They don't fill a specific niche, they cool down my Evo 970's NAND down to 50 degrees C under sustained load, and the controller to 65 max (from 82!). Given it costs under 15 EUR for end-consumers, (not exactly these, but same stuff from Aquatuning) that should be attached out of the box.
They heat themselves up to operating temperature by design, cooling them means they just continually dump out power trying to hit temperature, and will be operating with a lower endurance (simplified: higher operating temperature = lower energy input to set/erase cells = less degradation of each cell per write/erase cycle).
As long as this doesn't catch significant airflow, it might actually help a little by spreading heat from the controller / active dies to inactive dies.
If you're going to be running lots of benchmarks trying to hit the highest speeds possible, you will have to cool both the NAND and the controller. If you only care about performance during real-world usage, then you don't need to cool either (except for Optane 905P and arguably Phison E16 drives).
m.2 seemed like a great idea at first. The promise of PCIE speeds in a small form factor is certainly appealing. However, 2 things that weren't accounted for are cooling and physical space for the flash modules themselves. I expect that 4-8 tb will be the most we'll see with the current technology. You can't even effectly cool the drive, since the underside is almost always inaccessible. The heatsinks do help, but unless we figure out a better way to do things we are kind of stuck.
m.2 is great, just nobody thought about colling at the time. I've seen mobo with shields, both under and over the ssd. I guess with thick thermal pad it will work
The 'need' for cooling m.2 drives is massively overblown. Even in benchmark testing, it takes several tens to hundreds of GB of continuous writing for a controller to start throttling.
If you're buying them for cosmetics like RAMsinks then knock yourself out, but in terms of function they are about as useful as ramsinks.
The cooling and PCB space limitations are a non-issue because they only come into play when you're far past the practical cost limitations, or if you're just trying to get the highest benchmark scores without regard to real-world use cases. M.2 is fine for consumer systems, and the form factor is certainly not holding back the progress of consumer storage tech in any meaningful way.
They don't need to make a huge difference - 10 degrees is enough to stop most of these drives from throttling, and 15-20 should have a beneficial effect on service life.
Copper would change that by around 5 degrees at most.
interesting, I have similar solution from EKWB which looks to have better mounting solution but it's also bit more expensive I wonder how they compare.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
16 Comments
Back to Article
bENDEJO - Tuesday, December 3, 2019 - link
What, no LEDs"drexnx - Tuesday, December 3, 2019 - link
silicon fixtures aka rubber bands?hey, if it works, it works.
deil - Wednesday, December 4, 2019 - link
better than expected AND maybe it will fit into laptop?RoC_17 - Tuesday, December 3, 2019 - link
They don't fill a specific niche, they cool down my Evo 970's NAND down to 50 degrees C under sustained load, and the controller to 65 max (from 82!). Given it costs under 15 EUR for end-consumers, (not exactly these, but same stuff from Aquatuning) that should be attached out of the box.edzieba - Wednesday, December 4, 2019 - link
Don't cool the NAND dies themselves!They heat themselves up to operating temperature by design, cooling them means they just continually dump out power trying to hit temperature, and will be operating with a lower endurance (simplified: higher operating temperature = lower energy input to set/erase cells = less degradation of each cell per write/erase cycle).
Spunjji - Wednesday, December 4, 2019 - link
As long as this doesn't catch significant airflow, it might actually help a little by spreading heat from the controller / active dies to inactive dies.Billy Tallis - Thursday, December 5, 2019 - link
If you're going to be running lots of benchmarks trying to hit the highest speeds possible, you will have to cool both the NAND and the controller. If you only care about performance during real-world usage, then you don't need to cool either (except for Optane 905P and arguably Phison E16 drives).Thud2 - Tuesday, December 3, 2019 - link
Do you mean Silicone Fixtures?eek2121 - Wednesday, December 4, 2019 - link
m.2 seemed like a great idea at first. The promise of PCIE speeds in a small form factor is certainly appealing. However, 2 things that weren't accounted for are cooling and physical space for the flash modules themselves. I expect that 4-8 tb will be the most we'll see with the current technology. You can't even effectly cool the drive, since the underside is almost always inaccessible. The heatsinks do help, but unless we figure out a better way to do things we are kind of stuck.deil - Wednesday, December 4, 2019 - link
m.2 is great, just nobody thought about colling at the time. I've seen mobo with shields, both under and over the ssd. I guess with thick thermal pad it will workyetanotherhuman - Wednesday, December 4, 2019 - link
Maybe we should have gone with 1.8" SSDs instead, with a 12Gbps SAS connection. I vote for that. All aluminium housing with thermal pads..edzieba - Wednesday, December 4, 2019 - link
The 'need' for cooling m.2 drives is massively overblown. Even in benchmark testing, it takes several tens to hundreds of GB of continuous writing for a controller to start throttling.If you're buying them for cosmetics like RAMsinks then knock yourself out, but in terms of function they are about as useful as ramsinks.
Billy Tallis - Wednesday, December 4, 2019 - link
The cooling and PCB space limitations are a non-issue because they only come into play when you're far past the practical cost limitations, or if you're just trying to get the highest benchmark scores without regard to real-world use cases. M.2 is fine for consumer systems, and the form factor is certainly not holding back the progress of consumer storage tech in any meaningful way.Soulkeeper - Wednesday, December 4, 2019 - link
aluminum passive heatsinks make next to no difference ...Unless it's copper with some very thin fins, I don't see the point.
Spunjji - Wednesday, December 4, 2019 - link
They don't need to make a huge difference - 10 degrees is enough to stop most of these drives from throttling, and 15-20 should have a beneficial effect on service life.Copper would change that by around 5 degrees at most.
Eliadbu - Monday, December 9, 2019 - link
interesting, I have similar solution from EKWB which looks to have better mounting solution but it's also bit more expensive I wonder how they compare.