SSD/NAND Operating Temps

[thilanliyan]

I was wondering if there is a recommended operating temp (not environment temp) for SSDs?

Apparently data retention is better at low temps, but write/erase is better at higher temps. Discussion came up because I added some extra heatsinks to a NVME SSD I am using in a small single board computer.

Before putting the heatsinks on it was blazing hot (couldn't put on my finger on it for more than 1/2 second). Now, with heatsinks and a fan blowing on it it runs relatively cool. (Will report back with actual temps)

Should I not be adding the active cooling?

 

[VirtualLarry]

I'm less than convinced that somehow, higher temps are better for an SSD.

I've used SATA 2.5" SSDs, in a passively-cooled NanoPC with a C-60 APU, and the SSDs I put in there, brand-new, eventually, after a month or so of straight running, would develop bad sectors, slowly but surely, until they "ate" the OS files and it wouldn't boot any more.

I became convinced that the failure mode was caused by overly high temps.

I've also "lost" (temporarily) some M.2 SSDs, that were underneath some dual-card mining configurations, that simply overheated and crashed.

That could have been more caused by the controller overheating, rather than the NAND, hard to say. But you can't really isolate the temps that well between them, in current SSD configs.

Thirdly, Sandisk promoted their "Data Vault", and in their marketing material, they showed the Ahelius (sp?) equation, which links temp with degradation of electronics and in this case, NAND.

At least according to that equation (which is part of Physics), higher temps DO cause data degradation.

So color me rather skeptical, when someone claims that SSDs "run better at higher temps". Not in my personal experience.

 

[UsandThem]

It's tough to say without knowing what drive you are talking about.

Some deal with thermal throttling, while others don't. In a program like CrystalDiskInfo, what does it report it's temps as? "Blazing hot" can be subjective from one person to the next.

Anyways, it's likely the controller is the hottest part, and not the NAND.

 

[thilanliyan]

So color me rather skeptical, when someone claims that SSDs "run better at higher temps". Not in my personal experience.

Here is an article on the subject:
https://www.eeweb.com/profile/eli-t...al-temperature-and-nand-flash-in-ssd-products
Basically says that data retention at low temps is better but actually write/erase is better at higher temps.

Now, this refers to the NAND only. Doesn't say anything about high temps on other aspects (controller, PCB, etc) of M2 SSDs.

This is what I'm trying to figure out...what is a reasonable operating temp when taking all aspects into account?
I'm using an ADATA SX8200 drive in Armbian Linux btw.

 

[UsandThem]

As many years as NAND has been in consumer parts, I think any concerns with NAND would have already been discovered, and widely talked about.

That article just doesn't do it for me personally. It says the author works for WD business marketing, and has 10 followers and only has two articles.

 

[thilanliyan]

As many years as NAND has been in consumer parts, I think any concerns with NAND would have already been discovered, and widely talked about.

I don't disagree. I was just wondering if there was a recommended operating temp range for SSDs?

Oh and I looked at the SMART attributes on my ADATA SX8200 drive and it says it is operating at 35C right now at load (this is with extra heatsinks and a fan blowing on it). Is that "too low"?

 

[UsandThem]

I don't disagree. I was just wondering if there was a recommended operating temp range for SSDs?

Oh and I looked at the SMART attributes on my ADATA SX8200 drive and it says it is operating at 35C right now at load (this is with extra heatsinks and a fan blowing on it). Is that "too low"?

Not in my opinion. That's what my 970 EVO idles at, and I have zero concerns on its idle or load temps. I don't know about Adata specs, but Samsung lists the operating temperature at 0 - 70c.

 

[Glaring_Mistake]

I'm less than convinced that somehow, higher temps are better for an SSD.

I've used SATA 2.5" SSDs, in a passively-cooled NanoPC with a C-60 APU, and the SSDs I put in there, brand-new, eventually, after a month or so of straight running, would develop bad sectors, slowly but surely, until they "ate" the OS files and it wouldn't boot any more.

I became convinced that the failure mode was caused by overly high temps.

I've also "lost" (temporarily) some M.2 SSDs, that were underneath some dual-card mining configurations, that simply overheated and crashed.

That could have been more caused by the controller overheating, rather than the NAND, hard to say. But you can't really isolate the temps that well between them, in current SSD configs.

Thirdly, Sandisk promoted their "Data Vault", and in their marketing material, they showed the Ahelius (sp?) equation, which links temp with degradation of electronics and in this case, NAND.

At least according to that equation (which is part of Physics), higher temps DO cause data degradation.

So color me rather skeptical, when someone claims that SSDs "run better at higher temps". Not in my personal experience.

Just what kind of temperatures did they reach in your NanoPC?
I've got a number in drives that can get pretty hot because they're in an enclosure where they don't get a lot of breathing room and so far none of those have failed *knock on wood*.

Believe you're thinking of the Arrhenius equation but I thought that was about when the drive is powered down, not when it's in use.