICMAGIC_Renato says yes 

27/10/2004 

41 The rivet nuts don’t fit into the holes on the corner profiles 27/10/2004 

42 

43 

44 

45 

46 

47 

The M4x8mm bolts to fix the lateral panels to the primary structure are 

slightly too long for the inserts and therefore protrude a couple of mm 

past the lateral panels allowing them to move away from the satellite 

body 

Many of the bolts are a very tight fit because of tolerance problems 

potting the side inserts However on the whole EM structure only two 

bolts do not fit at all which is probably quite a good result These two 

however are the centre two on one side of the top-plate which is a little 

worrying as it means that the top-plate is flexing down at the corners 

CAM_Morten and OBC_Karl killed the FM CAM computer board 

Testing suggests that the FPGA failed 

The bolts are not really long enough to extend far enough into the rivet 

nuts after they have gone through the corner profiles 

There are not enough M6 nuts for the rest of the primary structure 

integration – we are short by one. 

The mounting bolts for the tensioning of the tank are not 25mm as 

specified by PROP, but only 13mm as specified by STRU_Antonio (so as 

not to interfere with the +z T-Pod e-box). 

27/10/2004 

27/10/2004 

28/10/2004 

28/10/2004 

28/10/2004 

31/10/2004 

48 Thruster inserts have M4 threads instead of M5 threads. 31/10/2004 

49 

50 

The conical low-pressure tubing mounts do not sit right down onto the 

baseplate. This is because the tubing was bent by hand and is not accurate 

to the millimetre. 

One of the low-pressure tubing mounts is not entirely touching the 

baseplate – this one will be slightly weaker than the others, but should be 

31/10/2004 

31/10/2004 

560






within acceptable limits. 

51 The PMS and of the high pressure tubing is slightly misplaced. 01/11/2004 

52 PROP_Hanno is not happy with the stress on the high-pressure tubing. 01/11/2004 

53 

54 

The magic box uses flash memory, which may not fare too well in space. 

OBC_Karl and MAGIC_Renato investigate ways to replace it. 

One-time-programmable chip not available as replacement for Texas 

Instruments DSI in the Magic box. Solution: MAGIC_Renato to 

investigate boot-loader options. 

04/11/2004 

05/11/2004 

55 The modem in the EM UHF box also appears to have flash on-board. 05/11/2004 

56 

57 

58 

59 

60 

EPS_Fulvio and ESA_Neil discover that when proper redundancy is 

applied to the connector leaving the PCU for the RBF connector, there 

are not enough pins 

The OBC flight proms have not been added, so it is required to load 

software into to memory upon every boot-up. This upload takes about 50 

seconds, which is longer than the EPS-OBC watchdog period. Therefore 

the OBC currently cannot be powered up directly from the PCU. 

When progressing through the SECOND (first is ok) cycle of recovery 

mode to nominal mode, EPS sends a shutdown command to OBC even 

after OBC responds to the watchdog ping. Current workaround: cycle 

power to the PCU 

The power levels seem dodgy. For example, when MAGIC is off it still 

gets 1.4 volts through the PIN 

The S-band unit fails to power up using the EPS. This is found to be due 

to a large inrush current on the DC-DC converter that EPS is treating like 

a latch-up. 

07/11/2004 

07/11/2004 

07/11/04 

07/11/04 

08/11/04 

61 The S-band units needs ground on the RS232 line. 08/11/04 

62 The S-band unit requires a ground on the audio line with UHF. 08/11/04 

63 

64 

65 

The OBC experiences a line driver failure on the EPS and UHF ports. 

This is probably due to an unconnected cable from the port acting like an 

antenna and running lots of powerful RF into the computer. 

AMS_Sam and AMS_David discover that the third harmonic from the S- 

BAND unit is only 30dB below the carrier. 

Packets downlinked from UHF or S-BAND are valid frames but 

gibberish. This is because the protocol is not implemented correctly on 

the OBC. 

08/11/04 

08/11/04 

08/11/04 

66 The holes in the LGAs do not quite match the mounting points on the 09/11/04 

561






67 

68 

69 

70 

71 

72 

73 

74 

75 

76 

77 

78 

mounting bracket on the top plate. 

The structure of LGA looks far to sensitive to mechanical loads and 

vibrations, we expect that they will fail during the vibration testing. 

A significant amount (about 10%) of the RF from the antennas is 

transmitted backwards into the satellite. This will cause problems for the 

on-board subsystems, and the compartments will act as waveguides. 

The nominal mode beacon has wrong call sign (AMS_Jason’s), and 

contains incorrect information about the website. 

The PROP team get back to IABG and find that the tank is now at only 

33 bar. It is leaking from below the tank, where an o-ring has popped its 

housing. 

The PROP team cannot torque the tank properly (step 5.2 from required 

modification listed above) because of bad access through the structure. 

The exact length of telecommand uplink that gets recognised by OBC is 

indeterminate and non-repeatable. OBC_Karl is made aware of this by 

ESA_Neil. 

Randomly, twice, a particular ASCII string is sent which causes the OBC 

to kill the nominal mode beacon (otherwise the OBC seems fine). This 

should not be possible. OBC_Karl is made aware of this problem by 

ESA_Neil. 

On the rare occasions where an attempt at uploading a telecommand is 

actually successful in placing the command into the flight plan, it 

overwrites any command that is already there – then resulting in a 

maximum of one TC in the flight plan at any given time – which is 

clearly not enough for the mission. OBC_Karl is made aware of this 

problem by ESA_Neil. 

EPS_Fulvio accidentally powers a 5V board from a 28V supply. This 

appears to damage either the MAX232 line driver or the PIC. After 

further testing it is determined that the PIC is partially damaged (still 

partly functional). 

Occasionally get looping "de-kissisifcation error" down the debugger 

upon startup, and once we do it doesn't get better. This seems to be 

resolved once the linux box is rebooted. 

When using TCINS to send commands to EPS while looking down the 

EPS port with a laptop the parameters get swapped around. Is this 

another least-sig / most-sig issue, or is this intentional? (It is probably 

general to all tcins, but this was the case I discovered.) 

Upon startup of OBC the bytes "0f 6f 6f" get sent to EPS. Why is this? 

(It even appears in the flight planner briefly, I caught it once.) This is 

09/11/04 

09/11/04 

09/11/04 

10/11/04 

10/11/04 

10/11/04 

10/11/04 

10/11/04 

11/11/04 

13/11/04 

13/11/04 

13/11/04 

562






79 

80 

81 

82 

83 

84 

85 

86 

87 

why EPS was shutting you down sometimes... now it just counts errors... 

I have just about found a way to uplink HEX-based TCs to the OBC via 

the UHF, but it is still a little shaky (is only half tested, the RF link is not 

strong enough to be THAT reliable, and it has to end in a carriage return). 

I can get stuff into the flight planner quite well, and the right numbers 

seem to be appearing in the right places. Issue is: if I send a TC starting 

with "00" (i.e.: subsystem ID is OBC) then it does not acknowledge, does 

not go into the light plan, the nominal mode beacon stops, and it ignores 

anything incoming from UHF completely from then onwards. (Although 

it all looks fine on the debugger.) 

TCs from UHF are overwritten, not stacked. This would make for a 

tedious mission... 

After adding commands on TCINS to be executed immediately (which 

they do), I have found them still listed in the flight planner later on, 

which is a tad odd. Doesn't a TC get taken out the planner when it's 

executed? This has not been a repeatable error... unfortunately. 

We notice some structural damage to the exposed core on the +y side of 

the baseplate. We have no idea when this occurred, but it looks like an 

impact by something like a screwdriver handle. 

Upon visual inspection it is found the high pressure tubing mount closest 

to the tank has been torn off of the baseplate. 

Upon visual inspection it is found that the inserts holding the tank 

mountings have been pulled into the centre compartment by a couple of 

millimetres, therefore damaging the honeycomb panels slightly by pulling 

the core away from the skin. This is because the washers used on these 

bolts were too small, and within the circumference of the inserts, 

therefore not transmitting loads onto the honeycomb in the proper way. 

This is highly concerning structural damage. 

ESA_Neil notices that the PMS box mountings also have washers that are 

too small for the inserts. 

The mid-pressure system has depressurised. This is bad, but not critical, 

since the SSTL requirements are on the high-pressure system. The leak is 

intermittent, and only seems to happen when the filling hose is 

disconnected, but no leak can be found in that area. We decide to 

proceed with the shake anyway. 

The clamp between the high- and low-pressure tubing has come loose. It 

is simple to torque it back up again, but this is concerning because it was 

torqued previously and uses a “self securing” nut, like most of the rest of 

the PROP system. 

13/11/04 

13/11/04 

13/11/04 

15/11/04 

16/11/04 

16/11/04 

16/11/04 

16/11/04 

16/11/04 

88 When requesting X units of the ALARM stack, if there are less than X 18/11/2004 

563






89 

90 

91 

units then it simply repeats until X units are transmitted. 

The timestamp on downlinked telemetry seems to fluctuate from a 

sensible time, to sometime in 2033. 

When requesting camera data the debugger receives “camera data sent” 

repeatedly and then the OBC crashes. 

Using the “tcins” command (from the debugger) to flush the alarm stack 

causes the OBC to crash. 

18/11/2004 

19/11/2004 

19/11/2004 

92 The GND does not successfully converse with OBC 22/11/2004 

93 ACDS magnetometer does not fit properly. 22/11/2004 

94 UHF_Lars needs a network analyser. 22/11/2004 

95 Still loosing packets (OBC, UHF, GND) and lots of erroneous data 22/11/2004 

96 

The branch valve is not responding properly, making leak testing 

impossible. 

23/11/2004 

97 A second line driver fails in the OBC. 23/11/2004 

98 

After extensive investigation OBC_Karl and CAM_Morten conclude that 

the byte errors could be coming from an overflow in the utility processor 

on the OBC. A new chip therefore needs to be programmed, burnt, and 

replaced. 

23/11/2004 

99 Some bytes dropped during transfer from CAM to OBC. 24/11/2004 

100 

101 

102 

The FM CAM PCB grounding plane connects to the bolts that hold it to 

the box, therefore forming a ground loop through the EPS harness and 

then back through the structure. This would be very problematic, 

especially with the RF antennas so close. 

The ACDS coils are too small for the mounting pattern prepared on the 


Signal is just too weak for reliable reception at GND. (Although works 

fine the other way.) 

24/11/2004 

24/11/2004 

24/11/2004 

103 MAGIC_Renato locks the FM MAGIC processor. 24/11/2004 

104 Upon EPS power up, no pings sent, OBC not powered up 24/11/2004 

105 MCC interprets the nominal mode beacon and it looks like rubbish. 24/11/2004 

106 The new EPS software doesn’t work. At all. 25/11/2004 

107 S-Band carrier-up command suddenly doesn’t work 25/11/2004 

108 OBC won’t power up properly. 25/11/2004 

109 Branch valve not responding. 25/11/2004 

110 ACDS power-up works but there is a slight bleed current when turned 25/11/2004 

564






111 

112 

OFF (as noticed before), which slowly charges up the capacitors and 

bleeds into the system. 

CAM power-up works but there is a slight bleed current when turned 

OFF (as noticed before), which slowly charges up the capacitors and 

bleeds into the system, causing some oscillations. 

OBC_Karl decided that he doesn’t like this (problems 110 and 111) very 

much either. 

25/11/2004 

25/11/2004 

113 The CAM doesn’t fit in to the hole on the top of the spacecraft. 25/11/2004 

114 GND miss packets – maybe this is still a utility processor buffer problem. 25/11/2004 

115 

GND and MCC interpret picture – there is so much missing it is not 

recognisable. 

26/11/2004 

116 OBC_Karl notices the occasional missing byte in EPS telemetry, 28/11/2004 

117 

The test solar panel emerges from the thermal vacuum chamber 

significantly damaged. 

30/11/2004 

118 Push-to-talk wired up like an RS232 in the UHF box. 03/12/2004 

119 We need 150ms turnaround on half-duplex for the groundstation 06/12/2004 

120 E-Box does not hold itself on like it should. 06/12/2004 

121 

No combination of specified wires produces the correct result. 23 and 24 

power all loads as normal, but also power T-Pods for 10 seconds initially. 

23 and 25 eat 32 milliamps permanently and do nothing else. 24 and 25 

power T-Pods for ten seconds and then eat 32 milliamps permanently and 

does nothing else. 

07/12/2004 

122 Black holes in picture (data lost during transmission). 07/12/2004 

123 

124 

The upper part of the +y lateral panel interferes with the release 

mechanism of the +z T-Pod. 

The +x and –x side protectors no long fit, as the T-Pods protrude more 

than planned. 

07/12/2004 

07/12/2004 

125 Power-pin on analogue multiplexer not connected on ACDS coil-driver. 07/12/2004 

126 

Back-shields do not fit through the holes in the +z LGA mounting plate 

and the top-plate, due to minimum radius in the corners of these holes 

when machining. 

13/12/2004 

127 LGA holes don’t line up very well with the +z LGA mounting plate. 13/12/2004 

128 

129 

The +z LGA mounting plate cannot be mounted to the top-plate as the 

washer on the +x+y lifting bolt is too wide. 

The battery uses one or two layers of thin plastic sheathing as insulation 

between cells. If this off-gasses and ‘dissolves’ then the battery will fail. 

13/12/2004 

14/12/2004 

565






130 

131 

132 

133 

134 

Quite a lot of plastic insulation is also used on the wires connecting the 

small circuit boards on the top of each cell. 

The OBC is not working at all, there is no TM, no flight plan execution, 

and a “tcins” causes it to crash. 

The OPEN_FILE command uses a target comms system parameter, 

which then means that only that comms system can be used with the 

GET_FILE command. This is too restrictive and should be the other way 

around. 

When S-BAND attempts to bring the carrier up and transmit right away 

the power spike is too high and EPS cuts it off. This happens when the 

carrier is down and the system in is voice config and attempts to send 

telemetry, and it happens when the carrier is down and the system is in 

data config and attempts to send telemetry. 

The battery box is at 14V instead of 0V. This is because the wall of one 

of the cells is touching the sides of the box through the kapton 

somewhere. 

Visual inspection reveals apparent damage during transportation. 

Situation reported to the NCube-II team and am awaiting appropriate 

response. 

15/12/2004 

16/12/2004 

17/12/2004 

17/12/2004 

12/01/2005 

135 The EM UHF antenna doesn’t fit in the mounting hole and is too long. 17/01/2005 

136 The UHF EM Antenna leans the wrong way compared to the FM. 27/01/2005 

137 

138 

139 

140 

141 

The PC in the radome controlling the antenna requires a login password 

and no-one in K-Sat or NCube can remember it. There is also a large 

time pressure, as the weather is closing in and the road back down from 

the plateau will soon be impassable. 

It seems that the only software on the PC is Nova and NetOp. The MCC 

team are not keen to interface to either of these programs, rather than 

creating their own system. 

It is not clear how the NCube guys are planning to control the radio, or 

how they are planning to stream the data back to Andøya. These two 

pieces of information are important to make sure that the SSETI Express 

MCC / GND interfacing solution does not interfere with the NCube 

setup, and vice versa. 

The various layers of firewall are getting in the way, and connection is 

not possible. The software at both ends is demonstrated to be correct by 

local connections. 

Although Nova can calculate the appropriate Doppler compensations 

during a pass, it does not appear to be able to control the radio directly. 

We therefore assume that NCube are not planning to adjust for Doppler, 

09/02/2005 

09/02/2005 

09/02/2005 

10/02/2005 

10/02/2005 

566






142 

143 

which leaves the SSETI Express teams with the requirement to come up 

with a solution. Subsequent telephone calls with the NCube team 

confirm this assumption. 

The firewall at K-Sat is also removing the possibility of using VNC for 

the Aalborg – K-Sat connection. 

SSETI Express cannot use the same solution as the NCube team, as we 

require control of the radio during a pass to correct for the Doppler shift, 

but Hyperterminal would not allow such control of a port it is connected 

to. This point is somewhat moot however, since we also require the data 

to be streamed real-time back and forth. A solution therefore needs to be 

found that can switch between the NCube software setup and the SSETI 

Express software setup. 

10/02/2005 

10/02/2005 

144 With the new OBC software no TM is generated, and no downlink at all. 15/02/2005 

145 

146 

147 

148 

149 

ESA_Neil boots up the OBC to test the latest software. The downlink 

works this time, but there is no TCS or ACDS TM. 

There seem to be a lot of “Fetch Queue Timeouts SID (4)” and “Hex 

dump” errors coming from the OBC to the debugger. 

Two-way communication with the S-band TNC cannot be established 

from an external laptop. 

After the baud rate and port change there is no data downlink from the S- 

Band unit. 

After the baud rate and port change we are only getting every second 

beacon from UHF (every 36 seconds). 

15/02/2005 

15/02/2005 

15/02/2005 

15/02/2005 

15/02/2005 

150 The uplink on UHF is extremely unreliable. 15/02/2005 

151 

The occasional packet in the picture chunk downlink is erroneous, 

therefore forcing the test groundstation software to automatically request 

the entire chunk again. The result is that it is usually 5 or 6 cycles before 

the entire chunk comes down trouble free. Suggestion is to only rerequest 

chunks once and then combine all good data. 

15/02/2005 

152 The spacecraft stops responding entirely to the UHF uplink. 15/02/2005 

153 S-band is producing almost no power to the antennas. 16/02/2005 

154 

When the S-band unit attempts to go from carrier down to transmitting in 

one go the EPS PCU appears to cycle the power as it trips the current 

limiting circuit. Therefore the transmission is not successful, and the unit 

ends up in an unpredictable state (since the power down is so short that 

the unit does not fully reset). This doesn’t make much sense though, as 

no real extra current is used when transmitting as when the carrier comes 

up alone – no explanation is forthcoming for some time. 

16/02/2005 

567






155 The PCU draws no current from the solar panel input. 17/02/2005 

156 

157 

158 

159 

160 

161 

162 

163 

164 

165 

166 

The PCU draws power right to the limit, and powers the shunt resister 

even though the input is only 25V. 

The downlinked thumbnails from the OBC are entirely black. Perhaps 

the parameters are not being sent correctly. 

Timers and voltage regulator may not be working properly, but we need a 

bigger current source to check for sure (which we don’t have). 

The antennas of NCube-2 have deployed during transit. This is a onetime-only 

physical mechanism, controlled by a one-time-only hardware 

timer, and therefore cannot be reset easily. It appears that the removebefore-flight 

pin is faulty, and the spacecraft must have activated during 

transit. 

The remove-before-flight pin on NCube-2 protrudes too much and 

interferes with the wall of the T-Pod, therefore not allowing integration. 

There are not quite enough bolts to integrate the S-Band baseplate (four 

M2x6mm short), or the S-Band lid (one M2.5x6mm damaged).. 

The power cut during eclipse is too long for the timers (tested for 40 

minutes), and the capacitor discharges, meaning that the timers would run 

again once in sunlight. 

The duration of the timers has been affected by the modification, it is 

now around 85 minutes instead of 74 minutes. 

The PCU collapses from recovery mode and consumes 150mA without 

powering any loads at all. Perhaps a loose cable touched a damaging 

potential. 

EPS_Tommy discovers that the linear voltage regulator powering the 

PDU is extremely hot. ESA_Neil switches off the power immediately. 

The PDU is consuming 235mA (it should be around 30mA). 

EPS_Tommy turns the power off. This implies that something on the 

PDU has failed, probably the PIC. 

18/02/2005 

18/02/2005 

22/02/2005 

24/02/2005 

24/02/2005 

03/03/2005 

04/03/2005 

04/03/2005 

05/03/2005 

05/03/2005 

05/03/2005 

167 The spare PICs do not have the bootloader on them. 05/03/2005 

168 EPS_Tommy continues to attempt to burn the PIC, but has no luck. 06/03/2005 

169 

170 

171 

We cannot test the safe-mode functionality as EPS team are not totally 

sure how to force and simulate safe mode at present. 

The recovery mode beacon is not being transmitted by the radio (there are 

no power spikes and it cannot be heard via a handheld). 

It turns out that problem 170 is due to an incorrect pinout on the PTT 

connector on the UHF FM enclosure. Tests show that pins 1 and 6 are 

07/03/2005 

07/03/2005 

07/03/2005 

568






172 

173 

174 

175 

176 

177 

ground and 2 and 7 are positive. Instead it should be 1 and 2 are positive 

and 6 and 7 are ground. This means that the PCU has been short 

circuiting itself every time it pushed the PTT and this could be why the 

PIC died. 

The +x lateral does not fit right, even when we loosen the corner profile. 

The same two bolts as the EM are misaligned (centre +z) and the also +y 

corner. 

Although two-way communication between the PCU and OBC is 

working (pings received by OBC and TC received by EPS), the PCU is 

not “hearing” the pongs sent by the OBC. This implies that the EPS 

software is incorrect. 

We appear to be loosing packets on the S-Band downlink, implying that 

the OBC is streaming data to the S-BAND unit too fast and we are 

overflowing the ‘leaking bucket algorithm’. 

The OBC cannot give data to S-Band at “any” speed, because the 

possible sleep times in between packets are only integer multiples of 

10ms (which is too long). 

The pings from the PCU are not answered by the OBC while it is busy 

storing the newly transferred picture into flash memory. 

ACDS_Lars needs an unobscured OBC tomorrow, but OBC cannot be 

unobscured until radio tunings are complete, but radio tunings cannot be 

completed while EPS is using the system 

07/03/2005 

07/03/2005 

08/03/2005 

08/03/2005 

08/03/2005 

08/03/2005 

178 Utility processor bottleneck cannot drive UHF past about 8kbps 08/03/2005 

179 

With transmission speed at just below 9k6 there is a significant error rate, 

probably partly from the internal conflicts and partly from the UHF 

loosing packets. 

08/03/2005 

180 We keep missing the acknowledge when we send a TC. 09/03/2005 

181 

182 

183 

184 

The UHF uplink reliability is extremely sensitive to other objects in the 

clean room. 

Part-way through a large downlink from S-Band the transmission stops 

and no more data can be sent without cycling the S-Band TNC. This is 

probably because the leaking bucket algorithm is overflowing the buffer. 

The OBC does not respond to the PCU pings while downlinking a picture 

on S-Band. 

If the OBC does not respond to two consecutive pings the EPS should 

send a “shutdown” command and THEN power it off. This shutdown 

command was not sent. 

09/03/2005 

09/03/2005 

09/03/2005 

09/03/2005 

185 Suddenly the spacecraft powers off for no apparent reason! Battery low? 09/03/2005 

569






Battery protected?! Then the battery is reading out at 18V, and seems to 

charge quite rapidly to 20V or so but no more. 

186 The nominal mode beacon data was not being sent correctly from EPS. 09/03/2005 

187 

188 

189 

190 

191 

192 

193 

194 

195 

196 

197 

198 

199 

200 

The OBC flight proms do not work as they should – there are some 

serious timing issues due to an unexpected difference between the flight 

and non-flight chips. 

The SSTL rear-solder jig is the wrong size and shape for the Tecstar 

cells. 

There is no nominal mode beacon data being received by the OBC from 

EPS. Instead there is an occasional MID 04 in the TM, and the 

occasional fetch queue timeout on SID 2. 


EPS. Instead there is an occasional MID 10 in the TM, and the 

occasional fetch queue timeout on SID 2. 


EPS, even though it is being sent. (Although there are no timeouts now.) 

MCC are not ready for remote testing of the camera and need some 

software modifications. 

Solar panel substrate 1 FM and 1 EM interfere with the sub-sensor, and 

do not reflect the ACDS coil driver location modifications. 

Solar panel substrate 7 FM has been dropped and has some damage at 

one corner. 

There is no attenuation on S-Band. This could be causing some TX 

reflections and radiation problems for personnel. 

During disintegration the head of one of the cross-head bolts used to 

mount the sun sensors is partially stripped. 

Solar panel substrate number 3 does not have the holes for the –y sun 

sensor mounting in the right place, and does not have the harness hole for 

the sun sensor harness at all. 

Solar cell string 3a interferes with the sub sensor since the original 

location has been modified by ACDS but not reported back to STRU. 

When STRU_Melro designed the solar panel substrates he did not know 

that there should be 1mm gaps between the cells, or that there were 5mm 

connectors on each end. Subsequently several of the strings are actually 

slightly too long for the panels. 

The holes in the base of the MAGIC box are not large enough to 

accommodate an M4 nut, only a bolt head. 

10/03/2005 

10/03/2005 

15 th March 

2005 

15 th March 

2005 

15 th March 

2005 

16 th March 

2005 

16 th March 

2005 

16 th March 

2005 

16 th March 

2005 

17 th March 

2005 

17 th March 

2005 

17 th March 

2005 

17 th March 

2005 

18 th March 

2005 

570






201 

202 

203 

204 

205 

206 

207 

208 

209 

210 

211 

212 

213 

214 

215 

216 

The +z-y bolt hole in the top-plate of MAGIC is slight misaligned (or the 

bolt is slightly bent). 

The bolts are too tight to integrate the MAGIC box into the FM structure. 

The combination of several tolerances must have led to a misalignment. 

(Probably the insert potting.) 

The nut on the batter charge stud is too tight and cannot be applied. 

A bolt from one of the mid height brackets is protruding through the 

honeycomb wall into the back of UHF (or it would be if it was there). 

A fit check with the -y lateral panel reveals that one of the bolts is a little 

too tight for comfort, it is marked for future modification. 

It is not possible to integrate the +x lateral panel after the -y lateral panel 

because the +x magnetorquer coil needs to slide behind the +x-y corner 

profile (attached to the -y panel), and the +x panel cannot move like that 

because the +x T-Pod is in the way. 

A fit check with the +x panel reveals that an extra bolt is too tight, in 

addition to the three already noted above (problem 172), they are marked 

for future modification. 

A fit check with the -x lateral panel reveals one bolt that is too tight (the 

one in the -y-z corner), it is marked for future modification. 

The battery, once covered in kapton, is too big for the battery box. It is 

not possible to close it. 

Cannot find the fixation bolts for the patch antennas. Using standard 

bolts for now as placeholders. 

The RTV is not properly securing the wires on the last point before they 

leave solar panel 1. 

STRU_Antonio notices that there are only four bolts holding MAGIC 

together, we need two more. 

We are still getting nominal mode beacons, but no TC are making it to 

the OBC. We have no idea why. 

The OBC has been shutdown by EPS and we are in recovery mode. This 

implies that two pings were missed. Why? 

After problem 214 ESA_Neil brings the OBC back up – but no pings are 

being received from EPS at all - why?! 

The uplink is so unreliable that it is unusable. 

18 th March 

2005 

18 th March 

2005 

18 th March 

2005 

18 th March 

2005 

20 th March 

2005 

20 th March 

2005 

20 th March 

2005 

20 th March 

2005 

20 th March 

2005 

20 th March 

2005 

20 th March 

2005 

21 st March 

2005 

21 st March 

2005 

21 st March 

2005 

21 st March 

2005 

21 st March 

2005 

217 The linux box finally dies. It sounds like it would work well as a coffee 21 st March 

571






218 

219 

220 

221 

222 

223 

224 

225 

226 

227 

228 

229 

230 

231 

grinder though. 2005 

Problem 213 rears its UGLY head again. We are getting nominal mode 

beacons, but cannot uplink at all. 

One of the small bolts securing the CAM PCBs to the enclosure snaps 

upon tightening. 

Problems 214 and 215 manifest themselves again. The OBC is power 

cycled regularly as if the ping/pongs are not functioning correctly. (This 

happened during a UHF full picture downlink, but previous experience 

suggests that this is not the reason.) 

EPS software version 7.4 still resets the OBC when the keepcount 

variable rolls over. After that it is stable though. This must mean it has 

been reduced to a boundary problem. 

The downlink on UHF is very dodgy all of a sudden. Whenever 

significant UHF downlink reaches the term.exe it crashes. 

There is not really enough space for the RF cable or the RS232 cable 

from UHF, as it is too close to the shear panel. 

The saver on the audio port (to S-Band) on the UHF is in the way of 

integration of the PIN. 

The holes in the PIN are large than they should be, causing potential load 

bearing problems. 

The downlink is not working at all. When we listen to it on a handheld 

radio there is a strange “squeak” whenever a packet is sent. 

The PCU software is coded such that it is assumed that the 0-5V reading 

of the battery voltage must be multiplied by 6 in order to get the actual 

value. In fact the BCR is only dividing the voltage by 5 in order to give 

the reading. Therefore the safe mode entry threshold, the safe mode exit 

threshold, and the battery broken threshold are all incorrect. This is not 

acceptable. 

The new antenna bolts are too narrow and do not securely hold the clips 

that fix the attenuation caps on. The old bolts were 7.67mm in diameter, 

but the new ones are only 6.88mm. 

The PCU is closer to the walls than we realised and it doesn’t fit because 

there are harness clamps in the way. 

The holes in the PCU L-Profile do not match those in the lateral panel. 

The PCU is consuming a very strange current (190mA), even when the 

load cable is not plugged in. This implies that it is not functioning 

correctly – which is practically irreversible after the conformal coating. 

22 nd March 

2005 

22 nd March 

2005 

23 rd March 

2005 

23 rd March 

2005 

26 th March 

2005 

27 th March 

2005 

27 th March 

2005 

27 th March 

2005 

27 th March 

2005 

29 th March 

2005 

29 th March 

2005 

2 nd April 

2005 

2 nd April 

2005 

3 rd April 

2005 

572






232 

233 

234 

235 

236 

237 

238 

239 

240 

241 

242 

243 

244 

245 

Once the thermistors are plugged in the OBC will not boot up. 

The +y sun sensor harness is not long enough. 

ESA_Jason’s glove melts onto the washer he is applying while bolting in 

the activation switch plate. This is because he accidentally connected the 

charge stud with the ground plate, shorting across the battery terminals. 

The spacecraft still cannot transmit properly, and there is the “beep” 

detectable on a handheld radio when it tries to, just as in problem 226. 

This implies that the lateral panels do not add enough attenuation to fix 

the problem. 

Around 50% of the time when the UHF unit is power toggled manually it 

comes up already transmitting at low power on an unknown frequency, it 

is then not possible to receive or transmit normally. 

Problem 235 reoccurs 

PROP does not have a hose that is proof pressure tested to 450 bar, so 

there can be no connection from the ground support equipment to the 

spacecraft. 

PROP_Sascha was not expecting such a lengthy decoding process from 

the MAGIC telemetry to actual data. The conversion done manually at 

each reading would necessitate an unacceptably long filling procedure. 

The high pressure system reads at 213 bar, but the gas bottle is fully open 

so we should expect around 267 bar. Also, the temperature is not 

climbing as high as expected. 

One of the kill switches from UWE-1 is found laying next to the +x 

thrusters cluster. It must have detached from the spacecraft and fallen 

through the access hole in the T-Pod. 

The NCube-2 antennas and gravity boom have deployed through the 

access port into SSETI Express. Their RBF pin is still present 

An accelerometer detached from the +Z antenna during the vibration, and 

has slightly dented the surface. This damage is probably superficial, but 

must be checked with COMMPL. 

One of the +x T-Pod cable ties has broken and the Teflon block is loose 

in the base of the pod. 

One of the -x T-Pod cable ties has broken and the Teflon block is loose in 

the base of the pod. 

4 th April 

2005 

4 th April 

2005 

5 th April 

2005 

5 th April 

2005 

9 th April 

2005 

9 th April 

2005 

10 th April 

2005 

10 th April 

2005 

12 th April 

2005 

12 th April 

2005 

13 th April 

2005 

13 th April 

2005 

13 th April 

2005 

13 th April 

2005 

573






246 

247 

248 

249 

250 

251 

252 

253 

254 

255 

256 

257 

The fill and drain valve cap has fallen from the PMS box. It had not been 

torqued properly. 

Xi-III has loose objects of significant mass inside it, it rattles loudly. It 

was probably this causing most of the noise in the last two low level 

sweeps. 

We cannot reinsert Xi-III as it is clearly damaged, but do not have any 

more mass dummies to replace it with. 

There is a lot of noise during the low level sine sweep, it seems to be the 

mass dummy in the +y T-Pod. This could be throwing out the control 

curves, and could potentially add increased loads to the local shear panel. 

Low sine sweep reveals a resonance at 44Hz (expected more like 50Hz). 

Using Miles formula (equivalent static load from a random test) we 

would get 21g at the first resonance. This would almost certainly cause 

damage to the spacecraft. 

During the random vibration a bolt unscrews itself and falls from the 

spacecraft. It is the upper-most (+z) bolt connecting the +y lateral panel 

to the +y+x small shear panel. This is not serious. 

Bolt a bit loose in the –y lateral panel / baseplate interface, below the +x– 

y shear panel. This is not serious. 

During run 7Y the fundamental resonance shifted lower and ended up 

halfway out of the notch. This is good in that the levels are more 

acceptable for SSTL, but bad in case some damage was done to the 

spacecraft. In the subsequent low-level sine sweep the fundamental rose 

slightly again, which is a good sign. 

During the vibration a bolt fell out, second from the top (+z), on the +y 

lateral panel interface to the +x+y shear panel. This is just below the one 

that fell out last time (problem 253), and can be explained by the heavy 

vibrations of the +x+y shear panel and the +Y T-Pod. This is not serious 

but it should be well glued and torqued for the flight. 

Bolt loose at the top of the +y lateral panel interface to the +x+y shear 

panel. This is the one that fell out last time and can be explained for the 

same reasons as problem 254. This is not serious but it should be well 

glued and torqued for the flight. 

Bolt loose from the +y lateral to baseplate interface, just below the –x+y 


The –x side of the +y T-Pod is scoured where it repeatedly impacted on 

the upper skin of the top-plate during the vibration. The +x side is also 

scoured at the +y corner for the same reason, but not at the –y corner, 

where the top bracket is holding it. 

13 th April 

2005 

13 th April 

2005 

13 th April 

2005 

14 th April 

2005 

14 th April 

2005 

15 th April 

2005 

15 th April 

2005 

15 th April 

2005 

15 th April 

2005 

15 th April 

2005 

15 th April 

2005 

15 th April 

2005 

574






258 

259 

260 

261 

262 

263 

264 

265 

266 

267 

268 

269 

270 

271 

The fundamental mode of the spacecraft on the x axis has divided into a 

38Hz resonance and a 32Hz resonance. 

The upper bolt on the mid-height bracket on the +x side of the 0x+y 

compartment is very loose (about to fall out). 

The bolt from the +x+y top bracket to the +x+y shear panel is very loose. 

The +y T-Pod current consumption wavers around during charging and 

the battery will not rise about 10.55V. 

The test connector for the pyro does not match the current harness. After 

a phonecall to MAGIC_Renato it is confirmed that the MAGIC pinout 

does not match the PROP pinout. Use of a DVM to buzz through the 

connector confirms this. 

ACDS_Lars reports that +y sun-sensor is not working. Suspect harness 

problem. 

Sampling of pressures and temperatures do not seem to be working 

properly during thruster firing, but this could just be operator error. 

A faster sampling of the acceleration data is requested, and then the 

results are queried. Instead of getting housekeeping data an item is added 

into the alarm stack from OBC, MID 0x88, 17 bytes long “CAM: Chunk 

Error”. After this it is necessary to power cycle both the OBC and the 

CAM before the CAM responds again. 

All downloaded pictures are black. 

OBC_Karl reports that the fast sampling command responds with an MID 

that looks like a picture, which is why OBC gets confused. This could 

only be changed with an update of the OBC software. 

It is impossible to upload parameters to the camera - the OBC software, 

rather stupidly, re-writes them all to zero. This means that it is 

impossible to take pictures. The only sensible way to fix this is to change 

the OBC software. 

The second incarnation of the groundstation laptop dies. It appears to be 

a RAM malfunction. It refuses to boot back up and the relevant CSV file 

from the DTMF decoding is probably lost. 

The electronics of the +y T-Pod have been damaged to the extent that the 

relay and heater block do not fire. This is almost certainly related to 

problem 261. 

One of the screw locks will not unscrew and simply rotates the mating 

half. The only way to get it off is to cut the head of the bolt. 

15 th April 

2005 

15 th April 

2005 

15 th April 

2005 

16 th April 

2005 

16 th April 

2005 

16 th April 

2005 

16 th April 

2005 

19 th April 

2005 

19 th April 

2005 

19 th April 

2005 

19 th April 

2005 

20 th April 

2005 

21 st April 

2005 

22 nd April 

2005 

575






272 

273 

274 

275 

276 

277 

278 

279 

280 

281 

282 

One of the heads of one of the screw locks on the OBC snapped off 

during torquing. It is not possible to remove the threaded shaft from the 

socket. Instead the connector is simply glued in place (also bolted at the 

other side). 

During the checkout the CAM occasionally stops responding. This later 

transpires to be because the OBC was being rebooted by EPS – overnight 

it booted a total of forty-seven times. 

Last night the OBC randomly reset the boot counter again. This could be 

due a momentary power drop during boot-up due to the connection of the 

external power suppy while the battery was still connected as well (due to 

lacking EPS documentation). 

Due to the accidental (lacking documentation!) dual-powering of the 

spacecraft the batteries have run down low overnight and are found 

locked (7V) in the morning. Unlocking them by charging reveals them at 

about 18.5V, which is dangerously low. However, they seem to charge 

back up without issue. 

Two washers are found in the +y0x compartment, and 2 washers and a 

long M5 bolt are found under the spacecraft. 

It seems form the above that several of the top-bracket bolts have 

loosened and fallen during the transport vibrations. This is very 

surprising and unexplained – since the bolts were torqued at integration. 

However, this was a long time ago and some settling would have taken 

place since then. 

The +x cap does not fit since the foam antenna was made before the 

attenuation foam was added and is not quite accurate. 

The Moxaets team requests that their solar panel protector is replaced 

while we perform our RBF work, since it is a large horizontal plane right 

underneath us. This presents us with a problem though, since one of the 

bolts on their protector puts their tooling perilously close to the +y sun 

sensor. 

The spacecraft is very difficult to turn and it is supported manually by 

SSY_Jörg and ESA_Neil on one top corner during most of the process. 

This is dangerous and a better system should be found for the return to 

the vertical. 

The fork lift impacted on the -x panel. It is not clear whether any damage 

has been done since viewing access is restricted. It is probably ok, but 

more luck than judgement. 

There are lots of strange mass spectrometer readings, many with 

inexplicably high AMUs. Possibly magnetorquers, tank coating or 

attenuation caps. The pressure rises to around 6.6x10-4 mill bar. 

22 nd April 

2005 

24 th April 

2005 

25 th April 

2005 

25 th April 

2005 

4 th May 

2005 

4 th May 

2005 

4 th May 

2005 

6 th May 

2005 

9 th May 

2005 

9 th May 

2005 

10 th May 

2005 

576






283 

284 

285 

286 

287 

288 

289 

290 

291 

292 

293 

294 

295 

296 

297 

Trying command S-Band up results in a jump in signal but no carrier, 

however, the current consumption continues as if it is up. Temperature in 

S-Band unit jumps up by a few degrees. 

The spacecraft stops responding, ESA_Neil debugs down the OBC port, 

it seems fine and responds to internal commands to transmit. This 

implies that the UHF TNC has locked up in the receive direction. 

UHF does not seem to be transmitting constantly... it is incrementing by 

about 10 packets at a time, and only up about 50%. 

The e-box tab snaps off due to work hardening. 

The T-Pod firing fails. 

A lose bolt is found in the NCube-II pod. It is from the e-box mounting 

on the back left. 

We only measure 240mW being delivered to the +x S-Band patch 

antenna, it should be 750mW. 

One rubber standoff has come loose from the +x T-Pod 

The standoffs do not quite hold the spacecraft tight enough (the push 

plate can still be moved by hand through the access port). 

It appears that the low-pressure tubing leaks, since the reading of pressure 

transducer three was 0 bar when the spacecraft was in the thermal 

vacuum chamber, and 1 bar when back at ambient pressure. It is not 

clear how serious this problem is. 

The spacecraft starts up at 1619, only 65 minutes instead of 74. 

Test 14: FAILED!! All public telecommands work even if the first 

parameter FF is used. This is potentially very dangerous as a command to 

downlink a full picture on UHF would effectively lock us out of the 

spacecraft for 30 minutes. OBC_Karl advise on reasons and possible 

solutions. 

S-band carrier only up about 50% of the time, with a period of 4 or 5 

seconds... (Or, if put carrier up full-time then modulation is only about 

50% of the time -> OBC problem) 

After 11 minutes the ACK is still coming down S-BAND, when it should 

have switched back to UHF by now. 

Between boots 18 and 19 there are more that 24 hours, but no 

telecommands were sent, so it should have rebooted after 24 hours. This 

11 th May 

2005 

13 th May 

2005 

13 th May 

2005 

16 th May 

2005 

16 th May 

2005 

18 th May 

2005 

20 th May 

2005 

24 th May 

2005 

24 th May 

2005 

26 th May 

2005 

30 th May 

2005 

1 st June 

2005 

1 st June 

2005 

1 st June 

2005 

5 th June 

2005 

577






298 

299 

300 

301 

302 

303 

304 

305 

306 

implies that the OBC failed the end of test 15. 

ACDS operation mode reports incorrectly. 

OPER discover that the marker on the GET_FILE command increments 

one packet too far on each downlink. This means that single packets will 

be missed between increments of GET_FILE, unless a FILE_SEEK of –1 

(FFFF) is used in between. 

There is no response from OBC on the debugging line when it is powered 

up with the new proms in place. 

There is no response from OBC on the debugging line when it is powered 

up with the old proms in place. 

Some thrusters are activated the wrong way 

Due to some HPR effects the mid pressure is rising over time. 

The OBC reboots after 1 day and 56 minutes. However, this appears to 

be an internal problem and was not a shutdown from EPS. Also, since 

this problem is certainly rare (first time we have seen it), and happened 

after 24 hours, it poses no significant danger to the mission. 

Upon shutdown the OBC is powered off before it writes the HK file, and 

before it marked the AL file as “alarm”. 

Upon a 24 timeout the OBC did not write the HK and AL stacks to flash, 

did not power cycle the UHF at all, and did not reset the PCU. On the 

debugging line it simply reported “Incremented ACDS mode counter to 

(7)” and then reset. 

6 th June 

2005 

12 th June 

2005 

13 th June 

2005 

13 th June 

2005 

15 th June 

2005 

16 th June 

2005 

16 th June 

2005 

16 th June 

2005 

17 th June 

2005 

578






6 Modifications 

1 STRU 

Small shear panel +x+y DMM side pockets 

enlarged slightly to make mounting strip sit 

flush 

ESA_Neil 

3rd October 

2004 

2 STRU 

Small shear panel -x+y DMM side pockets 

enlarged by 7mm 

ESA_Neil 

3rd October 

2004 

3 PIN 

A few solder joints touched up, tape 

removed, cable ties added, glue added 

ESA_Jason 

5th October 

2004 

4 STRU 

Glue and paper stuck to honeycomb skins 

on panels –x-y and –x+y removed with heat 

and scalpel 

ESA_Neil 

6th October 

2004 

5 STRU 

Cut away the excess glue just around the 

injection and air holes on all the –x side 

inserts of the DMM baseplate, the –x side 

inserts of the DMM top-plate and the –x 

side inserts of the FM baseplate. 

ESA_Neil 

7th October 

2004 

6 STRU 

Remove kapton and semi-cured excess glue 

from all potted centre inserts of FM +y 

panel, re-kapton and top-up as necessary 

ESA_Neil 

SYS_Joerg 

7th October 

2004 

7 STRU 

Remove all kapton and semi-cured excess 

glue from –x sides of DMM baseplate, 

DMM top-plate and FM top-plate 

ESA_Neil 

SYS_Joerg 

8th October 

2004 

8 STRU 

Inserts are carefully removed from the –x 

side of the +_y shear panel while the glue is 

still liquid. Inserts are cleaned . Small 

strips of tissue to remove some glue from 

the side pockets by wicking. Kapton to 

protect the panel from excess overflowing 

glue. Re-applies mounting strip. Gluing is 

redone and appears unproblematic. 

ESA_Neil 

SYS_Joerg 

9th October 

2004 

9 STRU 

The integration manual is wrong on page 

21; there was a typo in the CATIA model. 

Subsequently the wrong types of bolts were 

ordered, we must order some more. 

Luckily we already have M5x16mm bolts 

for other purposes in the spacecraft, so we 

use them to mount the base brackets. 

ESA_Neil 

SYS_Joerg 

9th October 

2004 

10 STRU 

We have to use an M4 bolt instead of an 

M5 on bracket #12 (page 30 of integration 

ESA_Neil 

SYS_Joerg 

9th October 

2004 

579






manual), since insert is M4. This needs to 

be checked by STRU to see whether or not 

it is acceptable. 

11 STRU 

We carry on without fixing bracket #13 to 

the –x0y panel because the holes do not 

align. We ask STRU to check where this 

error comes from. A possible solution 

would be to modify the bracket slightly by 

drilling a hole in the appropriate place. We 

ask STRU is this will be acceptable (the 

bolt would then be able to slide sideways a 

little in the –y direction). 

ESA_Neil 

SYS_Joerg 

9th October 

2004 

12 STRU 

Uppermost (+z) pocket on –x side of –y 

panel too high by 7mm. As with the 

precedent set on the engineering model 

correction is made by enlargement of the 

pocket in the right direction. 

ESA_Neil 

ESA_Marie 

14th 

October 

2004 

13 STRU 

For ease of manufacture it was decided not 

to have injection holes in new side inserts, 

since they are not really used in the potting 

procedure anyway. This will give a slightly 

reduced mechanical strength, which, 

although negligible, should be reflected by 

placing these inserts in positions where 

they bear the smallest loads. (On the –y 

panel.) 

ESA_Neil 

15th 

October 

2004 

14 STRU 

ESA_Neil ground the excess glue out of the 

holes in the ASAP inserts on the FM 

baseplate. 

ESA_Neil 

18th 

October 

2004 

15 STRU 

After extensive discussion ESA_Neil, 

SYS_Joerg, ESA_Marie and STRU_Melro 

decide to upgrade the –y, -x0y, and +x0y 

panels of the Engineering Model to flight 

hardware, and downgrade the –y, -x0y, and 

+x0y panels of the Flight Model to 

engineering hardware. 

ESA_Neil 

ESA_Marie 

SYS_Joerg 

STRU_Melro 

20th 

October 

2004 

16 STRU 

ESA_Neil arranges with ESA_Marcel to 

have the erroneous holes slotted in the 

number #13 aluminium brackets, and to 

have four of the M6x60 bolts cut-down to 

55mm in length so as not to interfere with 

PROP tubing. 

ESA_Neil 

21st 

October 

2004 

580






17 

OBC & 

TCS 

OBC to TCS pin-out: 1 Thermistor 1 

signal, 2 Thermistor 2 signal, 3 Thermistor 

3 signal, 4 Empty, 5 Ground, 6 Thermistor 

1 signal (redundant), 7 Thermistor 2 signal 

(redundant), 8 Thermistor 3 signal 

(redundant), 9 Power supply 3.3V 

OBC_Karl 

21st 

October 

2004 

18 OBC 

OBC_Karl disconnects power supply 

between boards (by de-soldering a hook-up 

wire) and powers the computer directly in 

order to check that the power supply is 

stable (no difference) 

OBC_Karl 

23rd 

October 

2004 

19 OBC 

OBC_Karl adds an appropriate resistor to 

terminate the CAN bus and re-solders the 

hook-up wire between the two boards. The 

system then functions normally. 

OBC_Karl 

23rd 

October 

2004 

20 OBC 

OBC_Karl needs to make the computer 

again, this will take approximately a day. 

The components will arrive on Monday 

night with ACDS_Lars and CAM_Morten. 

OBC_Karl 

24th 

October 

2004 

21 

STRU 

& TCS 

Excess kapton glue is removed from 

masking by “scrubbing” with a vinyl glove 

(to which it adheres better than to the 

panels), or, in extreme cases, with a scalpel. 

ESA_Neil 

ESA_Marie 

25th 

October 

2004 

22 TCS 

ESA_Neil and ESA_Marie carefully cut 

extra thermal paint away from the panel 

with a scalpel, until the area is large enough 

to encompass the titanium ring. The result 

is not too pretty, but the damage is only 

aesthetic. 

ESA_Neil 

ESA_Marie 

25 th 

October 

2004 

23 OBC 

OBC_Karl and CAM_Morten cut a hidden 

stray track on the FM OBC circuit board. 

Afterwards it works fine. 

OBC_Karl 

CAM_Morten 

26 th 

October 

2004 

24 MAGIC 

25 STRU 

OBC_Karl de-solders the CAN termination 

on the EM MAGIC. (Two pairs of parallel 

120 ohm resistors and a capacitor.) 

ESA_Neil enlarged the holes on the corner 

profiles to accommodate the rivet nuts. 

OBC_Karl 

ESA_Neil 

27 th 

October 

2004 

28 th 

October 

2004 

26 STRU 

ESA_Neil removes an M6 nut from under 

the EM lifting frame, taking care not to let 

the top bold fall into the satellite. This bolt 

ESA_Neil 

28 th 

October 

2004 

581






27 ACDS 

28 

OBC & 

TCS 

29 PROP 

30 PROP 

31 PROP 

32 PROP 

33 PROP 

34 MAGIC 

is needed to solve problem 46. 

ACDS_Lars and OBC_Karl re-configure 

the pinouts between OBC and ACDS 

It is decided to use the spare OBC analogue 

inputs as two extra thermistors to extend 

the TCS system. 

Developed method of tensioning of tank 

clamps when using shorter bolts and tank 

under pressure. 

An M4 bolt is used to mount the thrusters 

clusters to the thrusters inserts instead of an 

M5. The other bolts nearby are M5, so the 

system as a whole should be strong enough. 

The addition of washers to space the 

mounts, and then a large weight (7kg metal 

plate) placed on top of the tubing to hold 

them down while they dry overnight 

Due to the lack of one aluminium “patch”, 

a spare of the smallest mid-height bracket 

is used in the –x-y+z corner of the central 

compartment. This will add a little mass, 

but provide much better strength. 

ESA_Neil and PROP_Hanno temporarily 

remove the PMS box and bend the tubing 

into place. The PMS box is then replaced. 

During this operation the top-plate is 

temporarily replaced, in order to help make 

sure that the other panels do not move. 

We decide to leave the Magic box 

hardware as it is but make the software 

more robust by having a minimalist (and 

therefore small target) random number 

generator, which then directs to any of a 

large number (however many fit on the 

chip) of copies of the main code. Any 

particular copy of the main code will then 

report it’s own checksum to the OBC, 

which will toggle the unit if it doesn’t get 

the right response. 

OBC_Karl & 

ACDS_Lars 

OBC_Karl 

ESA_Neil 

PROP_Hanno, 

PROP_Nils 

PROP_Mattias 

PROP_Sascha 

PROP_Mattias 

PROP_Sascha 

PROP_Mattias 

ESA_Neil 

SYS_Joerg 

ESA_Neil 

PROP_Hanno 

ESA_Neil 

OBC_Karl 

MAGIC_Ren 

35 SYS We decide to cover all flash chips on-board ESA_Neil 5 th 

29 th 

October 

2004 

31 st 

October 

2004 

31 st 

October 

2004 

31 st 

October 

2004 

31 st 

October 

2004 

1 st 

November 

2004 

1 st 

November 

2004 

5 th 

November 

2004 

582






with small protective layers of tantalum to 

err on the side of caution. 

November 

2004 

36 EPS 

Decide to split harness from EPS RBF 

connector such that the battery lines come 

direct from the battery box and back, while 

the others come from the PCU. 

ESA_Neil 

EPS_Fulvio 

7 th 

November 

2004 

37 SBAND 

A line is added connecting pin 5 of the 

RS232 line to the ground arriving from the 

PCU. 

AMS_Jason 

8 th 

November 

2004 

38 SBAND 

A ground is taken from after the DC-DC 

converter (to keep galvanic isolation) to the 

audio link. 

AMS_Jason 

8 th 

November 

39 OBC 

40 SBAND 

OBC_Karl replaces the dead line driver 

from the UHF and EPS ports. 

A low pass filter is added to remove this 

problem of the third harmonic. 

OBC_Karl 

AMS_Sam 

AMS_David 

8 th 

November 

2004 

8 th 

November 

2004 

41 OBC 

OBC_Karl and AMS_Jason fix the AX25 

protocol implementation on the OBC. 

OBC_Karl 

AMS_Jason 

9 th 

November 

2004 

42 SBAND 

An inductance, diode, load resistor are 

tested as a choke on the S-BAND power 

input and work fine to reduce the current 

inrush. S-BAND can now be powered by 

EPS directly, although the flight 

implementation of the modification remains 

pending. 

AMS_David 

AMS_Jason 

AMS_Howard 

EPS_Fulvio 

ESA_Neil 

9 th 

November 

2004 

43 PROP 

The PROP team progress through to step 

5.1 inclusive from the list given as a 

required modification above. 

PROP team 

10 th 

November 

2004 

44 PROP 

45 EPS 

PROP team temporarily glue some metal 

blocks to the top of the tank to assist with 

applying the correct torque. 

EPS_Fulvio replaces the PIC with the only 

remaining spare. 

PROP team 

EPS_Fulvio 

11 th 

November 

2004 

11 th 

November 

2004 

46 PROP 

The high-pressure tubing mount is glued 

back down with “UHU” two-part resin (fast 

drying). 

PROP_Hanno 

16 th 

November 

2004 

583






47 

PROP 

STRU 

It is decided that washers “as big as 

possible” (quote STRU_Melro) should be 

added to the bolts on the tank mounting. 

Therefore two small stainless steel plates 

are made, each of which act as a washer for 

one pair of the tank mounting bolts. While 

the tank mountings are tightened again the 

honeycomb can be heard being pulled back 

into place. This seems like a very strong 

solution and is judged adequate to continue 

the testing. 

PROP_Hanno 

PROP_Sascha 

ESA_Neil 

16 th 

November 

2004 

48 PROP 

Larger M5 washers are added beneath the 

M4 ones on the PMS mounting points so as 

to distribute the loads properly. These bolts 

are torqued and glued again (UHU). 

PROP_Hanno 

ESA_Neil 

16 th 

November 

2004 

49 ACDS 

ACDS_Lars modifies EM lateral panels 

and coil-driver brackets to fit the coil driver 

PCB properly and to accommodate the sunsensors 

and associated harness, and 

ACDS_Lars replaces connector on 

magnetometer. 

ACDS_Lars 

24 th 

November 

2004 

50 ACDS 

ACDS_Lars drills new holes into the 

magnetometer mounting plate. 

ACDS_Lars 

24 th 

November 

2004 

51 OBC 

OBC_Karl once again replaces the dead 

line-driver. It is the same one as last time, 

therefore resulting in the third soldering of 

that part of the PCB. This is not really 

acceptable, but we have little choice. 

OBC_Karl 

24 th 

November 

2004 

52 ACDS 

53 MAGIC 

ACDS_Lars drills holes in EM lateral panel 

(+x) for the coil-holding-tie-wraps. 

ESA_Jason takes off the old magic 

processor. 

ACDS_Lars 

ESA_Jason 

24 th 

November 

2004 

25 th 

November 

2004 

54 MAGIC 

55 

EPS & 

OBC 

OBC_Karl solders the new MAGIC 

processor to the board. 

UHF_ON and UHF_OFF commands 

removed and replaced with UHF_CYCLE 

(much more sensible). This command was 

OBC_Karl 

OBC_Karl 

EPS_Stefano 

ESA_Neil 

25 th 

November 

2004 

25 th 

November 

2004 

584






then added to the “satellite has not heard 

form the ground for 24 hours” routine 

ensuring that the UHF will be power cycled 

if necessary. 

SYS_Joerg 

56 SYS 

ESA_Neil drills / mills new holes into the 

side protectors to fit the lower right hand 

corners properly. (And the lower left hand 

corner of the –y protector.) 

ESA_Neil 

2 nd 

December 

2004 

57 STRU 

58 ACDS 

ACDS_Lars and ESA_Neil drill holes in 

FM secondary structure to support ACDS 

integration. 

OBC_Karl adds modification to ACDS 

coil-driver to correct analogue controller 

problem. 

ACDS_Lars 

ESA_Neil 

OBC_Karl 

7 th 

December 

2004 

7 th 

December 

2004 

59 CAM 

60 OBC 

ESA_Jason mills sides of FM CAM down 

by 1mm each. 

ESA_Jason mills hole for crystal out of FM 

OBC box, and re-taps thread in OBC box. 

ESA_Jason 

ESA_Jason 

10 th 

December 

2004 

10 th 

December 

2004 

61 OBC 

OBC_Karl replaces two resistors on main 

board and IF board. 

OBC_Karl 

10 th 

December 

2004 

62 STRU 

ESA_Neil uses a Dremmel to “square” the 

corners of the holes in the EM and then the 

FM top-plates 

ESA_Neil 

13 th 

December 

2004 

63 STRU 

COMM_Damian uses a hand file to 

“square” the corners of the +z LGA 

mounting plate. 

COMM_Damian 

13 th 

December 

2004 

64 COMM 

ESA_Neil modifies an EM LGA with a 

Dremmel to fit the +z LGA mounting plate. 

Results are ok but COMM_Damian is not 

happy with the process. 

ESA_Neil 

13 th 

December 

2004 

65 STRU 

ESA_Neil uses a Dremmel to make a small 

cut-away in the FM +z LGA mounting 

plate so that it fits around the washer on the 

top-plate. 

ESA_Neil 

13 th 

December 

2004 

66 EPS 

EPS_Fulvio and EPS_Tommy take the 

battery apart and then re-integrate it, 

replacing plastic wiring with PTFE and 

EPS_Fulvio 

EPS_Tommy 

14 th 

December 

2004 

585






plastic sheathing with kapton tape. They 

take many pictures throughout the process 

and refer to these to ensure that the cells are 

correctly connected. 

67 EPS 

EPS_Fulvio and EPS_Tommy line the 

battery box with kapton and recharge the 

batteries, and touch up some of the 

soldering on the BCR. This probably fixes 

the problem, but only testing in full context 

(using the structure) will tell for sure. 

EPS_Fulvio 

EPS_Tommy 

17 th 

December 

2004 

68 UHF 

ESA_Neil modifies the EM UHF antenna 

by cutting off both ends with a hacksaw, 

grinding the edges smooth and grinding the 

corners rounded to fit into the insert on the 

top-plate. 

ESA_Neil 

17 th 

January 

2005 

69 UHF 

ESA_Jason redoes pin-out of PTT in FM 

UHF. 

ESA_Jason 

26 th 

January 

2005 

70 S-Band 

ESA_Neil and ESA_Marcel mill the lid of 

the S-Band FM enclosure so that it fits 

around the power connector properly. 

ESA_Neil performs a fit check on the 

hardware and reinstalls the TNC into the 

lid. 

ESA_Neil 

ESA_Marcel 

27 th 

January 

2005 

71 S-Band 

AMS_Sam opens the S-band FM box and 

de-solders the programming switch link. 

AMS_Sam 

15 th 

February 

2005 

72 OBC 

OBC_Karl uploads the latest version of the 

OBC software, including the S-Band / 

Debug hack to get around the data 

bottleneck problem. 

OBC_Karl 

15 th 

February 

2005 

73 S-Band 

74 S-Band 

AMS_Sam re-solders a link across the 

programming switch on the S-Band TNC. 

ESA_Jason removes the old MGA 82563 

and replaces it with one of the new ones. 

This does not solve the problem. 

AMS_Sam 

ESA_Jason 

15 th 

February 

2005 

16 th 

February 

2005 

75 Ncube2 

NCUBE_Bjørn cuts the end of the removebefore-flight 

pin off to shorten it. 

NCUBE_Bjørn 

24 th 

February 

2005 

586






76 EPS 

EPS_Tommy cuts the legs off of the 

original mosfet that is limiting the s-band 

current and temporarily solders a 

replacement mosfet with a lower resistance 

for the purposes of testing. 

EPS_Tommy 

2 nd March 

2005 

77 EPS 

EPS_Tommy removes the temporary 

mosfet by de-soldering the original pins, 

prepares a new mosfet and glues it into 

place on top of the old one. (There is no 

possibility to remove the old one from the 

board.) 

EPS_Tommy 

2 nd March 

2005 

78 EPS 

EPS_Tommy replaces a potentially faulty 

operational amplifier on the EPS FM PSU. 

EPS_Tommy 

3 rd March 

2005 

79 EPS 

EPS_Tommy removes the diodes and 

solders wire shorts across their previous 

locations instead. 

EPS_Tommy 

3 rd March 

2005 

80 EPS 

ESA_Neil and EPS_Tommy upload version 

6.8 of the PDU software to the PDU PIC. 

EPS_Tommy 

ESA_Neil 

3 rd March 

2005 

81 EPS 

EPS_Tommy modifies the timers on the 

flight board. 

EPS_Tommy 

4 th March 

2005 

82 EPS 

EPS_Tommy adds a safety resistor to the 

timer reset wire, and then crimps it to pin 1 

of the solar panel connector, from which it 

can run to the FPP. 

EPS_Tommy 

5 th March 

2005 

83 EPS 

EPS_Tommy removes the linear voltage 

regulator powering the PDU from the PSU. 

EPS_Tommy 

5 th March 

2005 

84 EPS 

EPS_Tommy removes the PIC from the 

PDU 

EPS_Tommy 

5 th March 

2005 

85 EPS 

EPS_Tommy re-solders the linear voltage 

regulator on the PSU for power feeding of 

the PDU. 

EPS_Tommy 

7 th March 

2005 

86 HARN 

Rather than modify the hard-to-access UHF 

box, the harness for the PTT link is 

redesigned. The new pinout uses ground 

on pins 1 and 6 and positive 5V on pins 2 

and 7. The new cable is symmetrical, and 

it is clearly labelled. 

ESA_Neil 

7 th March 

2005 

87 OBC 

OBC_Karl fixes problem 176 by creating a 

‘flash storage’ thread which loads the 

picture into flash memory but sleeps 

OBC_Karl 

8 th March 

2005 

587






occasionally to allow the other threads to 

deal with the pings from the PCU. 

88 OBC 

OBC_Karl removes the flash erasure 

commands. (They are replaced now by the 

filing system.) 

OBC_Karl 

9 th March 

2005 

89 OBC 

Karl takes TX_delay set out of the OBC 

boot-up sequence. 

OBC_Karl 

9 th March 

2005 

90 OBC 

OBC_Karl makes “whole picture” 

commands unfriendly. 

OBC_Karl 

9 th March 

2005 

91 ACDS 

It is decided that for successive periods of 

24 hours after the launch without reception 

of a valid TC the following configurations 

of ACDS will be used respectively: 

normal, off, off, off, off, off, off, inverted, 

off forever. 

ESA_Neil 

ACDS_Lars 

13 th March 

2005 

92 OBC 

OBC_Karl implements a byte stored in 

flash memory that controls what 

configuration will be used for ACDS after 

boot-up. A new telecommand to reset the 

byte is added, and the ACDS reading of the 

byte is implemented. 

OBC_Karl 

13 th March 

2005 

93 SYS 

The teflon stand-offs for the side protectors 

are shortened by 3mm to make space for 

M4 nuts that will hold the bolts in place. 

ESA_Neil 

14 th March 

2005 

94 UHF 

The EM dummy antenna is dismantled, the 

edges of the base ground inwards by 

0.5mm, the holes drilled out to 4.2mm and 

reassembled. 

ESA_Neil 

14 th March 

2005 

95 SYS 

ESA_Neil shortens the side protector 

spacers by 3mm and adds an M4 nut to 

each of the bolts to hold the spacers on. 

ESA_Neil 

15 th March 

2005 

96 EPS 

EPS_Fulvio adds delays between packets 

as well as between bytes, as some packets 

are being sent very close to each other. 

Version 7.1. 

EPS_Fulvio 

15 th March 

2005 

97 EPS 

EPS_Fulvio extends the inter-packet and 

inter-byte delay to 1ms instead of 1 us. 

Version 7.2 

EPS_Fulvio 

15 th March 

2005 

98 EPS 

EPS_Fulvio changes the period on the 

nominal mode beacon data to 59 seconds 

EPS_Fulvio 

15 th March 

2005 

588






instead of 60 so that it doesn’t collide with 

the telemetry packet so often. Version 7.3 

99 OBC 

OBC_Karl corrects the OBC software to 

rectify the “switch case fall through error” 

that was causing the EPS nominal mode 

beacon data to register as OBC data. 

OBC_Karl 

15 th March 

2005 

100 OBC 

OBC_Karl tweaks the UHF downlink 

dripping bucket. 

OBC_Karl 

15 th March 

2005 

101 OBC 

OBC_Karl tweaks the S-Band downlink 

dripping bucket. 

OBC_Karl 

15 th March 

2005 

102 STRU 

ESA_Neil and ESA_Andre remove a cutout 

from the 1 FM substrate panel and drill 

two new holes for the coil driver bolts 

(7mm down and 1mm left) 

ESA_Neil 

ESA_Andre 

16 th March 

2005 

103 STRU 

ESA_Neil files off the erroneous damaged 

corner of the solar panel substrate 7 FM 

ESA_Neil 

16 th March 

2005 

104 STRU 

ESA_Neil relocates the –y sun sensor holes 

in the lateral panel, using solar panel 

substrate 3 as a guide. 

ESA_Neil 

17 th March 

2005 

105 STRU 

ESA_Neil adds the –y sun sensor harness 

hole to solar panel substrate 3 and to the –y 

lateral panel, using the position of original 

harness hole relative to the original fixation 

holes in the lateral panel as a guide. 

ESA_Neil 

17 th March 

2005 

106 STRU 

ESA_Neil drills 8 holes into the –y lateral 

panel for the new interfaces with the first 

and third solar panel substrates. 

ESA_Neil 

17 th March 

2005 

107 STRU 

ESA_Neil drills 2 holes into the +x lateral 

panel for the new interfaces with the fifth 

solar panel substrate. 

ESA_Neil 

17 th March 

2005 

108 STRU 

ESA_André enlarges the cut-out in the top 

of the +y panel for the heating block and 

harness of the +y T-Pod. 

ESA_André 

17 th March 

2005 

109 STRU 

ESA_Neil drills 10 holes into the +y lateral 

panel for the new interfaces with the sixth 

and eighth solar panel substrate. 

ESA_Neil 

17 th March 

2005 

110 STRU 

ESA_Neil drills 2 holes into the -x lateral 

panel for the new interfaces with the ninth 

solar panel substrate. 

ESA_Neil 

17 th March 

2005 

589






111 SOLAR 

112 STRU 

113 SOLAR 

114 SOLAR 

115 SOLAR 

116 MAGIC 

117 MAGIC 

118 MAGIC 

119 STRU 

120 EPS 

It is decided that problem 199 can be 

alleviated by adding the final connectors to 

the side of the last cell instead of the end. 

Strings 4 and 8 are adapted to reflect this, 

and strings 5a, 6a and 9a are planned to be 

implemented with this change already in 

place. Similar modifications possibly 

remain to strings 3b, 3c and 6b. 

ESA_Neil adds a harness hole for the +y 

sun sensor to solar panel substrate 6 and the 

+y lateral panel. It is positioned relative to 

the fixation points in the same pattern as for 

the –y sun sensor. 

ESA_Bas solders an extra cell onto 3b to 

make it into a two cell string – hereafter 

dubbed 3d. 

ESA_Bas replaces 6b with a cell that has 

connectors on the side. 

For ease of configuration single cell strings 

3b and 3c are to be relocated to the second 

solar panel substrate as a two cell string, 

3d. 

ESA_Neil uses a milling bit on the 

Dremmel to enlarge the holes in the 

baseplate of MAGIC such that M4 nuts will 

fit. 

ESA_Neil uses the Dremmel to enlarge the 

hole in the MAGIC top plate 

ESA_Neil uses the Dremmel to slightly 

enlarge the four corner mounting holes of 

the MAGIC box. 

ESA_Neil grinds the end of the bolt flush 

with the shear panel, using the Dremmel. 

The various layers of kapton tape are 

removed from the interior walls of the 

battery box and replaced with a single layer 

of wider kapton tape. The battery box is 

then just possible to close, although two or 

three bolts still will not go in. 

ESA_Neil 

SOL team 

ESA_Neil 

ESA_Bas 

ESA_Bas 

SOL 

ESA_Neil 

ESA_Neil 

ESA_Neil 

ESA_Neil 

ESA_Neil 

17 th March 

2005 

18 th March 

2005 

18 th March 

2005 

18 th March 

2005 

18 th March 

2005 

18 th March 

2005 

18 th March 

2005 

18 th March 

2005 

18 th March 

2005 

20 th March 

2005 

121 STRU The appropriate bolt holes in the lateral ESA_Neil 20 th March 

590






panels are milled out slightly in the 

appropriate directions to fix problems 205, 

207 and 208. 

2005 

122 MAGIC 

MAGIC is turned around to make room for 

the CAM connectors next to the relay (the 

original 3D model of MAGIC was incorrect 

in terms of the top-plate orientation). 

ESA_Neil 

22 nd March 

2005 

123 EPS 

EPS_Fulvio changes the software so that 

the ticket is always “42” (0x2A). This is 

done since it is impossible for the OBC to 

get into a loop where it responds to pings 

automatically without “listening” to them, 

so the content of the ticket is irrelevant. 

EPS_Fulvio 

23 rd March 

2005 

124 EPS 

ESA_Neil slightly widens all the holes on 

the side panels of the battery box, and 

slightly shortens three of the bolts. It is 

then possible to close it properly. 

ESA_Neil 

25 th March 

2005 

125 UHF 

We add a right-angled SMA saver to the 

RF port. 

ESA_Neil 

AMS_Howard 

27 th March 

2005 

126 STRU 

For the PIN ESA_Neil uses M4x12mm 

bolts and two M4 washers instead of 

M4x10mm bolts with no washers. 

ESA_Neil 

27 th March 

2005 

127 HARN 

ESA_Neil cuts and dremmels the harness 

clamps by the PCU flat. 

ESA_Neil 

1 st April 

2005 

128 EPS 

ESA_Neil slots the holes in the PCU L- 

profile with a Dremmel 

ESA_Neil 

2 nd April 

2005 

129 ACDS 

ESA_Bas extends the +y sun sensor cables. 

ESA_Bas 

5 th April 

2005 

130 STRU 

In order to provide better access to the 

charging port on Xi-V, the +y lateral panel 

has a L shape dremmeled from it. The 

edges are tidied up with a sander and then 

protected with aluminium tape. 

ESA_Neil, 

STRU_Antonio 

10 th April 

2005 

131 T-Pod 

The +x t-pod is modified thusly: Teflon 

block roughed up with a screwdriver, cable 

ties cut to 80% width by hand with scalpel, 

two ties inserted and block glued to the 

back of the pod with AY138. Spring 

reattached and rails cleaned, cable ties 

tightened and UWE-1 (minus one kill 

ESA_Neil, 

STRU_Antonio, 

PROP_Sascha 

13 th April 

2005 

591






switch) is re-loaded. 

132 T-Pod 

133 T-Pod 

134 AIV 

135 MAGIC 

136 ACDS 

137 OBC 

138 T-Pod 

139 T-Pod 

The -x t-pod is modified thusly: Teflon 

block roughed up with a screwdriver, cable 

ties cut to 80% width by hand with scalpel, 

only one tie wrap inserted since one hole in 

the block was too narrow. Block glued to 

the back of pod with AY138. Spring 

reattached and rails cleaned, cable tie 

tightened and mass dummy (pretending to 

be ncube-2) is re-loaded. 

The +y t-pod is modified thusly: ay138 

applied to bases of centres of sides of 

Teflon block to try and secure it in position 

on the back of the pod. Rails and push 

plate cleaned for easy sliding. 

In order to protect the spacecraft it is 

decided to notch the input loads to the 

flight levels across the first resonance 

during the lateral random vibrations. This 

is not in accordance with the recent SSTL 

“regulations”, but that will be negotiated 

later. 

PROP_Sascha modifies harness on magic 

box pyro connector to correct problem 262. 

Moves pin 1 to 1, 2 to 2, 3 to 9, 4 to 10, 5 

to 5, 6 to 6, 7 to 13, 8 to 12. 

Pin 12 in the 25-pin ACDS connector is 

moved to pin 21, and the old pin 21 is 

removed 

The OBC is disintegrated and the PROMS 

replaced. 

ESA_Neil follows the instructions from 

CANX on how to replace the damaged E- 

box on the +y T-Pod 

ESA_Neil replaces the heater block of the 

+y T-Pod with one of the spare ones, 

making sure to cover the top of it in kapton 

tape to insulate the heater elements from 

the pod itself. 

ESA_Neil, 

STRU_Antonio, 

PROP_Sascha 

ESA_Neil, 

STRU_Antonio, 

PROP_Sascha 

STRU_Antonio, 

ESA_Wolfgang, 

ESA_Neil 

PROP_Sascha 

ESA_Neil 

ESA_Neil 

ESA_Neil 

ESA_Neil 

13 th April 

2005 

13 th April 

2005 

15 th April 

2005 

16 th April 

2005 

21st April 

2005 

22nd April 

2005 

16th May 

2005 

16th May 

2005 

140 T-Pod We add several layers of Kapton tape on ESA_Neil 18th May 

592






top of the rubber stand-offs to tighten the fit 

of Xi-V in the +y T-Pod. 

2005 

141 T-Pod 

ESA_Neil glues the loose rubber standoff 

back into position on the inside of the lid of 

the +x T-Pod. 

ESA_Neil 

24th May 

2005 

142 T-Pod 

ESA_Neil uses small squares of 5 layers of 

kapton tape to thicken the rubber standoffs 

in the +x T-Pod. 

ESA_Neil 

25th May 

2005 

143 OBC 

ESA_Neil replaces the OBC flight proms, 

which necessitates disintegrating the box 

and cycling the prom sockets. 

ESA_Neil 

13th June 

2005 

144 OBC 

A scalpel is used to scrape the 

discolouration off of the potentially 

problematic pin and to gently bend it back 

into alignment with the rest of the socket. 

ESA_Neil 

13th June 

2005 

PENDING MODIFICATION: The skin of the baseplate should be strengthened at the point 

where the Polyot deployment switch interfaces to SSETI Express. 

PENDING MODIFICATION: We need to add handles to the +x and –y side protectors for 

ease of manoeuvring. 

PENDING MODIFICATION: A way has to be found of encapsulating the bolts as they are 

applied to the –y panel, so that there is no way of dropping them. 

PENDING MODIFICATION: A way should be found of attaching the screw-driver to a 

wrist strap so that it cannot be dropped. 

593






7 Connector usage 

CAM Power Saver removed 17 September 2004 

CAM Data Saver removed 17 September 2004 

PIN All Savers removed and replaced once 6 October 2004 

MAGIC All All connectors cycled once as 7 November 2004 

MAGIC_Renato forgot to put savers 

SBAND Power Cycled once accidentally 24 November 2004 

PIN 25-pin Cycled once due to faulty saver 10 January 2005 

CAM Both Cycled once for dismantling box 31 January 2005 

SBAND Data Cycled once accidentally 15 February 2005 

OBC Debug / S- Cycled to place a saver in, since it is 15 February 2005 

Band now a flight socket 

MAGIC All on top Saver cycled 18 th March 2005 

MGM All one Saver cycled 18 th March 2005 

ACDS Floating 25 Saver cycled 20 th March 2005 

PIN UHF Saver cycled 25 th March 2005 

UHF SMA Cycled to change to right-angled 27 th March 2005 

saver (which is then cycled several 

times and should be replaced) 

UHF RS232 Saver removed and flight cable 27 th March 2005 

attached 

UHF Audio Saver removed and flight cable 27 th March 2005 

attached 

S-BAND +Z patch Saver added 27 th March 2005 

PIN UHF Saver removed and flight cable 29 th March 2005 

applied 

PIN OBC Saver removed and flight cable 29 th March 2005 

applied 

PIN S-BAND Saver removed and flight cable 29 th March 2005 

applied 

PIN CAM Saver removed and flight cable 29 th March 2005 

applied 

PIN PIC Saver removed and flight cable 29 th March 2005 

applied 

PIN ACDS Saver removed and flight cable 29 th March 2005 

applied 

-X POD Power Saver removed and flight cable 29 th March 2005 

applied 

+X POD Power Saver removed and flight cable 

applied 


594






+Y POD Power Saver removed and flight cable 29 th March 2005 

applied 

UHF RS232 Flight cable is changed for a longer 29 th March 2005 

and crimped one 

UHF Adapter Antenna end is cycled to place flight 29 th March 2005 

cable 

UHF PTT Saver removed and flight cable 29 th March 2005 

applied 

PIC CAN Saver removed and flight cable 29 th March 2005 

applied 

CAM RS232 Saver removed and flight cable 29 th March 2005 

applied 

UHF Antenna Cycled thrice 9th April 2005 

UHF Power Cycled twice 9th April 2005 

UHF RS232 Cycled once 9th April 2005 

UHF S-Band Cycled once 9th April 2005 

OBC ALL Cycled once from prom replacement 22 nd April 2005 

OBC Debug Cycled once for test 22 nd April 2005 

595






8 Arrivals and Departures 

0 OBC team arrive with large bag of thousands of separate pieces 02/08/2004 

1 EPS_Fulvio with software prototype board 12/08/2004 

2 Test UHF modems 16/08/2004 

DEP 1 EPS_Fulvio goes home 30/08/2004 

DEP 2 OBC team go home with OBC EM and test modems 30/08/2004 

3 

EPS_Tommasso arrives in ESTEC with EPS PSU components and empty 

PCBs 

08/09/2004 

4 PIN arrives in ESTEC 15/09/2004 

5 CAM arrives in ESTEC 17/09/2004 

6 ACDS MAGNETOMETER arrives in ESTEC 17/09/2004 

7 ACDS MAGNETS arrives in ESTEC 17/09/2004 

8 Magic EM arrives in ESTEC 17/09/2004 

9 Both titanium rings 23/09/2004 

10 Inserts arrive from STRU_Melro 23/09/2004 

11 New EPS boards arrive in ESTEC 27/09/2004 

12 Aluminium brackets 27/09/2004 

13 XI-V_Yuya arrives with the engineering model of Xi-V Cubesat 28/09/2004 

14 DMM shear panels +y and –y arrive in ESTEC 30/09/2004 

15 

ESA_Neil and ESA_Marie go and pick up Araldite AY103 and Hardener 

HY991 from Viba 

30/09/2004 

DEP 3 EPS_Tommasso takes EM hardware with him to Naples 02/10/2004 

16 ACDS COILS arrive in ESTEC 04/10/2004 

17 Flight structure (except plate 16) arrives in ESTEC 06/10/2004 

18 Lifting frame and panel protectors from Vienna 07/10/2004 

19 The EM e-box arrives in ESTEC from Canada 08/10/2004 

20 Two potential transport containers arrived in ESTEC 08/10/2004 

21 Two integration pillars arrive in ESTEC 08/10/2004 

22 The engineering models of the separation rings (two) arrive from SSTL 12/10/2004 

23 The spare transport boxes arrive in the office 12/10/2004 

24 Corner profiles from Westend 13/10/2004 

25 Lateral panels from Westend 13/10/2004 

26 All remaining inserts and the passive magnet housing from Westend 13/10 13/10/2004 

27 PROBA box 14/10/2004 

28 Final panels from ISF 14/10/2004 

29 The remaining nuts 15/10/2004 

30 The remaining side inserts from ESA_Marcel 15/10/2004 

31 The magnetorquer coil clamps from Westend BV 15/10/2004 

32 The adapter plate from ESA_Marcel 15/10/2004 

33 The FM Magic box arrives in ESTEC 19/10/2004 

596






34 UWE-1 mass dummy 20/10/2004 

35 New paint arrives from Map (base 21/10/2004 

36 Otto switches arrive from Westend BV 22/10/2004 

37 

38 

EPS_Fulvio arrives to complete the EPS subsystem with FM PDU boards and 

components 

ESA_Neil and ESA_Marcel go to Valkenburg airport to pick up the painted FM 

panels from the Royal Dutch Marines 

24/10/2004 

25/10/2004 

39 Second lifting frame and a dummy antenna arrive from INFRA_Lars 25/10 25/10/2004 

40 OBC FM replacement components 26/10/2004 

41 CAM spare components 26/10/2004 

42 Remaining ACDS components 26/10/2004 

43 ESA_Neil picks up the pyro connectors donated by ESA_Neil_Cable 27/10/2004 

44 The FM PDU components arrive for EPS_Fulvio 27/10/2004 

DEP 5 CAM_Morten takes all the CAM hardware (apart from the box) back to Aalborg 29/10/2004 

45 Remaining OBC components arrive 29/10/2004 

46 UHF EM arrives in ESTEC. Visual inspection, somewhat worrying. 01/11/2004 

48 

The FM S-BAND unit and a large amount of communications 

equipment arrives with the AMSAT UK team. 

08/11/2004 

49 The FM UHF box and antenna arrives with UHF_Holger 08/11/2004 

50 

51 

DEP 6 

DEP 7 

Low Gain Antennas, FS S-BAND enclosure, Microwave Cables and 

Antenna Caps arrive with COMMPL. 

AMS_Howard completes setup of a simple test groundstation that can 

be used for functional testing. 

AMS_Sam, AMS_David and AMS_Jason leave ESTEC taking with 

them the FM S-BAND unit, the FS S-BAND enclosure, ready-crimped 

lengths of 24 gauge PTFE wire, flight solder and rosin flux. 

EPS_Fulvio finally gets to go home, taking with him the FM BCR and 

the EM PDU. He will finish them in Naples and send them back. 

08/11/2004 

10/11/2004 

10/11/2004 

13/11/2004 

52 The flight structure and propulsion system arrive back in ESTEC. 18/11/2004 

53 The FM CAM returns, rebuilt, to ESTEC with CAM_Morten. 22/11/2004 

54 The current MCC and GND laptops and software arrive. 22/11/2004 

55 

S-BAND comes back to ESTEC with AMS_Graham. Visual inspection 

reveals some gluing issues, but nothing too serious. 

22/11/2004 

56 The T-Pods arrive with CANX_Fred. 06/12/2004 

57 The new Utility processor arrives with OBC_Karl. 06/12/2004 

58 LGA back-shields and reinforced LGAs arrive with COMM_Damian. 13/12/2004 

DEP 8 Damaged NCube-2 mass dummy is sent back to Norway for evaluation. 19/01/2005 

59 The flight battery arrives and looks fine. 27/01/2005 

60 The external BCR, and a spare, arrives, and they look fine 14/02/2005 

61 A spare laptop for the launch campaign is generously donated by 14/02/2005 

597






DEP 9 

AMS_Graham 

The S-Band FM unit returns to the UK with AMS_Graham and 

AMS_Sam. 

16/02/2005 

62 NCube-2 flight model arrives with Bjørn Pedersen. 24/02/2005 

DEP 

10 

The NCube-2 flight model leaves with Bjørn Pedersen. 24/02/2005 

63 The S-Band FM arrives (again) with AMS_David and AMS_Sam 01/03/2005 

64 

65 

66 

DS_Roland and DS_Fons generously donate some RTV and special 

solder to SSETI for the purposes of panel laydown. 

DS_Roland and DS_Fons generously donate 64 Tecstar solar cell interconnectors. 

The SSETI Express Solar Cell Rear Solder Jig arrives. A simple fit 

check proves that it will be perfect for the job. 

10/03/2005 

10/03/2005 

10/03/2005 

67 The first of the solar panel substrates arrives. 10/03/2005 

68 The last batch of solar cells arrive. 10/03/2005 

69 

Dutchspace generously donate three pots of RTV, two vials of hardener 

and a bottle of primer. 

70 Three solar panel substrates arrive, 1 FM, 1 EM and 7 FM. 

71 The rest of the solar panel substrates arrive 

DEP 

11 

The fit-check hardware leaves for Omsk 

72 The OBC flight proms arrive. 

73 The EPS flight PIC arrives. 

74 

NCube-II arrives with NCUBE_Åge. This time it has a metal protection 

box that is depressing the kill switches. The antennas have not 

deployed, and the RBF pin appears to be smaller, as required. 

75 The flight PIC (ver 1.3) arrives. 

76 The S-Band patch antenna bolts arrive. 

77 The new shunt resistor arrives. 

78 The L-profile for the PCU is manufactured. 

79 The fourth attempt at the flight EPS PIC arrives. 

80 

Filter / capacitor connectors and savers arrive from AMS_Howard for 

the purposes of troubleshooting the RF problems. 

14/03/2005 

16 th March 

2005 

17 th March 

2005 

18 th March 

2005 

23 rd March 

2005 

23 rd March 

2005 

24 th March 

2005 

29 th March 

2005 

29 th March 

2005 

29 th March 

2005 

30 th March 

2005 

31 st March 

2005 

8 th April 

2005 

81 The thermal foil arrives at ESTEC. 9 th May 

598






82 The thermal-vacuum chamber interface is manufactured. 

83 The flight battery arrives at ESTEC. 

84 The replacement e-box arrives at ESTEC 

85 

The Xi-V team arrive with the flight model and all associated ground 

support equipment. 

86 The UWE-1 flight model arrives with UWE_Yohko and UWE_Radu. 

87 The new OBC flight proms arrive. 

2005 

9 th May 

2005 

16 th May 

2005 

16 th May 

2005 

18 th May 

2005 

24 th May 

2005 

13 th June 

2005 

599






Glue batch usage 

No Time Date Application Vac Comments 

S1 1300 30/09/2004 DMM centre inserts in all 

small shear panels apart from 

No First attempt with 

AY103 

–x0y 

S2 1900 30/09/2004 DMM centre inserts in –x0y Yes 240g 

and base-plate 

S3 1530 01/10/2004 DMM centre inserts in topplate 

Yes 

and 16 

S4 1900 01/10/2004 DMM centre inserts in panel Yes 

15 

S5 1800 03/10/2004 DMM all side inserts in all 

small shears, +x side of +y 

panel, +x side of –y panel, 

+x side of top panel 

S6 1200 04/10/2004 DMM side inserts in +x side 

of base plate, plus S5 top-ups 

S7 1100 05/10/2004 DMM side inserts in –y side 

of base-plate, -y side of topplate, 

-x side of +y plate and 

–x side of –y plate 

S8 1200 06/10/2004 DMM side inserts in +y side 

of base-plate and +y side of 

top-plate. Top-ups in –x+y, - 

x-y, +y, -y 

S9 11:00 07/10/2004 Side inserts in –x side of 

DMM base-plate, –x side of 

DMM top-plate, -x side of 

FM base-plate. Centre 

inserts in majority of FM +y 

plate (see entry) 

S10 23:00 07/10/2004 Centre inserts of the FM 

+x+y panel, centre inserts of 

the FM -x+y panel, 

remaining centre inserts of 

the FM +y panel, topping up 

the S9 inserts 

No 

No 

Yes 

Yes 

Yes 

Yes 

S11 21:00 08/10/2004 Side inserts in +x side of FM Yes 

S12 21:00 08/10/2004 

baseplate, +x side of FM +y 

(15), sides of FM +x+y and – 

x+y shear plates 

Yes 

Trying without 

vacuum, result: more 

topping-up required 

Better – procedure 

adopted 

Extended vacuum 

exposure tested 

Glue MUCH better 

(less air), could be: 

New tins, new pots, 

cleaner pots, pouring 

Tried to mimic S10 

but not so good – 

reason unknown 

600






S13 10:00 09/10/2004 Side inserts in –x side of FM 

+y (15), +y side of FM 

baseplate, +y side of FM topplate, 

FM +x-y and FM –x-y 

S14 10:00 09/10/2004 The side insets in –y side of 

the FM baseplate and the –x 

side of the FM top-plate. 

S15 16:00 10/10/2004 The side inserts in the –y 

side of the FM top-plate and 

the centre inserts (not 

thrusters) of the FM 

baseplate 

S16 19:30 11/10/2004 The side inserts of the +x 

side of the FM top-plate 

S17 15:00 13/10/2004 The bottom layer of the 10 

inserts for the thruster 

mounting in the FM 

baseplate, and all the central 

inserts of the FM top-plate 

S18 17:30 13/10/2004 All the centre inserts of the 

FM –x0y, +x0y, +x-y, +x+y 

panels 

S19 14:00 15/10/2004 All the centre inserts of the 

FM –y panel, and the first 

halves of the thruster inserts 

in the FM baseplate 

S20 16:00 16/10/2004 The side inserts on the –x 

side of the FM –y panel 

S21 15:00 17/10/2004 The side inserts on the +x 

side of the FM –y panel and 

the remaining volume in the 

thruster inserts on the FM 

baseplate 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Yes 

Possibly better if do 

not pour 

601






9 Milestones 

1 

EPS prototype and OBC EM successfully progress through 

operational modes, boot-up and watchdog sequences 

2 First Cubesat arrives in ESTEC (Xi-V EM) 


28 th September 

2004 

3 Successful strip and mask test on all DMM panels 8 th October 2004 

4 Completion of EM primary structure 9 th October 2004 

5 FM OBC boots up successfully for the first time 22 nd October 2004 

6 Stage 1 of the flight structure leaves ESTEC for Stuttgart 25 th October 2004 

7 FM OBC downloads a picture from FM CAM 27 th October 2004 

8 EM structure completed (primary and secondary) 27 th October 2004 

9 Successful functional integration of FM OBC and FM EPS 31 st October 2004 

10 The point of no return. 1 st November 2004 

11 The FM Primary Structure is completed. 2 nd November 2004 

12 

13 

14 

15 

The flight model Pressure Management System passes its 

vibration and leakage test. 

Functional integration of OBC, EPS and MAGIC at the stage 

where the PROP payload is supported by the platform. 

The SSETI Express safe mode and nominal mode beacons are 

received via a radio for the first time. 

Spacecraft hardware is commanded via an RF link for the first 

time. 





16 SSETI Express transponds audio for the first time. 8 th November 2004 

17 

18 

The nominal mode beacon is received and decoded for the first 

time by the test ground station. 

SSETI Express receives, acknowledges and responds to its first 

RF telecommand. 

19 SSETI Express passes its pressurised vibration tests. 

20 Two-way link established between MCC, GND, UHF and OBC 

21 

The propulsion thrusters are fired via OBC and MAGIC for the 

first time and work perfectly. 



16 th November 

2004 


2004 


2004 

602






22 The FM PCU fires the T-Pods for the first time 6 th December 2004 

23 Successful S-BAND data downlink for the first time 

24 

The satellite is powered-up using the battery box for the first 

time. 

16 th December 

2004 

17 th December 

2004 

25 The S-Band sub-system is declared flight-ready 3 rd March 2005 

26 The OBC and EPS software are declared FLIGHT READY 15 th March 2005 

27 The integration is completed 11 th April 2005 

28 The spacecraft is ready for vibration testing. 12 th April 2005 

29 The vibration testing is completed 15 th April 2005 

30 The fit-check is completed successfully 7 th May 2005 

31 The spacecraft is ready for thermal vacuum tests 9 th May 2005 

603






10 Lessons Learned 

This section lists the lessons learned during the project at various levels. Where appropriate 

specific dates are given. 

GENERAL 

9. Take the time at the beginning of the design phase define which components and 

standards are acceptable for the project. Otherwise a varying standard of systems will be 

delivered with some undesirable components, then time will be lost deciding which to replace, 

and actually replacing them. 

Project management and system engineering should be at LEAST two full-time people. 

There simply are not enough hours in the day to do both, and both tasks have suffered 

significantly in Express due to the lack of manpower. This could easily lead to a sub-standard 

team or a sub-standard satellite (respectively). 

Projects like this should only be started with appropriate budget and mandate already in 

place. The SSETI Express project has lost a lot of time fighting to get the budget and 

mandate it needs. 

Projects like this should only start once all the tasks are accounted for by teams that 

have proven themselves to be capable and who state, in writing, that they will see the 

project through to the end. There are several examples in the Express project of “gaps” in 

the tasks that had to be filled by management or third parties, a few examples of teams who 

could not / did not follow the project through to completion, and a couple of examples of 

teams who simply were not capable enough for the task (still too far down the learning curve 

to be participating in a fast moving project with a launch date). 

20. Any modifications to the hardware (or software) should be “back implemented” into 

the design immediately. Otherwise future actions derived from the design may not fit with 

the current reality. (16/03/05) 

INTEGRATION 

14. Before a major system level test EVERY SINGLE COMPONENT should be 

checked and rechecked on ALL criteria. (16/11/04) 

There should be an Assembly, Integration and Verification team that is present at the 

integration site for the whole integration and testing period, reports to the project manager, 

and includes at least one of the system engineering team. 

604






As far as possible log everything you do, no matter how small. There will be a lot of it, 

you will not remember everything, and something you forget might be something you need to 

know. Going back through the log to find out WHY it happened might solve a problem. This 

is the reason that this logbook was started. 

Assemble and test entire engineering model before producing flight hardware. It is 

ONLY by building it that you will ever realise 90% of the problems with the design. If we 

had done this with Express then all of these lessons learned would have been incorporated 

into the flight hardware and the quality would have been significantly higher. Also – the 

same people that integrate the engineering models of everything should manufacture the flight 

models – there are always dozens of small and “insignificant” techniques and procedures that 

are developed and remembered. The manufacture and integration flight hardware should 

NOT involve a learning curve at all, that should already have been completed. 

Define procedures for everything. When working on the engineering model the procedures 

and techniques developed should be well documented, preferably with visual aids. This not 

only allows the working individual to crystallise and evolve a well-defined and 

comprehensive description of the task at hand (which is then to be followed for the flight 

model integration), but it also allows other people to perform the same task with a much 

smaller learning curve. 

5. All flight bolts should be cleaned VERY thoroughly, using dry tissues, then IPA, then an 

ultrasound bath, then acetone. This is to ensure smooth application and proper relevance of 

torque measurements. (16 th October 2004) 

7. Don’t do anything to flight model that you didn’t do to the EM. A fine example was in 

Express when OBC_Karl added a “handy” reset wire, which fried the whole FM computer 

when it accidentally touched a positive power supply. It could have already caused damage 

simply by acting as an antenna soldered directly onto a chip. (24 th October 2004.) 

Take the context of your models into account, and make them identical. It is common to 

mistakenly make an EM slightly different from the FM because they work in different 

contexts and you need a “hack” to get the EM working. This can cause problems with 

troubleshooting later on since you are blinded by the apparent, but false, similarity of your 

EM and FM models. A fine example is a day lost on the OBC troubleshooting (23 rd October 

2004) because the EM CAN bus was terminated locally, whereas the FM CAN bus is to be 

terminated in the Magic box. 

11. Before reaching any “point of no return” make ABSOLUTELY sure that everything 

“before” the point is complete, and demonstrate this by testing IN CONTEXT 

(10/11/2004). For example: even though the tank had passed a pressure test previously, the 

same tests should have been done again after integration to the structure, before the point of 

no return. 

605






16. Make sure that all required equipment is specified and provided before arrival and work 

of the teams, otherwise time will be wasted looking for it. (24 th November 2004) 

19. Close all valves properly before leak testing. (Obvious really – but we didn’t do it! 26 th 

November 2004) 

21. Never close a box without checking that it is properly documented. (16th April 2005) 

22. Explicitly check all interfaces before committing hardware. (16th April 2005) 

ELECTRICAL 

1. Before start of hardware manufacture hold small courses or workshops on PCB 

design, soldering, component selection, “flight worthy” rules, flight hardware 

methodology (23/09/2004). This is to make sure that everyone is clear from the start on what 

is required of them, and should lead to a fairly consistent set of hardware arriving without the 

same mistakes being repeated throughout. Much less time will therefore be spent ‘finishing’ 

hardware that is already meant to be finished. Express lost about 6 weeks in the schedule to 

this problem. 

Define redundancies early and make sure that it goes right to the PCBs, not just the cabling. 

Having redundancy in the cables is nice, but if you only have one solder joint on the PCB 

then the system is still quite unsafe. PCBs should have holes for each redundant cable. 

Strain relief. Cables should pass through a hole on the PCB and then be soldered to a second 

hole on the other side – this provides significant strain relief. 

1.1 Define minimum PCB print standard early (23/09/2004). This would ensure that all 

the PCBs turning up would be at least the required standard. In Express we lost time and 

606






money by having to replace sub-standard PCBs. Possibly centralise PCB manufacture for 

system-level standard and bulk discount 

10. Anything can act as an antenna (09/11/2004). Keep all cables connected at both ends 

and keep all at low level and at a reasonable distance from sensitive electronics. 

12. DOUBLE check ALL the connections to the power supply EVERY TIME before you 

turn it on (10/11/2004). 

13. At the end of any particular session remove all the test cables to the external power 

supplies, forcing them to be set up again the next time instead of relying on it being set up 

correctly already (maybe you were doing something slightly different in the last session and 

you forgot the details). (10/11/2004) 

17. Make sure you get the grounding scheme implemented properly during the testing! On 

SSETI Express we forgot that the plastic table doesn’t work the same way as the metal 

spacecraft and lost a lot of time and effort because of it. (25 th November 2004) 

18: Handle microcontrollers carefully. One in the PCU broke within about 5 minutes of 

getting it out the wrapper. (25 th November 2004) 

Ground all the components via the structure. This will reduce harness, allow radios to 

have a local grounding plane, and help to reduce the risk of ground loops. 

The solar panel definition should be done early and is part of the job of the EPS team 

themselves, including considerations of power-point-tracking, string length, efficiency, 

harness losses and diode losses. The thermal team (for efficiency degradation), the mission 

analysis team (for eclipse times), the attitude control team (for incident angle profiles), the 

simulation team (to bring all this together) and the structure team (for the physical 

configuration) will also need to be involved in the iterations. 

LESSON LEARNED 23: Side protectors should have handles 

LESSON LEARNED 24: All RBF and ABF should be BIG and preferably there should be 

no ABF items 

LESSON LEARNED 25: Side protectors should not need replacement bolts 

LESSON LEARNED 26: Submit drawings to launch authority after each update 

LESSON LEARNED 27: Request early-on the orientation for latest RBF activity 

LESSON LEARNED 28: RBF tooling should be small and minimised 

LESSON LEARNED 29: Set up document control system with launch authority 

607






STRU 

2. All inserts should be symmetrical in terms of mounting points (09/10/04). This is 

because there were several occasions when side inserts were nearly glued into panels the 

wrong way around, there are several hundred inserts and it is difficult and time-consuming to 

interpret the axes correctly that many times in a row. Symmetry would remove the problem 

entirely. 

3. All structural items should be symmetrical as far as possible (09/10/04). In the case 

where they are not symmetrical they should be obviously or explicitly not symmetrical. 

If possible “side insert” potting should be avoided. This process is very difficult, messy 

and time-consuming, partly because it is hard to position the inserts within their tolerances, 

and partly because side inserts have to dry before you can rotate to other sides (slow). Better 

would be entire side strips, or better yet would be side ‘cups’ that attach via centre inserts. 

Insert mounting masks should be provided. There are several types of insert in SSETI 

Express (thruster, top and lateral sides, base sides) which are extremely difficult to position 

correctly since they are entirely ‘floating’ in glue. For such configurations a mounting mask 

should be provided so that all inserts can be secured to it in their correct positions and angles 

while the glue hardens. 

4. Use kapton tape that is wide enough for the job. When insert potting no joins between 

different strips of kapton tape should be accessible to the glue being used. Otherwise glue 

will run down the joints on top of the panel skin. 15 th October 2004. 

6. Measures should be taken to ensure that mounting plates cannot be glued to the 

panels during insert potting. A better procedure should be developed for this process to 

ensure that the initial glue depth is insufficient to result in flow of glue on the surface of the 

skin. 17 th October 2004. 

8. Do not use “flight” kapton tape if you are going to want to take it off again for any 

reason, as the glue is often too adhesive to so easily. (25 th October 2004) 

15. Get every team to double check all mechanical interfaces explicitly with the relevant 

teams. (24 th November 2004) 

Configuration is a SYSTEM LEVEL task, and should be treated as such by capable people 

with a good overview of the entire system. 

All structural elements, including individual subsystem boxes, should be defined by the 

structures team. Otherwise you will end up with a wide variety of styles and configurations 

608






which will each require different bolts, torques and handling. Also, electrical engineers are 

often not particularly competent mechanical engineers, so the boxes may not be up to scratch. 

Represent your definition drawings as they are to be seen in real life - it avoids 

misunderstandings. 

DATA 

A data exchange definition, both hardware and protocol, should be decided very early 

on and should include some form of flow control (preferably hardware handshaking). This 

includes the connector type, pin-out, redundancy, protocol, voltage levels, baud rates, parity, 

start bits and stop bits. Recommendations are: no grounding in cables (to avoid loops, just 

use the chassis), hardware handshaking, double redundancy. 

If you use RS232, use the standard 5V logic values in order to avoid unnecessary failure 

points in line drivers, and then manufacture a set of debugging connectors which step up to 

inverted 12V and add a ground line. 

609






11 APPENDIX I – OBC report 

610






12 APPENDIX II – Lessons learned from INFRA / UHF 

UHF 

1) Protocols have to be 

a. defined early, 

b. versions of these protocols are IMPORTANT 

c. test dummies(like dummy modems) are important early 

2) Hardware-connections has to be defined early and fixed – 

a. Voltages 

b. Connector (size, male, female, wiring) 

c. changes have to be distributed with feedback !! 

3) Every hardware even the boxes have to be build at least two times 

a. Ask for some dummy PCB’s before machining (makes fit test much easier) 

b. Machining the box needs time (at least a week) 

c. Machine can break down and needs repair time, so think of other possibilities 

d. Machine can destroy the material so always get some spare material too. 

e. Always think of thermal behaviour when designing the box 

f. Don’t believe the calculated mass estimations of a program exactly add 20% 

safety. 

g. Normally 1mm wall thickness is not possible 3mm is OK (vibrations of the wall is 

to high when machining) 

h. 2mm corner radius is normally not possible- solution drill 5mm hole at the exact 

corners then you even can have an exact rectangular PCB. 

4) If we get hardware for free 

a. ask for the specifications 

b. ask for measuring instruments to test this hardware 

c. ask for more (second spare) 

5) do hard test early 

a. ask for the possibilities of vacuum chambers in your university (Sputtering 

machines can also be used) 

b. heating: we’ve seen, that the running PCB at 50°C burned one component but 

not at room temp 

c. To test connectors under high current and voltage in a vacuum chamber, we had 

to modify it – so ask if modifications are allowed, ask for the possibility of an 

assurance (for you, the accompanying students and the hardware, sometimes 

everything is included in the laboratory insurance sometimes NOT) 

6) ALWAYS do 

a. a fix pricing, 

b. a fix timing 

c. fix the number of produced things, before beginning to deal with professional 

manufacturers. 

d. Inform yourself about ALL needed parameters, data and formats before going to 

produce something. 

e. Always print BIG plans for the manufacturer 

f. Check and recheck all dimensions on these plans 

611






INFRA 

7) Check the servers and it’s services by hand as often as you can (sometimes things 

happen ;) ) 

8) Always do highest security first before implementing services 

9) Always do highest bandwidth and throughput !!! 

10) Always use virus detection at a most updated level 

11) Keep firewall as closed as possible 

12) Use UPS (uninterruptible power supply) on all your servers 

13) Do a backup as much as possible (especially before modifying things) 

14) Have distributed fallback servers (if network goes down) 

15) Always inform yourself as much as possible about the services/programs you install 

before installing them 

16) Do not promise times you cannot hold (even simple programs or modifications) 

17) Always test services for at least a week before making them available for all 

18) Keep the infra team as small as possible to keep track of the services and things going 

on. 

19) Do not use cheap hardware (this dies to early) 

20) Have at least 3 spare ventilators somewhere ;) 

21) Put your servers in a cold (air-conditioned) room, which can be locked 

22) All the mobile phone numbers of the infra team have to be given to the Managers for 

urgent requests 

23) Always have a backup person (trustable and knowing what he/she is doing) at the server 

when going into holidays, or workshops … 

24) Update and patch the services often (do backup before) 

25) Distribute independent services to different servers. 

26) Keep user restrictions to public services at a moderate level. 

27) Web pages have to be viewed by every type of browser (Explorer, Netscape, Opera, … at 

any versions – test this !!!) 

28) Keep track of the uploaded files (especially gif, jpg, avi … exe, com, …) if these files are 

allowed or illegal. 

29) Keep track of “automated” web services. 

30) Use a trustable Internet provider. 

31) Administer the domain name by Management. 

32) Don’t use “framed” forwarding of web pages. 

33) Don’t use html-frames at all … 

34) Web-page concept has to be easy and understandable by everybody ! 

35) Updating of information’s of teams and members has to be done by management, doing 

this by the team coordinators themselves doesn’t work!!!! 

36) Think of running the server for years not only for “months” 

37) Infra team has to be available for longer then just some months … (not normal students) 

38) Don’t change web page structure often (or ever) 

39) Keep track of the pwd’s and logins, iterate them every ½ year 

40) Do documentation of every service and modification 

41) Do a todo list (priority first) 

42) Don’t install services if they are not needed by a majority of users. 

43) Inform yourself about the legal issues of the servers, services … 

44) Be careful when getting pwd, information … 

612






613

SSETI Express SSETI Express Integration Logbook - Space.aau.dk

Create successful ePaper yourself

Delete template?

Save as template?