CodeMaster
Expert Member
For the last week my N40L Microserver has been randomly shutting down.
I originally thought it was just because I need to replace the UPS battery and just thought it was a dip in power causing it to shut down. When I got a gap to start looking into it, I started with the Event Viewer. Bit since it was just spontaneously shutting down with no BSoD, there was nothing in the events list and no dump files either.
This started making me think it may be hardware and not software.
So I installed "HW Monitor" and straight away saw that the Nvidia GT210 card was overheating. At that stage it was hovering around 70°C, but as soon as I fired up Kodi then it started climbing and quickly climbed over 100°C. I checked the specs on the card, and MAX temp is 105°C, so I guess that's the point at which the card shuts down the PC.
I thought that I had put a silent card in the machine, so I suspected that I may have needed to redo the thermal paste, as the card is over 2 yrs old.
When I whipped it out, it is actually a fan cooled card. So I stuck it in my desktop to check it, and it powered up fine. Fan was spinning no problem, with no extra noise or anything.
With the case still open, temps started out at around 42°C. If I stopped the fan, then temps climbed, and if I let it go then temps came down again. I didn't leave it in for a long time, and also didn't install any drivers. Put it back into the Microserver, and the temps were fine again at idle, but when pushed a bit they climbed.
I managed to keep it running overnight to get some downloading done, but only by keeping the front door open and temps this morning were just under 100°C.
I currently have it running in my desktop at work, using Windows 8.1 drivers, and it's running fine since 8am.
HW Monitor reports it as 55°C min and 61°C max. Fan speed seems to be on average 1% per 1°C - so at 50°C it's around 50% rpm, So it definitely looks like the fan is working fine.
Since it is running fine in my desktop, I am thinking it probably isn't the thermal paste, but I will redo it anyway.
My work desktop has 5 case fans, so the airflow is a hell of a lot better than the Microserver, but as you can see in the screenshot below, all the other temps are well within range, so I don't believe it's an airflow problem. Besides the 12cm rear fan in the Microserver is working, and nothing else has changed with the placement of the server.
I am also going to remove all the Nvidia drivers and software tonight, and just install the Windows certified drivers for the card.
Beside that, what else could I try?
Could it realistically be a driver problem that's causing the high temps in the Microserver?
I originally thought it was just because I need to replace the UPS battery and just thought it was a dip in power causing it to shut down. When I got a gap to start looking into it, I started with the Event Viewer. Bit since it was just spontaneously shutting down with no BSoD, there was nothing in the events list and no dump files either.
This started making me think it may be hardware and not software.
So I installed "HW Monitor" and straight away saw that the Nvidia GT210 card was overheating. At that stage it was hovering around 70°C, but as soon as I fired up Kodi then it started climbing and quickly climbed over 100°C. I checked the specs on the card, and MAX temp is 105°C, so I guess that's the point at which the card shuts down the PC.
I thought that I had put a silent card in the machine, so I suspected that I may have needed to redo the thermal paste, as the card is over 2 yrs old.
When I whipped it out, it is actually a fan cooled card. So I stuck it in my desktop to check it, and it powered up fine. Fan was spinning no problem, with no extra noise or anything.
With the case still open, temps started out at around 42°C. If I stopped the fan, then temps climbed, and if I let it go then temps came down again. I didn't leave it in for a long time, and also didn't install any drivers. Put it back into the Microserver, and the temps were fine again at idle, but when pushed a bit they climbed.
I managed to keep it running overnight to get some downloading done, but only by keeping the front door open and temps this morning were just under 100°C.
I currently have it running in my desktop at work, using Windows 8.1 drivers, and it's running fine since 8am.
HW Monitor reports it as 55°C min and 61°C max. Fan speed seems to be on average 1% per 1°C - so at 50°C it's around 50% rpm, So it definitely looks like the fan is working fine.
Since it is running fine in my desktop, I am thinking it probably isn't the thermal paste, but I will redo it anyway.
My work desktop has 5 case fans, so the airflow is a hell of a lot better than the Microserver, but as you can see in the screenshot below, all the other temps are well within range, so I don't believe it's an airflow problem. Besides the 12cm rear fan in the Microserver is working, and nothing else has changed with the placement of the server.
I am also going to remove all the Nvidia drivers and software tonight, and just install the Windows certified drivers for the card.
Beside that, what else could I try?
Could it realistically be a driver problem that's causing the high temps in the Microserver?
