Nand, temp vs endurance. Legitimate sources?

[MindArlekin]

I wonder, are there any sources from major manufacturers of NAND flash, which include clear definitions of temperature to endurance ratio.

Because in discussion about this topic i failed to prove my point of view on the questionable benefit of M.2 heatsinks.

In short, it looks like something like:
- "If it will be some document or something from Micron/Samsung/Toshiba - i will trust it. Otherwise, GTFO".

P.S.
I understand purpose behind M.2 heatsinks, i just think, that implementation (attaching heatsink flat both on a controller and nand chips) is kinda bad. You get an bonus of better temps on controller, but at a same time you loosing in endurance of your drive, because you lowering temperatures on nand chips.

 

This is what I referred to during the discussion:

Some whitepaper:

https://cdn.selinc.com/assets/Literature/Publications/White Papers/0015_NANDflash_IO_20141211.pdf?v=20170218-001047

Gamers Nexus:

https://www.youtube.com/watch?v=KzSIfxHppPY

 

Thank you in advance.

 

[Ryujin2003]

I can ask my buddy to test his, but both of us have the same MB, CPU, and m.2 combo: R7 1700x, MSI X370 Gaming Pro, and Samsung 960 Evo 500GB,

 

I didn't put a heatsink on mine. It was when Ryzen first launched, and everyone was complaining about the marketing of these heat shields and how it would absorb the heat from the GPU and keep the drive at a nice thermal throttle. He put it on because he thought it looked cool and figured it'd work.

 

I don't have the data in front of me, but I'm rocking CF RX 480's, and he's got dual 1070's. So there's extra heat from the GPU's right there. But our performance (from what I remember) was the same regardless of having the stock MSI heatsink vs not.

 

The airflow we have is pretty decent, and we both have fairly aggressive fan curves on our GPUs, so that may play a part of it as well.

 

Might be a thing on the 970 since this thing runs much hotter than the 960. Might be worth asking Gamers Nexus to actually do a test on this?

[Alex Atkin UK]

The thing is, NAND stores data better when running hot but the controllers run better when running cool, so its quite the balancing act.

[MindArlekin]

1 minute ago, Alex Atkin UK said:

The thing is, NAND stores data better when running hot but the controllers run better when running cool, so its quite the balancing act.

This could be balancing act, if heating/cooling both controller and memory would influence on a same thing, speed for example. But in reality cooling/heating a controller makes your drive faster/slower and cooling/heating nand chips makes lifespan of your drive shorter/longer.

 

I definitely understand, that, if we look at thermal performance of those heatsinks, they don't make a significant difference, so gains or damage is also minimal, if they even exist.

So we have situation, when by installing this thing on your drive it at best don't make enough difference to be bothered with, or harms your drive lifespan  in an unknown degree.

So, one and only thing, where those heatsinks make sense is in visuals.

 

I have an idea, but I do not have the resources to test it for now.
What if we take an heatpipe, from, for example, old cpu tower cooler, with part of it base and put it directly on controller chip. Those heatpipes quite long, so it will be possible to mount it some way to cool it with airflow from cpu or gpu cooler, without even possibility to harm nand chips.
This will be viable and useful thing. And, of course, it will never raise the question of improving one thing at the expense of another