Is 'Estimated Drive Remaining Life' a cause for concern?

[Inyerbones]

I have a XPG SPECTRIX S40G 512GB NVME PCIE 3x4 SSD : https://www.xpg.com/en/feature/610/

I've been using this SSD as my boot drive since August of 2019, so its relatively new

 

But I'm wondering If what I'm seeing in the screenshots attached below is worth worrying about? 

 

 

Few months ago I noticed it was around 70%, then it was 50% about 2 weeks ago, then I noticed it went down to 49%, followed by 48% and now 47% as days go by. 

Is my SSD simply just degrading that fast? 

Will it actually just die once it reaches 0%?

 

Thanks in advance.

 

[Bombastinator]

4 minutes ago, Inyerbones said:

I have a XPG SPECTRIX S40G 512GB NVME PCIE 3x4 SSD : https://www.xpg.com/en/feature/610/

I've been using this SSD as my boot drive since August of 2019, so its relatively new

 

But I'm wondering If what I'm seeing in the screenshots attached below is worth worrying about? 

 

 

Few months ago I noticed it was around 70%, then it was 50% about 2 weeks ago, then I noticed it went down to 49%, followed by 48% and now 47% as days go by. 

Is my SSD simply just degrading that fast? 

 

Thanks in advance.

 

Looks like it. 7tb of writes is a lot. You’re writing as much as you’re reading.

[Kilrah]

Looks like either the reporting is buggy, or you managed to buy the one SSD with the shortest life ever made.

 

It says you have 7.27TBW, and the datasheet rates it at 320TBW.

So according to the datasheet you're at 97.73% life left. 

 

So they likely botched their firmware reporting.

[kokosnh]

Post smart from crystaldiskinfo 

[Mihle]

On 3/1/2020 at 9:43 PM, Bombastinator said:

Looks like it. 7tb of writes is a lot. You’re writing as much as you’re reading.

7TB is not a lot at all, per month basis it's more than my 250gb one but still,  it's little compared to what the drive is rated for, as someone else said it's probably just the software used that is bugged.  

[NewMaxx]

Yep, check health with CDI. Also your temperature is rather high on that drive if it's idle, it throttles starting at 70C and idling at ~65C would be fairly...bad.

[Inyerbones]

On 3/2/2020 at 5:16 AM, kokosnh said:

Post smart from crystaldiskinfo 

 

On 3/3/2020 at 7:00 AM, NewMaxx said:

Yep, check health with CDI.

Not too familiar with crystaldiskinfo, but is this what you guys are talking about?

 

 

On 3/3/2020 at 7:00 AM, NewMaxx said:

Also your temperature is rather high on that drive if it's idle, it throttles starting at 70C and idling at ~65C would be fairly...bad.

That's something that's been bothering me slightly as well, I think that's probably caused by the way my case is set up, I have the PC-O11 Dynamic case with a 360mm AIO mounted up top for exhaust, and 6 120mm fans for intake, 3 on the bottom and 3 on the side. I have a custom fan curve set for all the fans, but I believe the main thing that's causing the SSD to heat up like this is because my graphics card is vertically mounted with the Cooler Master Vertical GPU Mounting kit. My GPU is the RTX 2070 Super Aorus, and it seems that it has quite an effective backplate for dissipating heat from the back of the card, and I think the heat is radiating from the backplate to the heatsink of my SSD which is located on the first M.2 slot on my motherboard, directly behind my graphics card.

I dont think it's getting much airflow from the bottom intake fans, because the riser cable for the graphics card is literally in between the SSD and the bottom intake fans, while the air from my side intake fans are being blocked by mainly my RAM sticks. For the most part it stays under 70C, but I have seen it peaking at 85C sometimes while checking HWinfo, not sure from what kind of a load though. I don't really do file transfers often, most of my games are stored in my HDD anyway. SSD only has OS and a few of the games I play most often which equates to less than 100GB total.

 

Is it a big problem to have my SSD running at these temps? Besides the transfer speed throttling, which is something I can live with because I dont really transfer much anyway. It's mainly a gaming rig.
 

[NewMaxx]

Good news and bad news. The good news: it doesn't look like it's damaged in any way. You still have 100% space blocks (64 in hex). Percentage/lifespan used is 35 (53 in decimal) which is where HWInfo is getting 47% (100 - 53 = 47) from, which is not normal. However the data units written matches up - 15411309x512B (sector size) = 7348GB. So I'm not sure why it's reporting that way, although CDI is ignoring it.

 

The bad news: 64C idle is rather bad. This is the temperature of the controller and it will throttle over 70C. With regards to the flash: NAND likes it hot when programming, however heat during data retention/reading will be detrimental (although, it will be reflected by more writes generally). I suspect temperature is the real culprit here.

[demonix00]

I think it's less to do with the GPU and more to do with the fact that you have several LEDs on that SSD which also generate heat when in use, so you're either going to have to turn that stuff off or get an SSD that isn't going to cook itself with the aid of several LEDs.

[Inyerbones]

On 3/5/2020 at 3:45 AM, NewMaxx said:

Good news and bad news. The good news: it doesn't look like it's damaged in any way.

That's really good to hear. Thanks.

 

 

On 3/5/2020 at 3:45 AM, NewMaxx said:

The bad news: 64C idle is rather bad. This is the temperature of the controller and it will throttle over 70C. With regards to the flash: NAND likes it hot when programming, however heat during data retention/reading will be detrimental (although, it will be reflected by more writes generally). I suspect temperature is the real culprit here.

Will this cause it to eventually just fail on me at some point in the future? Or am I just dealing with a risk of data loss? Since the rated operating temperatures of the drive is up to 70C.

 

On 3/5/2020 at 6:56 AM, demonix00 said:

I think it's less to do with the GPU and more to do with the fact that you have several LEDs on that SSD which also generate heat when in use, so you're either going to have to turn that stuff off or get an SSD that isn't going to cook itself with the aid of several LEDs.

I get that this was the case with the HyperX rgb sata drives, but It's not the case with this SSD I have. It does the same thing that the Team Group RGB sata drives do, which is that it is designed in a way where the heat from the LEDs don't get to the important parts of the SSD. 

 

To verify that the LEDs are not the issue myself, I turned the LEDs off, reset the sensor recordings on HwInfo and let it sit for 2.5 hours while I was out. 

As you can see, peak temps were at 69C. My graphics card has a fan stop mode, and its configured so that the fans don't spin unless it hits 63C. 

Minimums are at 62C but that's because I had the computer running for about 3 or 4 hours already before I reset the recordings, and it was at 63C at the time I turned off the LEDs, then I went out and just let it sit.
I haven't tested temps with a non-vertical gpu mount or a graphics card that doesn't have a backplate, but just wanted to make it clear that the temps aren't caused by the LEDs.

[NewMaxx]

The "drive remaining life" value is usually determined by the number of writes, e.g. in comparison to TBW, so I'm not sure why yours is off by that much. However, drives can and usually do survive well past 0% on that metric, sometimes ten times over. Which is why I was looking at the spare block metric instead - you haven't used any spare blocks yet, when you see that change is when you should be more concerned.

 

Long-term I have to believe those high temperatures are bad for the drive. Just not sure what you can do about it since you care about aesthetics.

[Inyerbones]

3 minutes ago, NewMaxx said:

The "drive remaining life" value is usually determined by the number of writes, e.g. in comparison to TBW, so I'm not sure why yours is off by that much. However, drives can and usually do survive well past 0% on that metric, sometimes ten times over. Which is why I was looking at the spare block metric instead - you haven't used any spare blocks yet, when you see that change is when you should be more concerned.

 

Long-term I have to believe those high temperatures are bad for the drive. Just not sure what you can do about it since you care about aesthetics.

I see, thanks a lot for the explanation  

 

I'll mess around with my fan curve settings I guess, and see if I can try to get it under 60C without being too loud.

[RasmusDC]

NVME disks like HEAT, the operate better under heat, i really don´t think you have to be nerveous about it.. 

 

my SN750 Kingston with the "headspreader" from my motherboard mounted on it, is ...

 

48 degrees, 9 GB reads 6 GB writes and 100% drive health. 578 power cylces. '

 

I have 5 other SSD apart from that NVME but those are all SATA they are between 28-40 degrees, the olderst a 830 Samsung... has drive remaining life. 92% (HWinfo64) it has had power on life of 1.32 Years and cyle of 3492 times it only has 19.4 TB written

[BlackManINC]

@Inyerbones

 

If you are concerned about losing everything because of your storage device failing in the near future, just back everything up on an external drive with a system image file, that way your mind can rest easy. I recently created a system image for my PC and my Dads laptop. I can sleep at night a lot easier now. 

[Godlygamer23]

Some info here for you:

https://www.atpinc.com/blog/ssd-data-retention-temperature-thermal-throttling

https://www.sciencedirect.com/science/article/abs/pii/S0026271418305079

 

I would definitely get your temperatures under control. The controller is only rated for so much, and in my case, my NVMe drive did exceed spec a few times which may also happen to you, which will definitely reduce the life of the drive. Even if the NAND flash like higher temperatures during writes, the same doesn't necessarily apply when you're trying to read that information. 

[NewMaxx]

2 hours ago, RasmusDC said:

NVME disks like HEAT, the operate better under heat, i really don´t think you have to be nerveous about it.. 

NAND/flash likes heat, not the controller or other components. And it only likes heat when programming/writing, otherwise it reduces data retention time. JEDEC rates consumer SSDs for 40C when programming with an idle temp of 30C. Although drives with a single sensor are reporting controller temperature which is only an estimate, however if he's regularly 60-70C idle with that sensor you can believe the NAND is also cooking which is most certainly not good.

 

Sources: JEDEC (pp.26-27) - Micheloni et. al. (2018). Inside Solid State Drives. 2nd ed. p. 222. Also Tiomkin: "data loss can occur if the NAND is stored or read over extended periods of time at high temperature."

[RasmusDC]

13 hours ago, NewMaxx said:

NAND/flash likes heat, not the controller or other components. And it only likes heat when programming/writing, otherwise it reduces data retention time. JEDEC rates consumer SSDs for 40C when programming with an idle temp of 30C. Although drives with a single sensor are reporting controller temperature which is only an estimate, however if he's regularly 60-70C idle with that sensor you can believe the NAND is also cooking which is most certainly not good.

 

Sources: JEDEC (pp.26-27) - Micheloni et. al. (2018). Inside Solid State Drives. 2nd ed. p. 222. Also Tiomkin: "data loss can occur if the NAND is stored or read over extended periods of time at high temperature."

that is just a "statement of high temperature" what is a high temperature? is it 60 degrees or is it 80? 

 

there is a lot of focus on data leak in the industry when storing "unpowered" drives, and that a 60 degree enviroment showed a faster data degradation than a 40 degree storage. but that is data degradation on a unpowered drive... 

 

“There is still a bit-rot factor for SSDs, but they have a lower bit-rot figure than tape: you have to sling a tape in the bin when it’s seven or ten years old. SSD is not as temperature-sensitive, it’s not vibration-sensitive in the way a hard drive is.” SSDs will start to include error correction to improve longevity, which will also help."

 

this is also a focus for data centers. and what is also seen is that in high temperature data enviroments, M.2 drives run at 60+ degrees daily in a long lifetime setup. does that mean that high temperatures are good, no.. but show me a "durability" graph of a large series setup with tempeture degredation, you need to get hot, before the durability slides a lot..

 

This issue however with prints, are that quick changeover from cold to hot and backwards, can have a negative impact on soldering, we saw in all electronics in mills, that if a mill in a cold enviroment, had to be restarted after 1 day of down time, we had to preheat everything, because the fast uncontrolled tempeture decrease could ruin electronics..

 

so in therory a drive going from 40-100 degrees in "stress cycles" fast, which some NVME does, could be harder on the solder joints and connections points, than one moving from 60-100 degrees. but this is just the "non memory" part of the board we are talking about...

 

i would not be nervous... however you will need to find some common ground on NVME, on having it cooled enough so it does not "throttle" at high loads.

[NewMaxx]

JEDEC measures endurance in a precise way via temperature with this equation. I'd link the full chapter but it's from a textbook. Even with their high point for a client SSD, 60C+ ("the maximum temperature ... for the client application class ... is 58C") for the NAND at idle would be exceedingly hot for reads long-term with 67C being the limit ("non-writing mode at a temperature not greater than 67C"). This is separate from their 10C-30C retention stress test. Bottom line: it's not good for NAND to be hot long-term with reads.

[Santiago Isra]

Hey! good news for all, i have the same ssd and my brother too. is a bug in life time measurement. maybe in the future xpg may correct this problem with a firmware update. Regarding the temperature it is not a problem, the normal operational temperature is 40 degrees and can vary up to 70 without problems.

 

[Andre_Resende]

Spoiler

 

On 3/1/2020 at 5:39 PM, Inyerbones said:

I have a XPG SPECTRIX S40G 512GB NVME PCIE 3x4 SSD : https://www.xpg.com/en/feature/610/

I've been using this SSD as my boot drive since August of 2019, so its relatively new

 

But I'm wondering If what I'm seeing in the screenshots attached below is worth worrying about? 

 

 

Few months ago I noticed it was around 70%, then it was 50% about 2 weeks ago, then I noticed it went down to 49%, followed by 48% and now 47% as days go by. 

Is my SSD simply just degrading that fast? 

Will it actually just die once it reaches 0%?

 

Thanks in advance.

 

Quote

Hello Friend,
I have an SSD of the same model, but in the 256 GB version, I live in Brazil and I have the same problem that you reported, after 520 hours of use and only 2 terabytes of writing on Cristaldiskinfo and also on Adata Toolbox which remains 72% of Estimated life, and according to the technical specification, the tbw of this model supports up to 160 terabytes, were you able to solve this problem? If so, could you guide me?
Thank you very much in advance.