Question FAKE Samsung 970 NVME??

JFaTaL12

Prominent
Dec 27, 2020
10
0
510
So, this is weird. bought from Best Buy 2 months ago and I randomly just check my hardware info and why is my M.2 Folder name ssdpeknw512g8 ???
Did I just get scammed or am I missing something
 
I have had three 970 EVO Plus drives that have needed firmware updates AND deleting them from the drives listing in device manager, plus a reboot, in order to get them to be identified correctly. I JUST went through this again this week. They didn't look (In my case) like they were some other drive brand, but two of them were identifying themselves as the wrong SIZE of 970 EVO Plus (Although they actually WERE the correct size. They were showing up as 250GB drives but clearly had 1TB of drive space) and one was showing just random letters/gibberish as the name of the drive in both disk management and device manager. In all three cases a firmware update on the drives using Samsung magician, plus uninstalling them from device manager and then restarting the system corrected the problem.

That doesn't mean it's the same problem as yours, but I'd certainly try it before throwing in the towel.

As to the problems with Samsung's 128 layer TLC NAND drives, I do not believe that. Sorry Fzabkar, but I own and have installed far too many of these drives to believe that this is anywhere near as widespread as you have been broadcasting it to be. Even guys who are pretty knowledgeable about a thing can sometimes see wolves where there are only cattle dogs. I have bought more than 20 970 EVO and 970 EVO Plus drives over the last two years and aside from the misreporting of the name and size, I've had no problems with ANY of them. I've also bought a number of other drives including the 870 EVO that uses their 128 layer TLC NAND, and none of those drives have had any issues either and they've all been in service long enough that if there were going to be problems as outlined, they'd have happened by now.
 
When a drive identifies itself with an incorrect capacity or a strange model number or "gibberish", it is most likely a sign that the firmware has panicked and the drive has reverted to "safe mode" or "ROM mode". The root cause is bad NAND, which in turn results in a corrupt FTL (Flash Translation Layer).

Here are several examples:

http://vlo.name:3000/hw/ssd/rommode/

I don't understand what happened in your case, or how a firmware update could have resolved it.

The 840 Evo is another product that was affected by bad NAND, according to Russian researchers. Samsung "fixed" the problem with a firmware update. The result of the update was that the drive continuously refreshed its NAND cells the whole time it was powered on. Most users would have been none the wiser.

Samsung 840 EVO - how to easily lose your data forever:

https://forum.acelab.eu.com/viewtopic.php?f=227&t=8735
 
Last edited:
  • Like
Reactions: Dark Lord of Tech
When Tom's Hardware reviewed the 970 Evo in 2018, it was using 64-layer NAND:

https://www.tomshardware.com/reviews/samsung-970-evo-ssd-review,5573.html

AFAICT, the more recent 970 Evo Plus is using 128-layer NAND.

Here is a SMART report from a 970 Evo in 2019:

https://www.techpowerup.com/forums/threads/samsung-860-evo-wear-level-question.266784/
https://www.techpowerup.com/forums/attachments/1590824283149-png.157206/

The Power On Hours count would suggest that the drive was bought before June 2019.

The Available Spares value has dropped by 1% (from 0x64 to 0x63), but there are no Media and Data Integrity Errors.

It seems that Samsung changed the BOM some time during the life of the 970 Evo, or perhaps it's just the Evo Plus that is/was affected.
 
I assure you, I don't lie nor do I "stretch" the truth. So it seems that, much as you know, there are possibilities that exist which you seem to discount at your leisure sometimes. This is not the first occasion.

Before the firmware update and removal from device manager. You can click the images to enlarge them.

VhT5qD4.png





After the firmware update, removal from device manager and restart of the system. Just did this a few days ago, on THIS one. The second 250GB, the one that is ACTUALLY 250GB, isn't listed in the second set of shots because I had to rob it for use in another board I just received from a fellow moderator for some testing. There were two others that happened previously with same result.

vTNyEPo.png
 
Last edited:
Thanks for that. It's really strange, but it's definitely not a "ROM mode" issue. Unfortunately, Samsung doesn't provide changelogs with its updates. :-(

FWIW, I have an 860 Evo 1TB SATA SSD which has racked up 20,000 hours with no bad sectors. I did have initial teething problems (because I let the drive sit on the shelf for 1 year before I installed it), but otherwise the drive has run well.
 
This review would suggest that a BOM change might explain the different reliability experiences.

Samsung 970 EVO Plus 250 GB (Phoenix + V5):
https://www.techpowerup.com/ssd-specs/samsung-970-evo-plus-250-gb.d60

This first version of the 250GB 970 Evo Plus had ...

Phoenix (S4LR020) controller​
K9DUGY8J5B-DCK0 92-layer TLC V-NAND​

2nd Revision might came with different Nand and the new "Elpis Controller" found in the Samsung 980 Pro.
 

Karadjgne

Titan
Ambassador
The 840 Evo is another product that was affected by bad NAND, according to Russian researchers.
So the Russian engineers and researchers claim to know more about a Korean based Leading manufacturer and designer of ssds silicon than the engineers and researchers of that same Korean company? I find that a little 'wrong' because if such was the case, we'd all be using Cyrillic ssds, not Korean ssds.

It's common practice to change stuff according to Region or Market, especially with patent concerns, so IMHO what the Russian people got wasn't the exact same as what the US or European markets got, so that could explain any discrepancy and why Only Russian researchers and not US or Euro researchers found corroborating evidence.
 

USAFRet

Titan
Moderator
None of my samsungs, from 250GB 840 EVOs through the 850-860 EVOs, to the new 980 Pro have died, or even slowed down from original out of the box performance.

The 840 EVOs were in 24/7 use from Nov-Dec 2014 to Dec 2021.
And yes, they got that original firmware update.


Looking just now, one of my 1TB 860 EVOs has 32,620 running hours.
 
Last edited:
So the Russian engineers and researchers claim to know more about a Korean based Leading manufacturer and designer of ssds silicon than the engineers and researchers of that same Korean company? I find that a little 'wrong' because if such was the case, we'd all be using Cyrillic ssds, not Korean ssds.

It's common practice to change stuff according to Region or Market, especially with patent concerns, so IMHO what the Russian people got wasn't the exact same as what the US or European markets got, so that could explain any discrepancy and why Only Russian researchers and not US or Euro researchers found corroborating evidence.

The Russian researchers (Ace Laboratory) write their own firmware ("loaders") to recover data from these SSDs. When these SSDs fail due to bad NAND, the firmware "panics" and denies access to the user's data. Ace Laboratory's data recovery tool, PC3000 SSD, uploads their own loader code to the SSD. PC3000 then builds a "virtual translator" (because the SSD's own translator is broken) and the tool then assembles and downloads the good and bad sectors to an image file. Data is then recovered from this image. I would say that this requires a very detailed knowledge of the firmware architecture, don't you think?

As for US researchers, there are none (except for Russian expats). When you send your broken drive to a data recovery shop in the US or in the Eurozone, the shop will most likely be using PC3000. The only other tool of note is MRT Lab, a Chinese product. Yes, it's true, the Russians and Chinese know a lot more about the firmware in storage devices than Americans and Europeans. Be afraid, be very afraid.

Here is an explanation (they are no doubt using PC3000):

https://www.dataclinic.co.uk/samsung-evo-840-ssd-faults-beware/

In order to communicate with this SSD our technicians were required to place the drive into safe mode. Safe mode on this model was instigated by shorting 2 pins on the PCB. Once safe mode was attained we then uploaded microcode to the RAM in order to simulate the normal drive initiation process, this allowed direct physical access to the NAND storage and allowed for basic read testing too be conducted. Once completed it was clear that although each chip was reading, 2 were suffering from read errors throughout.

After some further firmware fixes from our technicians we were able to gain logical access to the drive and begin a clone of the device over to a new stable HDD. After approximately a day our tools had produced the best possible copy of the failed SSD and from approx. 512,000,000 sectors we were unable to read 5,000,000 – these were the degraded cells. We then provided a full file list of both the good and bad files to our client who was over the moon with the result and completed the recovery case.


BTW, when I say "research", I'm referring to "reverse engineering".
 
Last edited:
None of my samsungs, from 250GB 840 EVOs through the 850-860 EVOs, to the new 980 Pro have died, or even slowed down from original out of the box performance.

The 840 EVOs were in 24/7 use from Nov-Dec 2014 to Dec 2021.
And yes, they got that original firmware update.


Looking just now, one of my 1TB 860 EVOs has 32,620 running hours.

Your 840 Evo's usage pattern was such that the charge never had a chance to "flow away", as the Russian researchers put it. Samsung's firmware update applied a bandaid over the problem, and it was then swept under the rug.

Here's another story, from a data recovery company:

https://www.dataclinic.co.uk/samsung-evo-840-ssd-faults-beware/

It explains the NAND problems with earlier Samsung models, plus it explains the technique used to recover the data.
 

USAFRet

Titan
Moderator
Your 840 Evo's usage pattern was such that the charge never had a chance to "flow away", as the Russian researchers put it. Samsung's firmware update applied a bandaid over the problem, and it was then swept under the rug.
From your link:
"This was then left unused for approx. 5 months, no reading/writing to the drive was completed nor was power supplied. When we later attached this up to our test equipment the partition and large amounts of the drive were unreadable proving if left long enough the voltage inside the cells dissipates and data is lost. "

That sort of 'fault' is/was not unique to Samsung drives.
You of course remember the brouhaha over an early "test" of treating drives like that.


So in this case, Samsung identified a potential issue, and published a firmware fix for it.
I fail to see the problem.
 
From your link:
"This was then left unused for approx. 5 months, no reading/writing to the drive was completed nor was power supplied. When we later attached this up to our test equipment the partition and large amounts of the drive were unreadable proving if left long enough the voltage inside the cells dissipates and data is lost. "

That sort of 'fault' is/was not unique to Samsung drives.
You of course remember the brouhaha over an early "test" of treating drives like that.


So in this case, Samsung identified a potential issue, and published a firmware fix for it.
I fail to see the problem.

It's like saying, "My car won't start if I leave it in the garage for a month, so I leave the engine running'".