Mobo issue? CPU overheating (9900K) and high temps on TMPIN2 in HWMonitor

[timmymagic07]

Good afternoon,

I put together a build with a 9900K and the Gigabyte Z390 Aorus Pro Intel Z390 in early February last year, all bought from OCUK. The cooler I used was a Eisbaer lt360, which I had to RMA first time round as the pump did not work correctly.

I did a quick and easy OC of the CPU to 5Ghz that was stable in day to day use and ran a little hot during benchmarking (high 80s) but never went above 75c during the most intense gaming session (playing Quake Champions for a few hours, throwing out frame rates of 300+).

A few months ago the PC started crashing randomly now and again, and after checking the temps I noticed they were now reaching over 90c during my benchmarking (I use PerformanceTest from PassMark and just run the CPU Tests, so this is far from a thorough test) and decided to downclock to 4.8Ghz. This fixed the stability issues and I no longer had random reboots.

Over the past few weeks I started to hear the CPU cooler quite a bit more than normal, like even the most mundane task on the Desktop the fans would shoot up more than normal and so I got a bit concerned the thing was on its way out. I did some tests and saw my temps were still very high and decided to get the Arctic Freezer 2.0. This appeared to help at first, I saw a 10c or so temp drop at 4.8Ghz and things seemed OK, so much so that I managed to get up to 4.9Ghz stable albeit with higher temps (high 80s).

I then installed a couple more RAM sticks and an RTX 3080 and the CPU has gone back to being quite hot - this might be coincidence, it might be relevant. When I say quite hot, I mean high 80s and hitting 90 on the test mentioned above (it runs for about 2 minutes and has breaks, so hardly intensive).

To give you an idea of the numbers and temps, here are two test runs with HWMonitor showing temps/voltages first with the Eisbaer cooler and the second with the Arctic Freezer 2.0;

Here the Arctic Freezer 2.0;

Note I am clocking higher in the second test (4.9Ghz) and so this is not a like for like comparison as such, particularly as I now have 64GB ram (4 sticks) and a 3080 RTX too. Although the 3080 runs cooler than the 1080ti I had before. That said, during a CPU test the GPU is going to be idle anyway, so I doubt it makes much of a difference.

In addition to all this and something I should probably have mentioned higher up in this write up, is the following I found on the back of my motherboard. Originally I thought it might be something serious, but on reading up it might just be thermal pad residue? I noted earlier the TMPIN2 reading in HWMonitor being higher than the other readings, perhaps there is something overheating on my motherboard? Here is a photo of what I am talking about;

---

Is it possible I have just lost the silicone lottery and my 9900K is just a bit rubbish for OC? Or is there something wrong with my motherboard or the voltage I am running through it? From what I can gather they do not look too high.

I am thinking of getting the CPU delided and replaced with a copper heat-sink, but I am wondering if there is something I should try before going down that road?

Any help would be greatly appreciated - apologies for the length of this post, I have tried to include everything I can.

 

[orbitalwalsh]

With the eisbaer , running pump at 100% all the time ?

Ran all my esibaer pumps like that and should be silent and at full speed. Although alphacool does have the problem of China not filling them up as much as they should be !

 

[pp111]

Firstly I think it is quite possible that your water cooler is on the way out. They often decline slowly over time rather than just fail, and during their decline they may show sudden changes in performance when disturbed or the pump speed is changed. You disturbed the block when you changed the paste, which may explain it's temporary recovery.

 

[timmymagic07]

Yes I think the Eisbaer may have been on the way out anyway, but having installed the Arctic Freezer 2.0 the temps have stayed roughly the same or only slightly lower. So I am wondering if it is something else.

 

[Hakanese]

looks like you a screw missing coming through your cpu cooler mounting plate

 

[timmymagic07]

I just opened the thing up again and took a look and that is just the angle of the photo, all four are the same. I did tighten up the screws a bit and I might reapply the thermal paste, but I cannot see that making a huge difference.

Also if there was a poor connection to the CPU on one corner I would expect much higher temps on a couple of cores? Regardless it looks OK to me.

 

[timmymagic07]

OK now I am confused, I when into the BIOS and loaded Optimised Defaults and it sets the CPU at 5Ghz (albeit with lower 'idle' clock or whatever the correct name is) and temps are much better albeit the scores are quite a bit lower than the prior run at 4.9Ghz.

Hmmm, maybe HWMonitor is mis-reading the clock speeds and only one core is running at 5Ghz? I am not sure, I guess I will have to go into the BIOS and check but this is all a tad confusing.

UPDATE - EMP was disabled, on enabling that the clock speeds are not 5Ghz on all cores. I am not sure where to go from here, I guess I will just leave it like this and monitor temps in games whilst running at 5Ghz on four cores and see how stable she is. Might delid/relid, there is a service based a few miles from me that does it for £35 with copper heatsink fitted, might be worth a shot.

 

[SpartaP]

I doubt it'll be your cpu, and delidding may be a waste of £35. Yes it'll help but I think the route cause is elsewhere such as cooler. May be worth renewing some paste and remounting and see if that helps.

 

[Mr Mcgoo]

i would use hwinfo as hwmonitor is known for random temp bugs. however, it does seem your pump is playing up. Is you rad above your pump?