NVME Controller Board (?) Easily Gets too Hot

[Exodus6124]

Hello, I own a 12th gen Dell Precision laptop. I recently upgraded its storage with a 2 TB WD Blue SN580. I specifically picked this model despite it lacking a DRAM chip, as I was worried about overheating. According to its datasheet, this model barely passes the 5W mark at maximum load (and it was also among the cheapest in my country in terms of capacity/cost).
My laptop has two NVMe slots: one is close to the GPU (which comes with a copper heat sink provided by Dell in the slot) and the other has no mounting screws for a heat sink and is located right next to the fan exhaust. I chose to plug it into the one closer to the fan exhaust, as Dell manual specified that it gets less heat. Everything is fine when the fan is working, but for instance, right now I am simply browsing the web, so the CPU is below 50 degrees and the fan never spins up. Despite not doing anything involving IO, what I assume to be the NVMe controller board gets extremely hot, even at a low-power state. According to the other sensors in my laptop, none of the other components are above 50 degrees, so this is most certainly its own heat. I don't understand; I thought these NVMe drives did not need active cooling.
I don't know if it would be any different with Windows, as I don't have it installed, but I imagine it would be the same case without the fan spinning. Any suggestions? For reference, I'm attaching a photo of the laptop's internals that I found online.

$ sensors nvme-pci-0300
nvme-pci-0300
Adapter: PCI adapter
Composite:    +67.8°C  (low  = -40.1°C, high = +83.8°C)
                       (crit = +87.8°C)
Sensor 1:     +84.8°C  (low  = -273.1°C, high = +65261.8°C)
Sensor 2:     +67.8°C  (low  = -273.1°C, high = +65261.8°C)

# nvme smart-log /dev/nvme0
Temperature Sensor 1                    : 85 °C (358 K)
Temperature Sensor 2                    : 67 °C (340 K)

# nvme get-feature /dev/nvme0 -f 0x0c -H                                   
get-feature:0x0c (Autonomous Power State Transition), Current value:0x00000001
        Autonomous Power State Transition Enable (APSTE): Enabled
      
# nvme id-ctrl /dev/nvme0       
ps      0 : mp:4.80W operational enlat:0 exlat:0 rrt:0 rrl:0
            rwt:0 rwl:0 idle_power:0.3000W active_power:4.80W
            active_power_workload:80K 128KiB SW
ps      1 : mp:3.50W operational enlat:0 exlat:0 rrt:0 rrl:0
            rwt:0 rwl:0 idle_power:0.3000W active_power:3.00W
            active_power_workload:80K 128KiB SW
ps      2 : mp:2.40W operational enlat:0 exlat:0 rrt:0 rrl:0
            rwt:0 rwl:0 idle_power:0.3000W active_power:2.00W
            active_power_workload:80K 128KiB SW
ps      3 : mp:0.0150W non-operational enlat:1500 exlat:2500 rrt:3 rrl:3
            rwt:3 rwl:3 idle_power:0.0150W active_power:-
            active_power_workload:-
ps      4 : mp:0.0050W non-operational enlat:10000 exlat:6000 rrt:4 rrl:4
            rwt:4 rwl:4 idle_power:0.0050W active_power:-
            active_power_workload:-
ps      5 : mp:0.0033W non-operational enlat:176000 exlat:25000 rrt:5 rrl:5
            rwt:5 rwl:5 idle_power:0.0033W active_power:-
            active_power_workload:-

# nvme get-feature /dev/nvme0 -f 2 -H
get-feature:0x02 (Power Management), Current value:0x00000004
        Workload Hint (WH): 0 - No Workload
        Power State   (PS): 4

NVME temperature when fan is running at 1-2k RPM:

$ sensors nvme-pci-0300                                                                                                                                         
nvme-pci-0300
Adapter: PCI adapter
Composite:    +39.9°C  (low  = -40.1°C, high = +83.8°C)
                       (crit = +87.8°C)
Sensor 1:     +50.9°C  (low  = -273.1°C, high = +65261.8°C)
Sensor 2:     +39.9°C  (low  = -273.1°C, high = +65261.8°C)

 

[RJ]

It has been my observation that the m.2's without DRAM (Orico 7400) run hotter than my DRAM m.2's, (firecuda 540 & 990Pro). The full sized 2280's also have more surface area to dissipate heat compared to yours. Your options in a laptop are limited but the cheapest and easiest course of action would be to stick a thermal pad on it and see if that helps.

 

[Wirko]

Your options in a laptop are limited but the cheapest and easiest course of action would be to stick a thermal pad on it and see if that helps.

It's not clear if the thermal pad would touch any metal surface. It would be very good if a 2280-sized copper or aluminium plate could be placed on top of the pad and pressed down somehow.

 

[ShrimpBrime]

Pull the sticker and install a little heatsink. Should help lower the temp a tad.

Here's my 980pro for example. Seems to do the trick. I have it this way so I don't have to use the motherboard heatsink because I do a lot of hot swapping.

 

[RJ]

It's not clear if the thermal pad would touch any metal surface. It would be very good if a 2280-sized copper or aluminium plate could be placed on top of the pad and pressed down somehow.

Thank you for pointing this out, I meant a sticky, non e-conductive thermal pad, not a metal heatsink but the OP can't read my mind. In theory, he should be able to add a 1-2mm thick thermal pad and still have clearance to close the laptop and put the back cover on.

One of the sensors reads warm temps but not excessively hot. It's that sensor #2 that reads hot temps that likely throttle and might shorten the lifespan.
Hopefully a thermal pad can bring that down. OP, if you have to replace that m.2, consider a full length, 80mm aka 2280.

 

[Wirko]

Thank you for pointing this out, I meant a sticky, non e-conductive thermal pad, not a metal heatsink but the OP can't read my mind. In theory, he should be able to add a 1-2mm thick thermal pad and still have clearance to close the laptop and put the back cover on.

One of the sensors reads warm temps but not excessively hot. It's that sensor #2 that reads hot temps that likely throttle and might shorten the lifespan.
Hopefully a thermal pad can bring that down. OP, if you have to replace that m.2, consider a full length, 80mm aka 2280.

Both you and me misunderstood the situation and the pic. It shows the original 2230 Kioxia SSD, but that one was later replaced by a 2280 SN580. The SN580 doesn't exist in 2230 size!

But regardless, that SSD comes without as much as a graphene sticker. I'd try a sticker (graphene/copper/combination) first because it's designed to distribute the heat over the entire surface of the SSD, over a distance of a couple cm, so that the entire surface becomes a half-decent heatsink. On the other hand, a thermal pad's job is to move the heat away from the SSD, over a very short distance of 0.5 mm or so, to the metal object it touches. That's why a thermal pad with no heatsink on top wouldn't have much effect.

To the OP: I took a look at Aliexpress and you can get several M.2 heatsinks designed specifically for Precision models. Try to identify one that would fit your laptop. Just guessing here: it may be fastened at the right side with a screw, and at the left, by locking under the black plastic edge that the red arrow is pointing at, and next to the "M2L3" mark.

Also, have you consulted the user manual? It's unlikely that Dell designed the right M.2 slot without an option for a heatsink.

 

[Assimilator]

I thought these NVMe drives did not need active cooling.

They don't. The drive will throttle if it hits too high a temperature to prevent damage, in exactly the same way your CPU does. As you would have found out with 5 minutes of using a search engine.

 

[Exodus6124]

Thanks for all the replies.

They don't. The drive will throttle if it hits too high a temperature to prevent damage, in exactly the same way your CPU does. As you would have found out with 5 minutes of using a search engine.

Obviously, I am well aware of that. Let me rephrase, I thought these NVMe drives did not need active cooling to stay at reasonable temperatures and not reach the throttling threshold at idle.

Thank you for pointing this out, I meant a sticky, non e-conductive thermal pad, not a metal heatsink but the OP can't read my mind. In theory, he should be able to add a 1-2mm thick thermal pad and still have clearance to close the laptop and put the back cover on.

One of the sensors reads warm temps but not excessively hot. It's that sensor #2 that reads hot temps that likely throttle and might shorten the lifespan.
Hopefully a thermal pad can bring that down. OP, if you have to replace that m.2, consider a full length, 80mm aka 2280.

Both you and me misunderstood the situation and the pic. It shows the original 2230 Kioxia SSD, but that one was later replaced by a 2280 SN580. The SN580 doesn't exist in 2230 size!

But regardless, that SSD comes without as much as a graphene sticker. I'd try a sticker (graphene/copper/combination) first because it's designed to distribute the heat over the entire surface of the SSD, over a distance of a couple cm, so that the entire surface becomes a half-decent heatsink. On the other hand, a thermal pad's job is to move the heat away from the SSD, over a very short distance of 0.5 mm or so, to the metal object it touches. That's why a thermal pad with no heatsink on top wouldn't have much effect.

To the OP: I took a look at Aliexpress and you can get several M.2 heatsinks designed specifically for Precision models. Try to identify one that would fit your laptop. Just guessing here: it may be fastened at the right side with a screw, and at the left, by locking under the black plastic edge that the red arrow is pointing at, and next to the "M2L3" mark.

Also, have you consulted the user manual? It's unlikely that Dell designed the right M.2 slot without an option for a heatsink.

Thank you both for the suggestions. I rechecked the service manual and the SSD slots themselves. There doesn't seem to be a screw hole for the right SSD slot at all. I also checked Aliexpress and for my specific model it only has heatsinks for the left slot. The screw hole you mentioned is elevated a bit, so even if I managed to find something to reach there, it'd probably strain the back cover. While it's janky, I ended up ordering one of the thin copper heatsinks with rubber bands. I'll post the new temperature measurements when it arrives.

 

[_roman_]

#Post 1

It seems the air temperature in your laptop is around +67.8°C

---

I would not trust those sensors too much.

sys-apps/lm-sensors-3.6.2
sys-kernel/gentoo-sources-6.13.1

30 minutes or longer - pc in power off state - with wall disconnected with hardware switch

nvme-pci-0400
Adapter: PCI adapter
Composite:    +23.9°C  (low  = -20.1°C, high = +83.8°C)
                       (crit = +88.8°C)
Sensor 2:     +74.8°C

Sienna_Cichlid /home/roman # uptime
12:10:32 up 12 min, 0 user, load average: 0.25, 1.38, 2.16

I highly doubt the trustworthiness of sensor 2

Infrared thermometer says 37.8°C - uncalibrated but I used it a lot and I think it is trustworthy.

nvme-pci-0400
Adapter: PCI adapter
Composite:    +35.9°C  (low  = -20.1°C, high = +83.8°C)
                       (crit = +88.8°C)
Sensor 2:     +74.8°C

thanks for showing me that nvme command (sys-apps/nvme-cli-2.11::gentoo)

for some reason someone wrote temperature in one field and next field with Temperature

Sienna_Cichlid /home/roman # nvme smart-log /dev/nvme0 |grep Temperature|tail -1
Temperature Sensor 2            : 75 °C (348 K)
Sienna_Cichlid /home/roman # nvme smart-log /dev/nvme0 |grep temperature
temperature                : 37 °C (310 K)

According to the other sensors in my laptop, none of the other components are above 50 degrees, so this is most certainly its own heat. I don't understand; I thought these NVMe drives did not need active cooling.

You have passiv cooling without a cooler.

Nvme need a passive cooler. The newer one needs an active cooler.

-- your hardware is very hot inside.

The minimum temperature is from the air temperature around the nvme itself.

--

Any suggestions?

Check if your dell uefi allows any fan curves. Or if you can do it via software or hardware Anyway it is a dell and it is a laptop which gets hot inside. Hardware constraints.

 

[FreedomEclipse]

You can get super small heatsinks for that form factor

Its not going to drop the temps to sub zero but It will still do some good

 

[qxp]

Try using "smartctl". Here is what I get for my setup (Dell Inspiron 7573, also a notebook)

sudo smartctl -a /dev/nvme0n1
[...]
Model Number: Samsung SSD 970 EVO Plus 2TB
[...]
SMART/Health Information (NVMe Log 0x02)
Critical Warning: 0x00
Temperature: 36 Celsius
Available Spare: 100%
Available Spare Threshold: 10%
Percentage Used: 0%
[...]
Power Cycles: 230
Power On Hours: 3,654
Unsafe Shutdowns: 9
Media and Data Integrity Errors: 0
Error Information Log Entries: 203
Warning Comp. Temperature Time: 0
Critical Comp. Temperature Time: 0
Temperature Sensor 1: 36 Celsius
Temperature Sensor 2: 34 Celsius

This is the second version of the Samsung EVO plus - the first generation had issues with overheating in some situations, but Samsung fixed that and now claims the drive is "Temperature proof", whatever that means. The rated power dissipation is less than original.

Link below, you have to scroll down and expand to see specs:

SSD 970 EVO NVMe® M.2 2TB Memory & Storage - MZ-V7E2T0BW | Samsung US

Learn more about Samsung SSD 970 EVO NVMe® M.2 2TB with the innovative Samsung V-NAND technology and a 5-year limited warranty.

www.samsung.com

 

[Exodus6124]

Try using "smartctl". Here is what I get for my setup (Dell Inspiron 7573, also a notebook)

sudo smartctl -a /dev/nvme0n1
[...]
Model Number: Samsung SSD 970 EVO Plus 2TB
[...]
SMART/Health Information (NVMe Log 0x02)
Critical Warning: 0x00
Temperature: 36 Celsius
Available Spare: 100%
Available Spare Threshold: 10%
Percentage Used: 0%
[...]
Power Cycles: 230
Power On Hours: 3,654
Unsafe Shutdowns: 9
Media and Data Integrity Errors: 0
Error Information Log Entries: 203
Warning Comp. Temperature Time: 0
Critical Comp. Temperature Time: 0
Temperature Sensor 1: 36 Celsius
Temperature Sensor 2: 34 Celsius

This is the second version of the Samsung EVO plus - the first generation had issues with overheating in some situations, but Samsung fixed that and now claims the drive is "Temperature proof", whatever that means. The rated power dissipation is less than original.

Link below, you have to scroll down and expand to see specs:

SSD 970 EVO NVMe® M.2 2TB Memory & Storage - MZ-V7E2T0BW | Samsung US

Learn more about Samsung SSD 970 EVO NVMe® M.2 2TB with the innovative Samsung V-NAND technology and a 5-year limited warranty.

www.samsung.com

That's an excellent temperature. I was initially considering your model, but in the end I decided not to due to the complaints regarding overheating regarding the first generation (+ max power consumption is listed as 9W at peak).

#Post 1

It seems the air temperature in your laptop is around +67.8°C

---

I would not trust those sensors too much.

sys-apps/lm-sensors-3.6.2
sys-kernel/gentoo-sources-6.13.1

30 minutes or longer - pc in power off state - with wall disconnected with hardware switch

nvme-pci-0400
Adapter: PCI adapter
Composite:    +23.9°C  (low  = -20.1°C, high = +83.8°C)
                       (crit = +88.8°C)
Sensor 2:     +74.8°C

Sienna_Cichlid /home/roman # uptime
12:10:32 up 12 min, 0 user, load average: 0.25, 1.38, 2.16

I highly doubt the trustworthiness of sensor 2

Infrared thermometer says 37.8°C - uncalibrated but I used it a lot and I think it is trustworthy.

nvme-pci-0400
Adapter: PCI adapter
Composite:    +35.9°C  (low  = -20.1°C, high = +83.8°C)
                       (crit = +88.8°C)
Sensor 2:     +74.8°C

thanks for showing me that nvme command (sys-apps/nvme-cli-2.11::gentoo)

for some reason someone wrote temperature in one field and next field with Temperature

Sienna_Cichlid /home/roman # nvme smart-log /dev/nvme0 |grep Temperature|tail -1
Temperature Sensor 2            : 75 °C (348 K)
Sienna_Cichlid /home/roman # nvme smart-log /dev/nvme0 |grep temperature
temperature                : 37 °C (310 K)

You have passiv cooling without a cooler.

Nvme need a passive cooler. The newer one needs an active cooler.

-- your hardware is very hot inside.

The minimum temperature is from the air temperature around the nvme itself.

--



Check if your dell uefi allows any fan curves. Or if you can do it via software or hardware Anyway it is a dell and it is a laptop which gets hot inside. Hardware constraints.

My laptop also has a sensor named ambient sensor which is for instance 35 degrees right now. I'm assuming that it's measuring the internal air temperature of the device. So I suppose then it'd makes sense that the NVMe can get quite hot.

Thank you all for the replies once again. While my NVMe is still hot, at least with a copper plate heatsink the high temperature sensor only reaches 80 degrees at load which is still within operating temperatures (0-85 degrees).