M.2 NVME external enclosure hot even at idle

DWD1961

[H]ard|Gawd
Joined
Nov 30, 2019
Messages
1,314
I have a Samsung 970 Evo in an aluminum USB enclosure, and it seconds as my hand warmer, even at idle. Anyone know why that is? It seems to me that without the need for power, it would be cool.
 
Well anything that contains a microprocessor generates heat and the faster the device the greater the delta t.
 
Not a personalized answer but I can tell you my internal nvme idles at 41c-50c/105f-122f (usually on the upper end, this is the first time I've seen it at 41c). And that's with one of those heatsinks -- whose effectiveness is unfortunately questionable according to people online
I hope that's normal!

Well anything that contains a microprocessor generates heat and the faster the device the greater the delta t.
Fascinating. So your phone contains fewer chips than a nvme since it doesn't act as a handwarmer?
What's even the point of your answer? We know electronics generate heat. OP is asking if it's normal to be that hot or not.
 
Last edited by a moderator:
Not a personalized answer but I can tell you my internal nvme idles at 41c-50c/105f-122f (usually on the upper end, this is the first time I've seen it at 41c). And that's with one of those heatsinks -- whose effectiveness is unfortunately questionable according to people online
I hope that's normal!


Fascinating. So your phone contains fewer chips than a nvme since it doesn't act as a handwarmer?
What's even the point of your answer? We know electronics generate heat. OP is asking if it's normal to be that hot or not.
Yeah it's idling and warm as hell. However, I think what learns permit meant is that even when a CPU idles, it's pretty damn warm. Like you said, your M2 idles at 115F. That's enough to warm up an external enclosure.

BTW, yeah, a heat sink will work really well. But there is a trick to it. I have two LED / M2 internal coolers, plus my stock Gigabyte Aorus MB cooler. The LED /Coolers both cool better than the stock cooler, but one of them initially did not. I had to change the heat pad in order to get it to cool effectively.

Make sure you read up on thermal conductivity of heat pads because they are not all equal--not in the least. There are actually some out there that have a higher thermal conductivity than paste, much, much higher. Also. make sure you get enough pad in-between the bottom and top of the drive in relation to the heat sink itself.

You are actually suppose to take the sticker off of the drive before applying the pad, but I didn't do that. You don't want to do that on a Samsung drive because the bottom sticker really isn't a sticker., It's a stick on copper heat sink from the factory. So leave that.

Test parameters: Crystal Disk Mark set to "Peak Performance" for 4 runs on SEQ1M/Q8T1. This was measures with HWiNFO and no one seems to know what the two numbers are. I also measured heat weith Crystal Disk Info and Samsung Magician (All fo the lower numbers reported the same). The best information I can find, regarding HWiNFO's two temp numbers is that one is the controller. When compared to the single temp read out in Samsung Magician, the lower number is given. I deduce from that that the lower number is the drive temp, and if not, in any case, the lower number is what matters. I'd like to get some realy nice pads and try those.

Stock cooler
Max Temp
52C / 73C

Asia Horse cooler
50 / 68C

EZ cooler
59/74C -- with included thermal pads
52/67C -- With the Asia Horse thermal pads, becasue I don't have any others.

I also set the run for 10 runs and oh boy does that MFer get hot. The specs says the max operating temp is 70C, but it gets WAY over that if you hammer it with ATTO or Crystal Peak Performance with multiple runs. Last, from what I have recently read, you don't have to worry about the drive actually overheating and harming itself. It's self limiting. Even when operating at 73C+, I haven't noticed any throttling, so I have no idea what Samsung has going on there.

Conclusion
If you really want to cool your M2 internal, yeah, do get a cooler. Get a nice one, not something like I have, which is more for looks. Then use a really good thermal pad, and make sure it actually fits tight onto the thermal pad. I'd also recommend a cooler with screw down not screw side, since you can aply m,ore pressure to teh drive and pads with the screw down type, opposed to the side screw type.
 
They can safely get up to 90c iirc (almost 200f, about 20°f shy of boiling water). Samsung says 70c max, which is still hot enough to cook food on. If anything, you may want to avoid placing them on finished wood and some plastics, but the drive should be fine.
 
They can safely get up to 90c iirc (almost 200f, about 20°f shy of boiling water). Samsung says 70c max, which is still hot enough to cook food on. If anything, you may want to avoid placing them on finished wood and some plastics, but the drive should be fine.
They get a lot hotter than that and that's why I don't understand the 70C Samsung temperature. They sure don't throttle at 70C.
 
anything over 46c is going to feel hot, thats where its idling so....
 
I'm not surprised it runs hot. The last build I completed actually had a thermal pad and heatsink for the M2 drive. Pretty slick
 
So here is some information:

This is the external enclosure that gets really warm, even when idling.
AMPCOM

I just received this external enclosure from Sabrent, which is a name brand I've used several times in the past years. I've always had good luck with them:
Sabrent

The Sabrent enclosure runs cool, cool, cool, even when hammering it with Crystal Disk Peak performance for 9 runs. I think the first drive is just running a cheap ass, fuck you shit controller. The controller is what is heating it up.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
For data points, I have a gen 4 and a gen 3 in my box. The gen 4 has the mobo heat sink on it, the gen 3 has a think copper heat sink held on via o-rings. They idle at 34C and 31C respectively. The gen 4 is the boot drive. (both have thermal pads though)

I don't think they will ever not be warm. The question is how warm yours is getting. "warm" isn't a very specific term.
 
For data points, I have a gen 4 and a gen 3 in my box. The gen 4 has the mobo heat sink on it, the gen 3 has a think copper heat sink held on via o-rings. They idle at 34C and 31C respectively. The gen 4 is the boot drive. (both have thermal pads though)

I don't think they will ever not be warm. The question is how warm yours is getting. "warm" isn't a very specific term.
I can't remember what I did or if I failed to find a way to show temp on an external drive. The Sabrent runs really cool. I think it was the cheaper enclosure's controller that is getting hot, not the drive. BTW, using a high thermal conductivity pad will really help. In fact, pads can far exceed pastes. Arctic MX-4 has a thermal conductivity of 8.5 W/(mK). Pads can go over 60. Or, you can get a cheaper pad that equals paste. They do differ. I have a couple different LED M2 internal coolers, and one of them cam with pads that were so poor it was like a heating blanket over the drive.
 
I can't remember what I did or if I failed to find a way to show temp on an external drive. The Sabrent runs really cool. I think it was the cheaper enclosure's controller that is getting hot, not the drive. BTW, using a high thermal conductivity pad will really help. In fact, pads can far exceed pastes. Arctic MX-4 has a thermal conductivity of 8.5 W/(mK). Pads can go over 60. Or, you can get a cheaper pad that equals paste. They do differ. I have a couple different LED M2 internal coolers, and one of them cam with pads that were so poor it was like a heating blanket over the drive.

Where are you seeing thermal pads that are blowing way paste? I know that thermal grizzly claims their pads are 65w/mk, but I have yet to see any tests show it actually beating paste. Even Puget said they got the same results as with paste and only so much reuse out of it.
 
Where are you seeing thermal pads that are blowing way paste?
Where are you seeing that i said thermal pads blow away thermal paste?

The thermal conductivity of pads can far exceed most pastes, but for CPUs, they do not perform as well (although they can come close for the higher end pads). I would still use paste for my CPU. I don't know about GPUs. I do know that you can't use thermal paste on a internal M2 cooler. It's not a tight enough fit. So, you're stuck with nothing or pads. The Sabrent enclosure I wasn't even using thermal pads that came with it. I just installed it. The other enclosure was getting toasty to the touch at idle. The Sabrent is cool to the touch, working the drive, or idling. Who knows?
 
Back
Top