Question Team Group SSD Possibly Misadvertised?

[NightHawkRMX]

I was looking for a cheap SSD, and I came across this SSD.
Team Group T-FORCE VULCAN G 2.5" 512GB Internal SSD - Newegg.com
I saw this review and figured the Vulcan G would just be a different variation. Boy was I wrong.
Team Group T-Force Vulcan SATA SSD Review: Flashy, Fast and Affordable | Tom's Hardware

It arrived today and I was curious if this SSD had DRAM, so I opened the SSD up.

Here is a shot of the inside: https://ibb.co/vcQjG5p
Another: https://ibb.co/wBnDPrs
Outside: https://ibb.co/KGWD83L

The controller is a Silicon Motion SM2258XT, and there is no dram, which makes it worse than the Vulcan G, but no claims were made about dram.
en.siliconmotion.com/download/product-brief/SM2258XT_Product_Brief_ENG_Q1109.pdf

HOWEVER, I noticed this SSD was advertised as having an SLC Cache. I am 99% sure it does not.

The SM2258XT is the same one used in the TLC Crucial BX500. The datasheet says "The SM2258XT is a complete merchant ASIC/firmware solution supporting 1z nm TLC and 3D NAND from all major NAND suppliers," indicating the NAND used is not SLC, and no reference is made to support for any sort of SLC cache, nor does their appear to be any SLC chips on the PCB.

The fact I am not getting the SLC advertised is bad enough, but not as bad as what I discovered next, assuming this is accurate.

If I google the numbers written on the NAND, I cannot find anything, however, I shared this with a few discord servers I am in. Immediately one person flagged them and said they were likely fake Toshiba, then linked me this "https://www.chinahao.com/product/580665668642/" and said it seems they are only available from a few Chinese sellers on sites like aliexpress. Toshiba does not have this product listed on their official website like they do for most products. It appears to be TLC.

I asked in another server and the conversation went as follows:

-Other person "Found the chips"
-Other person "TLC"
-Me "Yea I knew that since the controller only supports tlc" (this is not entirely true, now that I have done more research)
-Other person "And they are subgrade sales"
-Me "Which means"
-Other person "lol basically meant for USB Flash Drives since they didnt meet the BIN reqs"
-Other person "Discards"
-Me "Idk what to do about this"
-Other person "removed they are sold to KIOXIA Exclusively who is known to fake a bunch of removed [stuff] for China markets"
-Other person "Hahah"
-Other person "I would not trust the integrity of that drive"
-Me "Then how did I get them"
-Me "Or team rather"
-Other person "Remember, Team... is ran by a team of Human Beings"
-Other person "I would not doubt it is them that got jipped"
-Me "My ram is teamgroup"
-ME "It's proper hynix djr tho"
-Other person "Hahah it even says on the datasheet in Korean "Not meant for Permanent Storage Use"
-Other person "Emulated TRIM Support, Not full TRIM!!!!"
-Other Person "Lol the pre-controller on the flash is even gimped amazing"

This other person claims to work in the IT field and he wanted to link me where he found this information including a datasheet, however I told him not to since he said the website was a Russian forum and had questionable advertising. Whether or not this is legitimate or not, I do not know, however the fact this SSD is already not delivering what is advertised, I would not be surprised if this is the case.

I WILL NOT BE BUYING ANY MORE TEAMGROUP PRODUCTS.

 

[NightHawkRMX]

I have been informed Toshiba engraves their chips where cheap fakes often use printing. Tomorrow I will pick up some more alcohol and see if it comes off.

 

[fzabkar]

If you reduce the levels in a TLC from 8 to 2, then you create a "pseudo SLC".

https://www.electronicdesign.com/in...es-criticalapp-reliabilitywhile-cutting-costs

 

[NightHawkRMX]

This is very interesting. I do not see any mention of this functionality of the controller, however. I wonder if this functionality could be programmed into the firmware?

 

[rabbidreader]

I was looking for a cheap SSD, and I came across this SSD.
Team Group T-FORCE VULCAN G 2.5" 512GB Internal SSD - Newegg.com
I saw this review and figured the Vulcan G would just be a different variation. Boy was I wrong.
Team Group T-Force Vulcan SATA SSD Review: Flashy, Fast and Affordable | Tom's Hardware

It arrived today and I was curious if this SSD had DRAM, so I opened the SSD up.

Here is a shot of the inside: https://ibb.co/vcQjG5p
Another: https://ibb.co/wBnDPrs
Outside: https://ibb.co/KGWD83L

The controller is a Silicon Motion SM2258XT, and there is no dram, which makes it worse than the Vulcan G, but no claims were made about dram.
en.siliconmotion.com/download/product-brief/SM2258XT_Product_Brief_ENG_Q1109.pdf

HOWEVER, I noticed this SSD was advertised as having an SLC Cache. I am 99% sure it does not.

The SM2258XT is the same one used in the TLC Crucial BX500. The datasheet says "The SM2258XT is a complete merchant ASIC/firmware solution supporting 1z nm TLC and 3D NAND from all major NAND suppliers," indicating the NAND used is not SLC, and no reference is made to support for any sort of SLC cache, nor does their appear to be any SLC chips on the PCB.

The fact I am not getting the SLC advertised is bad enough, but not as bad as what I discovered next, assuming this is accurate.

If I google the numbers written on the NAND, I cannot find anything, however, I shared this with a few discord servers I am in. Immediately one person flagged them and said they were likely fake Toshiba, then linked me this "https://www.chinahao.com/product/580665668642/" and said it seems they are only available from a few Chinese sellers on sites like aliexpress. Toshiba does not have this product listed on their official website like they do for most products. It appears to be TLC.

I asked in another server and the conversation went as follows:

-Other person "Found the chips"
-Other person "TLC"
-Me "Yea I knew that since the controller only supports tlc" (this is not entirely true, now that I have done more research)
-Other person "And they are subgrade sales"
-Me "Which means"
-Other person "lol basically meant for USB Flash Drives since they didnt meet the BIN reqs"
-Other person "Discards"
-Me "Idk what to do about this"
-Other person "removed they are sold to KIOXIA Exclusively who is known to fake a bunch of removed [stuff] for China markets"
-Other person "Hahah"
-Other person "I would not trust the integrity of that drive"
-Me "Then how did I get them"
-Me "Or team rather"
-Other person "Remember, Team... is ran by a team of Human Beings"
-Other person "I would not doubt it is them that got jipped"
-Me "My ram is teamgroup"
-ME "It's proper hynix djr tho"
-Other person "Hahah it even says on the datasheet in Korean "Not meant for Permanent Storage Use"
-Other person "Emulated TRIM Support, Not full TRIM!!!!"
-Other Person "Lol the pre-controller on the flash is even gimped amazing"

This other person claims to work in the IT field and he wanted to link me where he found this information including a datasheet, however I told him not to since he said the website was a Russian forum and had questionable advertising. Whether or not this is legitimate or not, I do not know, however the fact this SSD is already not delivering what is advertised, I would not be surprised if this is the case.

I WILL NOT BE BUYING ANY MORE TEAMGROUP PRODUCTS.

I was looking for a cheap SSD, and I came across this SSD.
Team Group T-FORCE VULCAN G 2.5" 512GB Internal SSD - Newegg.com
I saw this review and figured the Vulcan G would just be a different variation. Boy was I wrong.
Team Group T-Force Vulcan SATA SSD Review: Flashy, Fast and Affordable | Tom's Hardware

It arrived today and I was curious if this SSD had DRAM, so I opened the SSD up.

Here is a shot of the inside: https://ibb.co/vcQjG5p
Another: https://ibb.co/wBnDPrs
Outside: https://ibb.co/KGWD83L

The controller is a Silicon Motion SM2258XT, and there is no dram, which makes it worse than the Vulcan G, but no claims were made about dram.
en.siliconmotion.com/download/product-brief/SM2258XT_Product_Brief_ENG_Q1109.pdf

HOWEVER, I noticed this SSD was advertised as having an SLC Cache. I am 99% sure it does not.

The SM2258XT is the same one used in the TLC Crucial BX500. The datasheet says "The SM2258XT is a complete merchant ASIC/firmware solution supporting 1z nm TLC and 3D NAND from all major NAND suppliers," indicating the NAND used is not SLC, and no reference is made to support for any sort of SLC cache, nor does their appear to be any SLC chips on the PCB.

The fact I am not getting the SLC advertised is bad enough, but not as bad as what I discovered next, assuming this is accurate.

If I google the numbers written on the NAND, I cannot find anything, however, I shared this with a few discord servers I am in. Immediately one person flagged them and said they were likely fake Toshiba, then linked me this "https://www.chinahao.com/product/580665668642/" and said it seems they are only available from a few Chinese sellers on sites like aliexpress. Toshiba does not have this product listed on their official website like they do for most products. It appears to be TLC.

I asked in another server and the conversation went as follows:

-Other person "Found the chips"
-Other person "TLC"
-Me "Yea I knew that since the controller only supports tlc" (this is not entirely true, now that I have done more research)
-Other person "And they are subgrade sales"
-Me "Which means"
-Other person "lol basically meant for USB Flash Drives since they didnt meet the BIN reqs"
-Other person "Discards"
-Me "Idk what to do about this"
-Other person "removed they are sold to KIOXIA Exclusively who is known to fake a bunch of removed [stuff] for China markets"
-Other person "Hahah"
-Other person "I would not trust the integrity of that drive"
-Me "Then how did I get them"
-Me "Or team rather"
-Other person "Remember, Team... is ran by a team of Human Beings"
-Other person "I would not doubt it is them that got jipped"
-Me "My ram is teamgroup"
-ME "It's proper hynix djr tho"
-Other person "Hahah it even says on the datasheet in Korean "Not meant for Permanent Storage Use"
-Other person "Emulated TRIM Support, Not full TRIM!!!!"
-Other Person "Lol the pre-controller on the flash is even gimped amazing"

This other person claims to work in the IT field and he wanted to link me where he found this information including a datasheet, however I told him not to since he said the website was a Russian forum and had questionable advertising. Whether or not this is legitimate or not, I do not know, however the fact this SSD is already not delivering what is advertised, I would not be surprised if this is the case.

I WILL NOT BE BUYING ANY MORE TEAMGROUP PRODUCTS.

I am not doubting anything you have said, BUT, if this is true, it is pretty damning. I purchased some Team Group memory in the past and now I will Have to reconsider doing any repeat business with them, if this is the type of practice they are following. It pays to research and have a good group like this one to be informed.

 

[fzabkar]

This is very interesting. I do not see any mention of this functionality of the controller, however. I wonder if this functionality could be programmed into the firmware?

I would think that this would be a firmware function. In fact I've been wondering if it would be possible to convert an entire QLC NAND array into a pseudo-MLC. The total capacity would be halved, but there might be reliability benefits.


Edit: Seagate's Nytro enterprise SSDs are built with eMLC NAND flash.

https://www.ni.com/en-au/support/do...tanding-life-expectancy-of-flash-storage.html

eMLC NAND: Enterprise MLC flash that stores one bit of data in an MLC architecture, offering a midpoint between SLC and MLC in cost, endurance, and capacity. May also be called pSLC (Psuedo SLC) or iMLC (industrial-grade MLC)

 

[NightHawkRMX]

I will have to stick the drive in a spare pc to see what behavior it exhibits over a larg load

 

[fzabkar]

I found this document informative:

https://nebula.wsimg.com/d88bcbf18f...D046C045384969673&disposition=0&alloworigin=1

It would appear that many (all?) manufacturers are using pseudo SLC (pSLC) in some way. In particular, SSDs store their firmware in NAND, so there has to be a high reliability area set aside for this purpose. This means that some SSDs, eg Toshiba, will convert MLC (or TLC or QLC) cells to pSLC for their firmware modules, and they do this transparently. Hard drives also increase the reliability of their platter resident firmware modules by recording them at a lower bit rate.

 

[NightHawkRMX]

That's very good to hear, so I guess I was somewhat wrong. It would have been nice if they said PSUDO SLC cache, and not just SLC cache, but I guess it really doesn't matter either way.

The NAND is still very questionable but so is the information from that Russian site. Either way, the Vulcan G SSD is going back since the Samsung 860 EVO 500g was basically only $10 more.

I also bought this since its literally $67 for 1tb SSD and will disassemble it. I have a feeling its similar.
Team Group CX2 2.5" 1TB SATA III 3D NAND Internal Solid State Drive (SSD) T253X6001T0C101 - Newegg.com

 

[NightHawkRMX]

I just ran some alcohol on a tissue over the printing on the nand chips and the printing did not come off. I still find it awfully sus that these chips are not listed on toshibas site

 

[fzabkar]

You can relax. Seagate's Ironwolf 110 SSD for NAS uses the same Toshiba flash (TH58TFT1V23BA8H).

https://www.linuxlookup.com/review/seagate_ironwolf_110_ssd_nas_review


Part Number Decoder for Toshiba NAND Flash (Sept 24, 2010):
http://yourcmc.ru/wiki/images/2/2a/Toshiba_Nand_Flash_naming.pdf

TH58TFT1V23BA8H

• TH58 = multichip
  T = Toggle mode NAND
  F = supply voltage
  T1 = 2 (= 2^1 ) Tbits = 256GB
  V = cell levels
  2 = page/block size = >4KB/512KB
  3 = design rule
  BA = BGA package, lead free, halogen free
  8 = Single / Dual channel, 4 CEs, TSOP, BGA,LGA
  H = package dimensions

 

[NightHawkRMX]

Well now I just feel stupid for making such accusations.

 

[fzabkar]

The IC manufacturers bring suspicion upon themselves and their clients by not making their datasheets publicly available. In fact I'm suspicious about every claim that comes out of a marketing department.

 

[NightHawkRMX]

https://imgur.com/xfWpNPm

View: https://imgur.com/xfWpNPm

Just a quick benchmark.

I tried to run an extended file copy from my NVME to the Team SSD to test the cache and given how fast it expired, but that didnt go well and I have no good way to test it.

 

[NightHawkRMX]

You can relax. Seagate's Ironwolf 110 SSD for NAS uses the same Toshiba flash (TH58TFT1V23BA8H).

https://www.linuxlookup.com/review/seagate_ironwolf_110_ssd_nas_review


Part Number Decoder for Toshiba NAND Flash (Sept 24, 2010):
http://yourcmc.ru/wiki/images/2/2a/Toshiba_Nand_Flash_naming.pdf

TH58TFT1V23BA8H

• TH58 = multichip
  T = Toggle mode NAND
  F = supply voltage
  T1 = 2 (= 2^1 ) Tbits = 256GB
  V = cell levels
  2 = page/block size = >4KB/512KB
  3 = design rule
  BA = BGA package, lead free, halogen free
  8 = Single / Dual channel, 4 CEs, TSOP, BGA,LGA
  H = package dimensions

Wait. If they are using pseudo SLC, they would be disabling some of the capacity on at least one of the chips. This drive has 2x256gb chips but is 512gb total capacity, so how is it possible for them to have the pseudo slc and still have the full capacity?

 

[fzabkar]

An SSD always has overprovisioned space. The pSLC would live in this space.

The capacity of the drive is 512GB (= 512 x 1000 x 1000 x 1000), but the capacity of the chips is 512GiB (= 512 x 1024 x 1024 x 1024). The difference is about 7%.

 

[NightHawkRMX]

Hmm, so that would mean the SLC portion would have to be pretty small, though. Better than nothing.

The SLC on my Sabrent 1tb is 250g i think

 

[fzabkar]

If TRIM is enabled, then the OS can tell the SSD that certain parts of the user area are not in use by the file system. I would think that this free space could then be dynamically assigned to pSLC. It's only when the drive is full that the pSLC would need to be totally confined to the overprovisioned space. Maybe that's the difference between dynamic and static pSLC cache.

 

[Maxxify]

Yes, static pSLC comes from the original cut of flash - you have an amount of flash that's reduced first by bad blocks, then reserved for spare/ECC, then part for native OP (7% for binary -> decimal) and marketed OP (additional) with some set aside potentially for static SLC. Dynamic SLC is made from all TLC, OP or not, and is shifted/rotated based on wear, however the drive will try to always maintain some when empty. Typically this is 6-12GB per TB and likewise static tends to be either 6GB or 12GB per TB (e.g., 970 EVO vs. SN750).

The "disabling of some flash" is accurate as static SLC is set aside per die. This is a bit complicated as there's a balance between flash used for ECC, OP, and static SLC, and is in fact the subject of much study. If you interested in learning more about these subjects, check me out on Reddit including my guides (check my signature).

Re: your 250GB SLC on Sabrent

This is known as full-drive SLC caching, e.g. 250GB out of a 1TB drive with QLC, and is NOT necessarily a good thing.

Re: TRIM

A modern drive will have dynamic overprovisioning based on TRIM + free user space, dynamic SLC is a bit more complicated as there are algorithms that determine its behavior. In fact it's gotten very complex on newer drives, for example I have a patent for Crucial's P5 that explains behavioral profiles and treatment of data types.

Re: pSLC in general

It's not real SLC and neither does it have the same properties due to process node, wiring, etc. Real SLC may be 100K P/E while pSLC is 30-40K P/E (static) or much lower (dynamic). It's a complicated subject, read my Basics + posts for more or hit me up on here/Reddit.

 

[fzabkar]

@Maxxify, I was hoping you would drop by.

I believe this is your resource:

http://ssd.borecraft.com/

 

[NightHawkRMX]

It's not real SLC and neither does it have the same properties due to process node, wiring, etc. Real SLC may be 100K P/E while pSLC is 30-40K P/E (static) or much lower (dynamic). It's a complicated subject, read my Basics + posts for more or hit me up on here/Reddit.

I guess that's why it says "SLC caching" not SLC Cache?

I still feel its a bit misleading, since it does not say it is not proper SLC.

 

[Maxxify]

@Maxxify, I was hoping you would drop by.

I believe this is your resource:

http://ssd.borecraft.com/

I will have to update everything for the new year as we have a new crop of drives coming in, and actually quite a bit new technology as well. I'm also leaning more heavily enterprise these days as I'm writing white papers for a major manufacturer which has expanded my knowledge. But yep, the general links are there on my website!

 

[Maxxify]

I guess that's why it says "SLC caching" not SLC Cache?

I still feel its a bit misleading, since it does not say it is not proper SLC.

Yes, technically. There are drives with actual SLC (or even 3D Xpoint) for caching/tiering but they operate a bit differently.

Full-drive SLC is still manufactured but generally for storage class memory (SCM) as an alternative to 3D XPoint technologies (which if course means pSLC doesn't match native SLC performance, either). When I say true/native SLC is different than pSLC it's because the standards required are much higher. However, static pSLC specifically can get very high endurance - an order-of-magnitude higher than native flash - to the point there are retail pSLC drives made entirely from TLC in SLC mode. That topic is not for consumer usage and can get complicated as previously there were different types of flash modes (like fast page) and pSLC even with higher endurance has issues that deviate from native SLC, for example with error correction (esp. over time) which is also dependent on flash type (FG, CTF).

 

[NightHawkRMX]

My TEAM CX2 1tb came in today. It is identical to the GX2 512gb except for 2 more NAND chips on the reverse of the PCB.

I guess these would be on par with a BX500? Same controller and similar nand just from Toshiba not Micron