For funsies, I bought a couple of microSD cards that were obviously "fake flash" (where the card misrepresents itself as a higher capacity than it actually is) and ran them through some tests to see what capacity they really were and how reliable they were. I bought this card as my control.
My rig for doing these tests was an Orange Pi 4 running Armbian (a Ubuntu-based distro tuned towards single board computers). I used AltraMayor's f3 suite of tools, along with ncw's stressdisk tool (both of which you can find on Github).
Step 1 was to see how the card represented itself to the operating system. You can see from the dmesg logs that it presented itself as a 256 GB card to the OS. (It technically presented itself as having 500,695,040 512-byte blocks -- which would come out to 256,355,860,480 bytes. Most computer programs measure a gigabyte as 1024^3, whereas most products are advertised using 1000^3. So by the latter definition, this card is 256 GB; but by the former definition, it's only 239 GB. It's annoying, but not unexpected.)
Step 2 was use to use f3probe to see what the card's actual capacity is. (This program works by writing pseudorandom data to different places on the card, then reading it back to see if it matches what's expected. Genuine flash will always return the data that's expected, whereas fake flash will return the expected data up to a point -- after which it'll start returning bad data.) The good news here is that it only took f3probe 29 seconds to tell me that yes, this card is actually 256 GB. (For comparison, it took it over three hours to tell me that my two "fake flash" cards -- which were advertised as 1024 GB -- were actually only about 8 GB in size.)
Step 3 was to use f3write and f3read. (These programs fill up all of the available space on the card with pseudorandom data -- as opposed to f3probe, which just writes data to strategic points on the card -- and the read it back to see if it matches what's expected, to check that the flash cells are working properly.) Again, nothing but good news here -- this card passed with flying colors.
Step 4 was to run a stress test on the card using stressdisk. (This program fills up the card with pseudorandom data, then reads it back a couple of times to check to see if it matches what's expected. It then writes new data to the card, reads it back a couple of times, and repeats this process over and over again for 24 hours -- watching to see if any errors crop up.) Once again, this card passed with flying colors -- not a single error during the entire 24 hours. (One of my "fake flash" cards started showing errors about halfway through the test; the other one started showing errors before the 2-hour mark.)
The last thing to look at is read/write speeds. Both f3read/f3write and stressdisk will tell us what read/write speeds it was able to attain -- so all we have to do is look at the output from those programs. This card carries the UHS-3 mark -- which means that it must be able to sustain read/write speeds of at least 30 MB/sec -- and this card blew that out of the water. Both tools said that it was able to attain over 67 MB/sec write speed and over 87 MB/sec read speed. Heck, that's enough to qualify for the next highest speed rating, V60 -- which requires read/write speeds of at least 60 MB/sec. (Samsung advertises that this card can get over 90 MB/sec write speed and 100 MB/sec read speed -- I'm going to chalk up the difference to limitations in my rig.)
So if you're looking for a good quality SD card...in my opinion, this is it!
| Special Feature | water_proof, magnetic_proof, x_ray_proof, temperature_proof, shock_proof |
|---|---|
| Read Speed | 95 |
| Write Speed | 20 |
| Hardware Interface | microSDHC |
| Secure Digital Association Speed Class | Class 10 |
| Product Dimensions | 0.59"L x 0.43"W |
| RAM | 32 GB |
