Unusual temperatures - what to do?

Ant84

Ant84

Ant84

Associate
Joined
20 Feb 2007
Posts
1,094
So I tried overclocking for the first time yesterday with my i3 530 chip and GA-H55-UD2H board. All seemed to be going quite well as I'd managed 4.07Ghz just by adjusting the multiplier and bclk, as well as setting the QPI/VTT to 1.29v. Ran Prime95 for half an hour and it was stable.

However, the temperatures seemed far, far higher than they should. I'm using an Arctic Cooler v2 and the cores were hitting 96 degrees at 100% load. Correct me if I'm wrong, but surely the chip should be nowhere near those temps with an aftermarket cooler? I read a walkthrough on another site using the very same chip and board, and they never went above 55 degrees even at 1.5v.

Any ideas where I might have gone wrong here, or what I could do to remedy the situation?
 
Not just yet. I ended up lowering the QPI/VTT right down in an attempt to lower temperatures, now it won't POST. Inexperience, I guess!

When I open it up to reset CMOS, I'll try reseating the cooler. Do you think it might be worth buying some decent thermal paste, or is that not the likely cause?
 
Some better thermal paste wouldn't hurt, IC Diamond paste seems to be the way to go, your going to need some arcticlean or similar too to clean up the chip and cooler face.

The only other thing it could be is very poor airflow or a dust clog, what case are you using?

You should also try incremental overclocking, gradually increase the speed then stress test, check everythings fine then take it a notch higher, rinse and repeat.
 
That's what I did, really. At around 3.6Ghz, the max temps were around 75 degrees, so I might just revert to that.

The case is a Coolermaster 335.
 
I've not had experience with those chips but after some quick searching I've found people with that chip @4.4ghz on air only hitting 72c. Though other evidence does suggest that these chips can run very hot under a large OC, especially if the coolers not up-to the job.

I would try a re-seat with some fresh compound, try increase airflow a little in your case. If that doesn't help you may want to look into a new CPU cooler. The Freezer Pro 7 rev.2 coolers are good, but not brilliant for a heavy overclock.
 
Cheers, I'll give that a go and report back later.

How far do you think I should be able to push the chip with this cooler (reseated)?
 
If you get the temps to settle down I'd say 3.8-4ghz would be well within its capability based off other overclocks i've seen of the same chip. Just keep an eye on the temps on your overclocks and try keep the voltage as low as you can while stable.

by all means take some pics once you have removed the cooler before you clean the chip, we should be able tell (roughly) if it has been seated correctly.
 
Well so far I've removed the battery to reset CMOS. The overclock remained at 4.07Ghz even after I got it back up and running again (I thought CMOS reverted everything back to standard?), and it seems stable at this speed.

If anyone would be kind enough to take a look at this and see if any of my settings look way off, I'd be most grateful. Not yet removed the cooler as I'm planning on buying a few bits tomorrow so I'll open it up properly then and take a few pics of the bottom of heatsink/CPU etc.

The max temps in Real Temp state 84 and 80 degrees. After a further ten minutes running Prime95, they increased to 88 and 85 respectively.

OCDesktop.png
 
Have you tried getting it stable on a lower vcore? Your at 1.424v, looking at some other overclocks at 4ghz on that chip that are stable at 1.3/31v. Infact the only clocks I've found at a similar voltage are running 4.6ghz+. Reducing voltage would significantly reduce temps...
 
When I was in the BIOS, I never adjusted the CPU vcore at all. The only voltage I touched was the QPI/VTT setting, from 1.1v to 1.29v.

That said, on the main settings screen, the vcore setting kept rising on that page without me doing anything. I wonder if it's doing it automatically?
 
If you didnt change it the Vcore will be set to Auto, so the mobo will be automatically controlling he Vcore. Iv found that with Vcore set to auto the mobo will use more volts than required, so id keep an eye on it :)
 
Ahh, well that makes sense.

I'll restart in a few minutes, set vcore to manual and adjust the setting to 1.3v, if that would improve things?

Are vcore and CPU vcore the same thing, or are they different settings?
 
They should be the same thing, there may be other voltage settings to change but I don't know too much about those chips or chipsets (someone else will!). Try drop it just a shade for now, say 1.38v, test it, check stable and temps, then drop it again, rinse and repeat until the OC fails, then bump it back up to the last stable voltage.
 
Ok, so I managed to get the vcore hovering around 1.3v and it's still running at 4.07Ghz. Prime95 been running for 15 minutes so far without a problem, so fingers crossed it's stable.

Max temps under 100% load are now 78 and 76 degrees respectively, down from 96 and 93 degrees. A fair improvement, but do I need to be looking at lowering the temperatures even further?
 
It wouldn't hurt to try, was 1.3v the lowest stable vcore?

Those temps are a huge improvement though, your going in the right direction
 
I reseated the cooler today when installing my new SSD. Used some Chill Factor 3 after cleaning the existing compound from the CPU and heatsink. Installed a 120mm fan at the front of my case, too.

Things have somehow gotten worse. Max temps now are at 85 and 82 degrees, despite only having improved thermal insulation and airflow.

No idea what's going on here.
 
Last edited:
You make sure the push pins on the cooler are sitting tight against the board?

How much compound did you apply and which method did you use?
 
You must log in or register to reply here.
Back
Top Bottom