I recently bought a MikroTik L009UiGS-RM router. I tried using two brand new SanDisk Ultra Flair 64GB USB 3.0 flash drives with it (one after the other).
Both USB sticks get extremely hot (over 60°C) when plugged into the router and end up failing/broken after a short time. When I later connect them to a PC to reformat them to ext4, the computer no longer recognizes the drives at all.
The drives come pre-formatted as FAT32/NTFS. I suspect this is caused by a bug or power-delivery issue in the MikroTik L009’s USB port that is damaging the flash drives.
Is the there anyone experienced before? Or any idea?
It Is perfectly possible that your specific Mikrotik device has a defect, but AFAIK there haven't been similar reports.
I would check with a USB voltmeter/amperometer which voltage out you have, and, if It Is the standard 5V (with no load) I would try connecting briefly another (different make/model) stick and check that also amps are correct ( It depends on the specific stick model, but should be less than 1 A) .
Then, depending on the results I would RMA (if possible) either the L009 or the sticks.
it is also perfectly possible that what you thought was a genuine SanDisk Ultra Flair 64GB USB 3.0 flash drive was in fact a counterfeit.
If you didn't test with at least GRC's quick test with validrive or a thourgh test with h2testw, you don't even know that the drive isn't reporting more capacity than it really has.
Getting hot is common for the counterfeit drives. Some have quite convincing packaging, China prides itself in making clones.
This is for other people that want to test their drives while they can still return them.
h2testw is the "gold standard". It can fill the entire device with psuedo random data (that it can easily reproduce and verify when it reads it back). Here's a youtube video describing its use
This is the latest place I have downloaded from. After downloading, test with virustotal to verify that it is the real deal before running.
GRC (Steve Gibson's Gibson Research Corporation) has a freeware utility that checks only a portion of the drive (from random locations). It also gives an indication of the quality of the drives. MicroCenter branded drives aren't fake, but they are really slow switching from read to write mode, and veridrive takes a long time to run when testing them. I had some pny drives that were also really slow, but still faster than the MicroCenter drives. https://www.grc.com/validrive.htm and youtube video describing how it is used (example with a counterfeit drive that is overstating its capacity) Validrive: Verify Your Micro SD Card!
Just in case, there is a fast checker, designed as first thing to use to check a drive (it will only detect fake partitioning/allocations, but it will do it very quickly), written by the Author of RMPrepUSB, that also put together a comprehensive page about the issue of fake storage devices, with links to various suitable programs, including Validrive and H2testw: https://rmprepusb.com/tutorials/007-all-about-fake-sd-cards-and-usb-flash-drives/
Loosely, there are two kinds of "fakes":
small capacity devices "cleverly" partitioned/formatted to show as much bigger at OS/filesystem level
full capacity devices using "bad" memory
Type #1 can be quickly detected, #2 need a slower "full" read/write cycle.
So, voltage is perfectly in line with standard 5V. (typical is +/-5%, so 4.75-5.25V)
Allow me to doubt that you are actually measuring Amperes with probes connected in parallel to + and -, though (they should be in series with the load).
A charging phone @5V should normally use between 0.5 and 1 A, anyway, even if trickle charging, something in the tenths of A, not in the hundredths.
It's resuidal current that flows via multimeter that should have megaOhm resistance for voltage measurement and very low one for current measurements to not "steal" too many volts/ampers from the source.
I was recently bored one evening so got out all all 15 of my USB pen drives obtained from all over the place. Mainly because I’d recently use Rufus to burn a Windows ISO and it took ages. What surprised me was the massive variation in speeds. Okay so USB 2 was never going to be fast with consistent raw read speeds usually around 20-25MB/s mark. Write speeds varied from horrible <1MB/s on a 16GB thumb sized drive, to average around 5-10Mb/s.
USB 3 varied more. The fastest was a Kingston 127GB data traveller with impressive 242MBs read/118MBs write. Another 64GB Kingston got 107MBs read/52MBs write which isn’t bad. A cheap 16GB unbranded USB 3 from eBay got 123MBs read but only 30MBs write.
@good2see1
no, there is still something that is "off", with the tester set to mA, you are measuring 0.48 mA, i.e. 0.00048 A, please read as "nothing", no surprise that the phone battery is not charging.
Amazon does not manage stock for each seller and item. Items/products with identical ASIN (or whatever makes the product identifiable) are “mixed up” in their fully automated warehouses. When there is only one marketplace seller sending counterfeit products. Their were many reports on this in the past. Even one time I remember a manufacturer collecting evidence by ordering his own products from Amazon, which he invisibly marked in factory to prove it.
I ordered a new USB stick (KIOXIA Transmemory U366 64GB USB 3.2). Tomorrow, I’m going to test it as mentioned on my computer. I’ll plug it in and wait to see if it works.
@good2see1
Sure, the important test was the voltage one (verifying that there was not a short of some kind or other board circuit failure that gave out 12, 24 or 48V on the USB) once checked that the port is at 5 V, the actual amperage drawn from it depends on the device attached to it, hopefully your new USB stick will draw something like 20-80 mA at idle and 200-400 mA (maybe 500) when performing read/write operations.
Firstly i plugged into the pc and test it. Result seems ok.
After that, i plugged into the mikrotik. Usb Drive got hot and not working now. Pc doesnot recognize anymore. Not listed in windows+x > disk management.
This is concerning. I doubt that MikroTik will replace the ruined flash drives.
Lesson. Buy cheap flash, test on known good PC, then test with cheap tested usb on new device to verify the USB isn't fried. Finally use expensive, high performance USB on new device.
All that being said, what actual tests did you do on the PC? a 1 second crystal mark test isn't going to be very useful.
If the results you posted were for the KIOXIA, it appears to have poor random write performance, but that's not uncommon for flash drives. Did you run h2testw or GRC's validrive? I would have expected validrive to be slow given the results you saw, because what validrive does; small reads and writes without any queuing is similar to the Q1T1 results are based on.
What function on the L009 was the flash being used for?
For comparison with a MicroCenter "USB 3.1 64GB flash drive" here are the results from my last purchase. This drive passes the tests, but the performance for small random reads and writes is terrible (but it was cheap). validrive took 527 seconds to complete (checking 576 "random" locations on the drive). A good drive takes a short time (seconds). First the results for a good flash (this was a microSD in a USB 3 reader connected to a USB 3 port. Same computer and USB port used on all tests below.
This is 256GB SanDisk extreme microSD with GRC validrive (completed test in 2.7 seconds).
Report #1
test date and time 10/10/2025 at 4:39 PM
declared drive size 255,869,321,216 (256GB)
validated drive size 255,869,321,216 (256GB)
highest valid region 255,869,321,216 (256GB)
hub or drive vendor realsil
hub or drive product rtsuerlun0
serial number 0000
performance details
read write
samples 1,152 1,152
minimum 1,163 1,742
maximum 2,160 8,928
average 1,427 2,346
median 1,425 2,046
std dev 88 1,051
variance 0.062 0.448
total time 1,644,582 2,703,119
percent 37.83% 62.17%
time measurements in microseconds
Compare above to the MicroCenter USB 3 drive which took nearly 528 seconds to complete.
MC 64GB USB 3 flash drive purchased 2025-06-14 (Purple)
h2testw
Writing 59132 MByte
1:30:38 h
10.9 MBytes/s
Verifying 59132 MByte
13:22 min
73.7 MByte/s
Test finished without errors.
You can now delete the test files *.h2w or verify them again.
Writing speed: 10.9 MByte/s
Reading speed: 73.7 MByte/s
H2testw v1.4
GRC validrive results
Report #1
test date and time 6/14/2025 at 8:57 PM
declared drive size 62,026,416,128 (62.0GB)
validated drive size 62,026,416,128 (62.0GB)
highest valid region 62,026,416,128 (62.0GB)
hub or drive vendor
hub or drive product usb_disk_3.0
serial number 071c516937abc162
performance details
read write
samples 1,152 1,152
minimum 1,020 2,289
maximum 4,359 1,658,307
average 1,649 458,240
median 1,621 157,602
std dev 342 625,022
variance 0.208 1.364
total time 1,900,318 527,893,483
percent 0.36% 99.64%
time measurements in microseconds
