MinisForum MS-01 : heating problem
- Thread starter caplam
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
In case anyone else comes across this post about thermals and the MS-01. Here are how my two MS-01 changed with the PTM7950 Pads from LTT. (Blue Line)
*These sit in my garage.
Proxmox1:
Runs Frigate with M2 Coral on it with 2 Cameras
2 Dockers
Minecraft Server(no one really using it)
Pfsense and Transmission (seedbox)
Proxmox2:
2 Dockers
Home Assistant with USB Pass through
Code Server
Minecraft Server 2
Proxmox 1 had way too much thermal paste on it, Proxmox 2 had a normal amount before the swap.
*These sit in my garage.
Proxmox1:
Runs Frigate with M2 Coral on it with 2 Cameras
2 Dockers
Minecraft Server(no one really using it)
Pfsense and Transmission (seedbox)
Proxmox2:
2 Dockers
Home Assistant with USB Pass through
Code Server
Minecraft Server 2
Proxmox 1 had way too much thermal paste on it, Proxmox 2 had a normal amount before the swap.
Hi all,
I've followed this thread for some time and now caught back up with it, read through the whole thread over again.
I own a MS-01 for about 4 months now, running Unraid on it migrated from my bigger AMD 5950X box and have been getting consistently HIGH temps, in the range of 90c+ and recently over 100c:
It's been hot but not been a problem until recently where the MS-01 unit has started crashing (Last few weeks, increasing in frequency now daily), and I'm considering replacing it under warranty as this seems unrasonable, the motherboard temps are consistently at 115c+ and CPU is ~95c sometimes 100c+. I've temporarily removed the outer case to aleviate the airflow, but this isn't a sustainable solution.
Temps with the casing off aren't much better (Fluctuates between ~75-100c+ with case off)-:
In my system I am running 3 NVME disks:
- Samsung 980 Pro 2TB with Heatsink - temps between 40-70c
- Samsung 980 Pro 1TB (without heatsink) - temps between 40-60c
- Samsung 970 Evo Plus 1TB (without heatsink) - temps between 40-60c
The PCIe slot is unpopulated currently.
The system runs 24/7 and at the moment is about 20-24% CPU usage.
I am considering if I should replace the (Likely badly applied) heat sink compound with LM or a Graphene/PhaseChange thermal pad, but I think I prefer the maintenance-free potential of the thermal pad over LM...
I'm seriously considering purchasing the Thermal Grizzly KryoSheet or Thermal Grizzly PhaseSheet PTM, but I understand the KryoSheet seems more attractive as I will never need to touch it again.
I'm looking for some advice as to if I should go ahead and replace the unit first in case of damage or if I should just go with the KryoSheet and run with it as it is now?
I've followed this thread for some time and now caught back up with it, read through the whole thread over again.
I own a MS-01 for about 4 months now, running Unraid on it migrated from my bigger AMD 5950X box and have been getting consistently HIGH temps, in the range of 90c+ and recently over 100c:
It's been hot but not been a problem until recently where the MS-01 unit has started crashing (Last few weeks, increasing in frequency now daily), and I'm considering replacing it under warranty as this seems unrasonable, the motherboard temps are consistently at 115c+ and CPU is ~95c sometimes 100c+. I've temporarily removed the outer case to aleviate the airflow, but this isn't a sustainable solution.
Temps with the casing off aren't much better (Fluctuates between ~75-100c+ with case off)-:
In my system I am running 3 NVME disks:
- Samsung 980 Pro 2TB with Heatsink - temps between 40-70c
- Samsung 980 Pro 1TB (without heatsink) - temps between 40-60c
- Samsung 970 Evo Plus 1TB (without heatsink) - temps between 40-60c
The PCIe slot is unpopulated currently.
The system runs 24/7 and at the moment is about 20-24% CPU usage.
I am considering if I should replace the (Likely badly applied) heat sink compound with LM or a Graphene/PhaseChange thermal pad, but I think I prefer the maintenance-free potential of the thermal pad over LM...
I'm seriously considering purchasing the Thermal Grizzly KryoSheet or Thermal Grizzly PhaseSheet PTM, but I understand the KryoSheet seems more attractive as I will never need to touch it again.
I'm looking for some advice as to if I should go ahead and replace the unit first in case of damage or if I should just go with the KryoSheet and run with it as it is now?
Last edited:
My MS-01 dropped temps by an average of 20c just by pulling the bower fan off and re-pasting using PTM7950. Shocking that it helped that much. The factory paste they use is just crap, crappily appied.
It still runs warm but never quite touches 90c under heavy synthetic stress tests and happily runs up to about 75c with my normal loads.
It still runs warm but never quite touches 90c under heavy synthetic stress tests and happily runs up to about 75c with my normal loads.
Thanks, it would seem I might have a particularly egregiously bad thermal paste application out of the factory.
The temps I'm getting are frankly ridiculous, and appear to be worsening over the 4 months I've had it.
Any advice for which product I should use?
Which PTM9750 are you using? The LTT version?
Is the Thermal Grizzly PhaseSheet PTM comparable?
The temps I'm getting are frankly ridiculous, and appear to be worsening over the 4 months I've had it.
Any advice for which product I should use?
Which PTM9750 are you using? The LTT version?
Is the Thermal Grizzly PhaseSheet PTM comparable?
I used a rather generic PTM9750 pad from Amazon. The factory job is so bad you really can’t do it worse no matter what you use.
I've been looking around for Honeywell PTM9750 for a few days, but it is understandably expensive but more importantly will take nearly a month to ship to me, and I really need to fix this.
Conversely the Thermal Grisley PhaseSheet PTM and KryoSheet I can get next day.
I think I'll order both of the Thermal Grisley as its cheap enough, and take a chance on the PhaseSheet PTM...
Edit: In the end I ordered the Thermal Grisleys from Amazon, and the PTM9750 as well from MODDIY and if the Thermal Grisley PTM doesn't work out, I have only lost 20 schmeckles and will wait a few weeks. I'll also have spare for my GPU and other systems if need be I guess. Probably won't hurt to have some PTM9750 spare.
Does anyone know the die size of the 13900-H in the MS-01?
Conversely the Thermal Grisley PhaseSheet PTM and KryoSheet I can get next day.
I think I'll order both of the Thermal Grisley as its cheap enough, and take a chance on the PhaseSheet PTM...
Edit: In the end I ordered the Thermal Grisleys from Amazon, and the PTM9750 as well from MODDIY and if the Thermal Grisley PTM doesn't work out, I have only lost 20 schmeckles and will wait a few weeks. I'll also have spare for my GPU and other systems if need be I guess. Probably won't hurt to have some PTM9750 spare.
Does anyone know the die size of the 13900-H in the MS-01?
Last edited:
Well... the MS-01 crashed TWICE again this morning with no warning, and Unraid has already started another Parity-Check and found more errors... The last 2 Parity-Check after daily crashes also did not completed, so I'm getting into a loop here. This is ridiculous.
So I contacted Amazon (Thankfully I purchased through Amazon) for replacement and they have helpfully processed a return and refund for me.
Apparently Minisforum are a "3rd party" on Amazon so they said there's nothing else to do but issue refund. So I'm going to print a shipping label and get refund within 21 days, once they receive it.
For the record, I emailed Minisforum about the issues at the beginning of this week explaining everything and quoting unit serial number and they still have not even acknowledged or responded at all.
I've also noticed the MS-01 is further discounted on Amazon, a good deal or bad omen?
That being said I REALLY like this system, I think there's nothing like it, it fits my use case and my Thermal Grisley (Amazon also) AND PTM7950 (MODDIY) is on its way, so that being said I have a few options...
Should I:
- Just repurchase another MS-01 and apply the proper cooling IMMEDIATELY before use?
- Find another comparable system, is there even one? When I looked around originally, there isn't much that really compares to the MS-01.
- Refund EVERYTHING when they arrive tomorrow and give up?
Looking for some honest advice.
Edit: Created a new thread to discuss MS-01 alternatives, but I'd still appreciate ANY advice...
So I contacted Amazon (Thankfully I purchased through Amazon) for replacement and they have helpfully processed a return and refund for me.
Apparently Minisforum are a "3rd party" on Amazon so they said there's nothing else to do but issue refund. So I'm going to print a shipping label and get refund within 21 days, once they receive it.
For the record, I emailed Minisforum about the issues at the beginning of this week explaining everything and quoting unit serial number and they still have not even acknowledged or responded at all.
I've also noticed the MS-01 is further discounted on Amazon, a good deal or bad omen?
That being said I REALLY like this system, I think there's nothing like it, it fits my use case and my Thermal Grisley (Amazon also) AND PTM7950 (MODDIY) is on its way, so that being said I have a few options...
Should I:
- Just repurchase another MS-01 and apply the proper cooling IMMEDIATELY before use?
- Find another comparable system, is there even one? When I looked around originally, there isn't much that really compares to the MS-01.
- Refund EVERYTHING when they arrive tomorrow and give up?
Looking for some honest advice.
Edit: Created a new thread to discuss MS-01 alternatives, but I'd still appreciate ANY advice...
Last edited:
I use Unraid as well so I can share some of the things I have learned. Use PTM9750 don't mess with anything else. It will drop temps considerably without the risks of LM. Use powersave for scaling_governor I have noticed big temp swings in performance.
You can put this in user scripts:
echo powersave | tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor > null
If you are using Unraid 7 beta it has crashing issues you may want to go back to stable. Take the cover off MS-01 and put a fan on it just for testing see if it still crashes. If not you know its cooling and using powersave along with PTM9750 should fix things. Also putting MS-01 somewhere that has good airflow. If it still crashes even with lid off and fan you know its an Unraid issue. You can mirror syslog to flash for troubleshooting and join Unraid Discord or go to the forums and post your diagnostics. The guys on there are super smart. I have had VM's and Dockers crash my server sporadically over the years.
One thing I noticed from an earlier post is your cpu fan should be full blast at those temps, mine runs at 2400 idle.
Try this fan curve:
Just take one thing at a time.
Last edited:
Yes I saw this as well, but I saw there is no mention of the measured size of the die.
I'm still wondering, I may have to do it myself.
My PTM7950 is on order from MODDIY at this moment, and I anticipate it will arrive next week.
I also have some TGP8000PT pads on order for the VRMs, but I'll see how that is when it arrives.
I applied the powersave user script, and I'll see how it goes thanks!
I think can already see some minor drops, but will monitor it.
I'm not using Unraid 7 or the beta, tbh I stay away from beta releases on my main Unraid box... and currently this is my only Unraid.
I'm using the same stable 6.12.10 version I've been using without issue for a long time, migrated from my old stable setup, I know its stable.
I'm very sure its not Unraid causing the instability, I have quite a few things running, but my Unraid is well tuned and never had this instability.
The box is always been in a well ventilated, open area with a large fan right next to it, but not blowing directly on it. The air is circulated well.
I know the envionment is stable temp, the room maintains about 22-24c consistently, in the middle of summer. I'm sure its not the environment.
Currently the MS-01 has the case off, ever since the crashing increased. The MS-01 has been stable for about 2 days and 3 hours now, as I'm writing this.
I haven't adjusted the fan curves, but I will check it.
I wish the Intel vPro worked more reliably though, as I would be able to adjust that right away.
I'll shortly check the fan curve as you advised and adjust it to your settings.
I'll update once done that.
@wadup if you don't mind my asking, I'm really wondering if you think I should return this MS-01 for a replacement (As I have the refund filed with Amazon already, but I have until end of next week to ship it) and oder a new kit next week, or keep this one and apply the PTM7950 to this?
If you think something is wrong replace it, nothing hurt by using your warranty. If it were me I would troubleshoot before I return it because it could take awhile to get a new one.
Hey thanks, I decided to order a replacement while it was at a good sale price on Amazon, and just arrived today.
Thankfully, I ordered through Amazon which is cheaper (739 euros vs 779 on their own site), faster (Next day with Prime) and I think better support (Amazon processed my refund request even after 4 months). I'd highly recommend using Amazon for this machine, not to big them up but their support network is so much better.
Anyway, the PTM7950 and TGP8000PT from MODDIY is arriving this week, so I'll be ready to set install it immediately and migrate to that. I'll take pics for anyone interested.
Then I'll send the current unit back to Amazon for the refund.
I applied the same CPU fan curve as you, mine was a bit different:
Temp1: 40
Temp2: 50
Temp3: 65
Temp4: 85
PWM1: 50
PWM2: 70
PWM3: 90
PWM4: 150
VS your curve:
Temp1: 20
Temp2: 30
Temp3: 50
Temp4: 60
PWM1: 50
PWM2: 80
PWM3: 110
PWM4: 255
I may look at applying a similarly more aggressive curve to the other fans.
Also thanks for the powersave script, I applied the script to run at array start on my Unraid and I've seen low temps, even as the summer temps rose a little bit over the last few days in the heatwave, and its been very good.
Idle temps are much lower, even with all my services "running".
The MB temp is still alarmingly high at 117c, which wasn't the case when I got the unit at first... Not sure what's going on there.
If I look at all the sensors, I see the that the components including NVMEs are not this hot, but the board temps are sky high:
Next weekend I expect I'll be able to have everything done, and will be able to see the difference.
I'm fairly certain the the NCT6798 doesn't have all of the sensors connected (or at least they are not the standard sensors for that chip and need appropriate but unknown offsets applied). I get the same kind of readings on my MS-01 but a scan with a thermal camera can't find anything that might be that close to boiling water. Its also telling that those readings (MB Temp and AUXTIN0..2) never seem to change.
I wouldn't worry about them too much.
Well that's interesting because 1) These MB sensors were not at this level when I first got the MS-01, and have gone up this way around the same time as the instability I have reported and 2) The readings do change, as you can see from the various screenshots the "MB Temp" currently ranges between 115-117c.
I'd be interested to see screenshots of the temps of YOUR MS-01.
The temps of mine were reported much lower before, and I haven't adjusted anything related to this, so it seems to me that yes something is wrong but as to what is wrong either the sensor is broken/incorrect or the temp is correct and the board is reaching this temp and there is a hardware issue.
I know for a fact that The MB Temp was not 117c when I first got this.
Here it is again as I'm writing this to show that the MB temp does change as well as the CPU:
Something is wrong, and I suspect that there will be different readings when I swap into the newly arrived MS-01 later this week/weekend.
All things being thre same (Same hardware, but with the PTM7950) it will be the best test comparison I can make.
Last edited:
Not the same format but here you go. You can clearly see the same pattern of ridiculous and obviously incorrect readings for SYSTIN (what you report as ('MB temp') and AUXTIN0, 1, 2, 3 & 4. As I said before, I've been all over the MB with a thermal imager and NOTHING on the board is running over 100c - not even close!
Code:
nct6798-isa-0a20
Adapter: ISA adapter
in0: 368.00 mV (min = +0.00 V, max = +1.74 V)
in1: 1.03 V (min = +0.00 V, max = +0.00 V) ALARM
in2: 3.34 V (min = +0.00 V, max = +0.00 V) ALARM
in3: 3.34 V (min = +0.00 V, max = +0.00 V) ALARM
in4: 1.09 V (min = +0.00 V, max = +0.00 V) ALARM
in5: 168.00 mV (min = +0.00 V, max = +0.00 V)
in6: 152.00 mV (min = +0.00 V, max = +0.00 V) ALARM
in7: 3.34 V (min = +0.00 V, max = +0.00 V) ALARM
in8: 3.17 V (min = +0.00 V, max = +0.00 V) ALARM
in9: 1.02 V (min = +0.00 V, max = +0.00 V) ALARM
in10: 176.00 mV (min = +0.00 V, max = +0.00 V) ALARM
in11: 144.00 mV (min = +0.00 V, max = +0.00 V) ALARM
in12: 1.01 V (min = +0.00 V, max = +0.00 V) ALARM
in13: 168.00 mV (min = +0.00 V, max = +0.00 V) ALARM
in14: 1.27 V (min = +0.00 V, max = +0.00 V) ALARM
fan1: 2307 RPM (min = 0 RPM)
fan2: 1862 RPM (min = 0 RPM)
fan3: 0 RPM (min = 0 RPM)
fan4: 0 RPM (min = 0 RPM)
fan5: 0 RPM (min = 0 RPM)
fan7: 0 RPM (min = 0 RPM)
SYSTIN: +109.0°C (high = +80.0°C, hyst = +75.0°C)
(crit = +95.0°C) sensor = thermistor
CPUTIN: +47.5°C (high = +80.0°C, hyst = +75.0°C)
(crit = +95.0°C) sensor = thermistor
AUXTIN0: +106.0°C (high = +80.0°C, hyst = +75.0°C) ALARM
(crit = +100.0°C) sensor = thermistor
AUXTIN1: +107.0°C (high = +80.0°C, hyst = +75.0°C) ALARM
(crit = +100.0°C) sensor = thermistor
AUXTIN2: +108.0°C (high = +80.0°C, hyst = +75.0°C) ALARM
(crit = +100.0°C) sensor = thermistor
AUXTIN3: -40.0°C (high = +80.0°C, hyst = +75.0°C)
(crit = +100.0°C) sensor = thermal diode
PECI Agent 0: +47.0°C (high = +80.0°C, hyst = +75.0°C)
AUXTIN4: +103.0°C (high = +80.0°C, hyst = +75.0°C) ALARM
(crit = +95.0°C)
PCH_CHIP_CPU_MAX_TEMP: +0.0°C
PCH_CHIP_TEMP: +0.0°C
PCH_CPU_TEMP: +0.0°C
PCH_MCH_TEMP: +0.0°C
intrusion0: ALARM
intrusion1: ALARM
beep_enable: disabled
nvme-pci-0100
Adapter: PCI adapter
Composite: +37.9°C (low = -20.1°C, high = +83.8°C)
(crit = +88.8°C)
acpitz-acpi-0
Adapter: ACPI interface
temp1: +27.8°C
coretemp-isa-0000
Adapter: ISA adapter
Package id 0: +47.0°C (high = +100.0°C, crit = +100.0°C)
Core 0: +44.0°C (high = +100.0°C, crit = +100.0°C)
Core 4: +43.0°C (high = +100.0°C, crit = +100.0°C)
Core 8: +43.0°C (high = +100.0°C, crit = +100.0°C)
Core 12: +41.0°C (high = +100.0°C, crit = +100.0°C)
Core 16: +43.0°C (high = +100.0°C, crit = +100.0°C)
Core 20: +40.0°C (high = +100.0°C, crit = +100.0°C)
Core 24: +41.0°C (high = +100.0°C, crit = +100.0°C)
Core 25: +41.0°C (high = +100.0°C, crit = +100.0°C)
Core 26: +41.0°C (high = +100.0°C, crit = +100.0°C)
Core 27: +41.0°C (high = +100.0°C, crit = +100.0°C)
Core 28: +45.0°C (high = +100.0°C, crit = +100.0°C)
Core 29: +44.0°C (high = +100.0°C, crit = +100.0°C)
Core 30: +44.0°C (high = +100.0°C, crit = +100.0°C)
Core 31: +44.0°C (high = +100.0°C, crit = +100.0°C)
nvme-pci-5800
Adapter: PCI adapter
Composite: +37.9°C (low = -20.1°C, high = +74.8°C)
(crit = +79.8°C)That's interesting that you're experiencing high temps, but I don't think you're listening to me.
I communicated clearly that I was not seeing 100c temps on anything when I first got this unit, and I wish I recorded that, but of course I had no reason to do so at the time.
If I got this MS-01 and was immediatlely seeing reports above 100c, I would be alarmed. Who wouldn't?
Anyway that aside, there are other people who have posted details in this very thread which show "System" temps far below 100c.
From what I've seen on this forum, "System" temps should REPORT between 50-70c.
I think it may be worth considering that my unit and your unit might be the odd ones out here. There could be a defect here.
Unfortunately our sample size is small, and I would plead for others to post their MS-01 temps in this thread, because it's the only way we're going to find out. I am convinced this is not normal.
I have the replacement MS-01 right next to me here, still in the Amazon box, so I'm excited to get to the bottom of this.
So I have just received one of these units, and so far at idle I am seeing between 50-60c.
I'm going to use it as a Win 11 Desktop PC, so will push it a bit harder and keep an eye on the temps...
I'm going to use it as a Win 11 Desktop PC, so will push it a bit harder and keep an eye on the temps...
Thanks and PLEASE let us know how you get on.
Very interested to see your CPU and System temps.
So it's all built up and in use......
CoreTemp reports at idle most cores 56-58c.
Max temps, most cores 72c, but the odd few cores hit 92c (like 2).
CoreTemp reports at idle most cores 56-58c.
Max temps, most cores 72c, but the odd few cores hit 92c (like 2).