M.2 SSD Overheating

Ian

Ian

Administrator
Joined
Feb 23, 2002
Messages
19,878
Reaction score
1,509
I've recently purchased a pair of Crucial MX300 M.2 SSDs to use in a new QNAP NAS (with the aim of keeping constant access files on these drives, to minimise drive noise). However, I've noticed that they run surprisingly hot, even under relatively modest drive access.

I'm sure part of the problem is down to the minimal space inside a NAS, however these M.2 SSDs are designed to be used in much more enclosed spaces than this, so that can't be the only issue. The overheating warning appears when the drives are >65C, after which there is some performance management that kicks in to keep the temperature until control. It's not like I'm hammering the drives for hours at a time, I simply performed a relatively modest drive snapshot that should take a few minutes.

In the end, I bought a load of 10mmx10mm heatsinks and thermal pads - so I've covered the chips in those :D. I'm going to monitor how the temperatures change over the coming days. So far it appears to have reduced the mean temperature from 47C to 38C - which is a decent drop.
 
So, my question is ... how are you measuring the temps?

I ask 'cos from what I've read, that's quite normal for that chipset. :)
 
I'm just using the onboard S.M.A.R.T diagnostics, so a reading from the chip itself. The max operating temperature is 70C, which I get really close to under load (68C this morning).

I'm happy with the normal operating temps of ~47C, but the peaks are too high - I suspect the thermal protection kicks in so that I don't exceed 70C.
 
I have a D-Link DNS 320L with 2 x WD drives in it. When I first set it up the drives were in RAID and when I moved large files like movies you could feel the heat coming off the box. I replaced one drive and now they are set up as 2 x single drives and the box doesn't get near as hot as it used to moving large files.
 
Cheers Evan, I've not got these drives as part of the RAID array (at the moment), but I've been doing more digging and it looks like the overheating is particularly bad when TRIM commands are run. Tomorrow morning I should know if the heatsinks have worked, as I'll have enough data :).
 
you are not manually invoking TRIM are you?

Win 10, despite some FUD, will do a good job left alone. :)
 
muckshifter said:
you are not manually invoking TRIM are you?

Win 10, despite some FUD, will do a good job left alone. :)
Click to expand...
Oh no, they're in a NAS running a custom Linux build from QNAP - there's an overnight garbage collection facility that runs :).
 
OK, turn off the background garbage collection and use 'foreground' collection instead ... you shouldn't suffer any performance issues.

Also note: Today’s TRIM doesn’t work in most RAID environments because current RAID drivers generally don’t yet support it.


:user:
 
As the NAS OS/Drivers are heavily customised, I've just left TRIM enabled as they've recommended that for their latest release - I guess however they've written their RAID implementation they must take it in to account somehow. Since adding the headsinks and watching the log cleanup process overnight, the problem has stopped - temps didn't cross 60C at any point :D.
 
Back
Top