Intermittent fault on CPU01 cooler.

[Capablanca]

I noticed the fan speed beginning to pick up and wondered why as the workstation was not under high load. HP Performance Advisor reported CPU1 temp at 82C, whilst CPU0 was at 41C. All the fan speeds were identified as being normal.

I rebooted the workstation and now CPU1 temp is just 51C, 10C more than CPU0.

I've had this situation at least a couple of times now.

Has anyone had a similar intermittent fault, or could you provide any advice as to what the problem might be?

Any feedback would be most welcome.

Have a great day!

Andrew

[Brian1965]

I would recommend trying the following;

1.   Unmount the CPU coolers from both CPU's (leave the CPU's in-place if possible)

2.   Carefully clean off the old thermal grease from both CPU's and CPU coolers

3.   Re-apply fresh thermal grease to the back of both CPU's

4.   Re-mount the CPU coolers but swap them over

If CPU0 then runs hot and CPU1 seems fine then the CPU cooler (now mounted on CPU0) is probably damaged (heat pipe?) and needs replaced. If CPU1 still runs hot then the CPU may be damaged. If both CPU's run fine then the issue was likely due to dried out thermal grease (which loses most of its thermal conductivity if dry).

Try installing Open Hardware Monitor and Prime95 (both are free). Just search for the latest versions online. With  Open Hardware Monitor running on the desktop, run the Prime95 torture test (Small FFT's = maximum heat!). Run the test for a good 10 minutes while monitoring the CPU temperatures. You might want to try this first, before swapping the CPU coolers so you have a baseline to measure any changes.

[DGroves]

the heatpipes in a air cooled cooler can sometimes fail without any warning, and while this is rare it does happen

depending on your installed cpu wattage, you may have the standard coolers (2 pipes) or the high performance 130 watt (3 heatpipes) make sure to only replace a 130 watt cooler with same as a stock low power cooler will not work right 

[Capablanca]

Brian,

Thanks for the quick response, most appreciated.

I’ve already tried what you recommend, in a roundabout way.

I had this problem three weeks back and (before I knew about HP Performance Advisor) thought the problem was with CPU0. Hence I bought a replacement off eBay and swapped it out with new paste, without removing the CPU.

 I still got a warning at boot (527 – CPU Liquid Cooling Pump(2) not detected) so replaced CPU1 Cooler with the one just removed, again with new paste, without removing the CPU.

This solved the warning at boot, but now I’m getting this intermittent temp differential on CPU1.

I note the original cooler on CPU1 had 4 wires, the replacement only 3, I wasn’t sure if that has any significance.

My only other thought was to get a replacement cooler (with 4 wires on the connector).

Thanks for your performance testing software recommendations.

Regards,

Andrew

[Capablanca]

I've got the 2 cooler pipes version and replaced like for like, with the exception of the number of wires in the connector, as explained above.

Thanks for taking the time out to help, appreciated.

[DGroves]

(527 – CPU Liquid Cooling Pump(2) not detected)  is a error that should only appear if you have the optional 150 watt water cooling option thes AIO coolers have a small pump at the base of the cooling plate and use the two optional cpu 3 pin headers instead of just relying on the cooling fans in the main shroud

DO YOU HAVE THE AIO COOLERS??

IF NOT THEN YOU MUST USE A 4 PIN PWM BASED FAN

[Capablanca]

Yes, I have the liquid coolers with the pump in the base of the cooling plate as you describe.

Z820 Liquid CPU Cooler 

Note this cooler only has 3 wires. For some reason, the existing cooler on CPU1 has 4 wires, with the additional wire connecting to the base plate.