Terrible temps with new loop.

The paste itself cant be great either tbh. Take the chips in the ocuk extreme bundles, theyre delidded and the paste is replaced with liquid metal. But the glue that holds the ihs in place is left on so the ihs is still the same distance from the die. Temps are still upto 20c lower than a standard chip.
 
Sounds like you have a restriction in your cpu block, either an air lock or some other obstruction. What coolant have you used in the loop? You may have some built up coolant gunk or plasticiser from your tubing blocking the cpu block micro channels.
 
I'd drain the loop and take the cpu block apart and check the micro channels, if you find red gunk that'll be the culprit and can be easily rinsed away.

Has the coolant been in the loop long? Also how old is the tubing or is the loop completely new?
 
The build got finished yesterday, I will try draining it tomorrow morning and having a look at the block. I did flush the rad but I might have left something behind.
 
Otherthan that you could have the classic air lock in a roof mounted rad. Does the rad have a bleed screw by any chance?
 
What are temps like with something like Asus real bench? This is a bit less demanding on haswell chips than the likes of prime/aida64.
 
I did an hour of XTU at 4.7ghz @ 1.250v and it was 65-75, Real Bench is around the same. The air coming off my rad with the temps at 75+ is cold.

JeffyB said:
Otherthan that you could have the classic air lock in a roof mounted rad. Does the rad have a bleed screw by any chance?
Click to expand...

I have tipped the case upside down, on its front and back, I have shaken it and no air has came out. All with the pump turned on its max speed with it held outside the case.
 
Last edited:
Can be tricky to bleed a loop with the rad in the roof, you have to chase the air from component to component ending up at the reservoir. Whilst doing this you'd have to keep the res vertical to stop air from leaving the res. When you move the case around can you hear liquid sloshing around in the rad? It so there's air in the rad.
 
I believe Aida64 FPU stress test uses AVX coding which absolutely hammers Haswell and also adds a extra 0.1v to the vcore. Try using IBT which doesn't use AVX. I saw temps 15-20 degrees less with IBT.
 
Where are you reading the 1.2 volts from? BIOS or desktop? Could voltage be reporting wrong? Unlikely but just a suggestion to double check.
 
pastymuncher said:
I believe Aida64 FPU stress test uses AVX coding which absolutely hammers Haswell and also adds a extra 0.1v to the vcore. Try using IBT which doesn't use AVX. I saw temps 15-20 degrees less with IBT.
Click to expand...

Yep, same as p95 ver 28, AVX on it too, whereas ver 27is a lot less demanding.
 
I have no issues with Haswell systems and Aida64. On systems that would cause a 4770k to go past 100Degrees with IBT or Prime they work fine with Aida64 Stress test. Aida64 is recommended by Intel to test stability with Haswell CPUs as a matter of fact the press/review samples went out with a pack containing Aida64 as well as a few CPU monitoring tools rather than using IBT/Prime etc..
 
That would be about right then. The extra heat when using Aida would be from the AVX coding hammering the cpu. That's normal. If I use Linx (which uses AVX) to stress test the cpu it hit's the high 70's-low 80's as well as drawing a extra 0.1v. If I use IBT (which doesn't use AVX) the cpu is in the low-mid 50's.
 
If it's not loop restriction/gunk/airlock I hope he has a bad ihs contact (sorry as it's your cpu) and this is not common for all 4xxx chips, I may have to go the direct contact route if I upgraded.
 
JeffyB said:
If it's not loop restriction/gunk/airlock I hope he has a bad ihs contact (sorry as it's your cpu) and this is not common for all 4xxx chips, I may have to go the direct contact route if I upgraded.
Click to expand...

I am going to try to remove the IHS tomorrow and put some Liquid Pro on it tomorrow.
 
You must log in or register to reply here.
Back
Top Bottom