High CPU temps

Toss3

Toss3

Toss3

Associate
Joined
13 Oct 2009
Posts
778
So I've got the following components in my loop:

Pump: Laing DDC-1 3.25
Reservoir: XSPC Acrylic Tank Reservoir
Rads: EK-Coolstream XT 120 mm + Phobya Xtreme 200mm
Fans: EK-Waterblocks Vardar F4-120ER 2250RPM + Bitfenix Spectre Pro 230mm (both running full speed)
Tubing: Tygon 16/13mm Noprene R6016
CPU Block: EK-Waterblocks Supremacy EVO Acetal (I1 insert + J2 jet)
GPU Block: EK-FC Titan X - Acetal


Cooling the following components:

CPU: i7 4770k 4.3Ghz at 1.19V (delidded + Naked IVY mounting set from EK-Waterblocks)
GPU: GTX 980ti at 1495mhz

The loop is set up as follows:
RES/PUMP -> 120mm RAD -> CPU -> 200mm RAD -> GPU -> RES/PUMP

Problem:

During gaming the GPU stays under 40c(1495mhz) while the CPU shoots up to around 60-70c

Idle: 39-40c CPU and 29c GPU

During stress-testing the CPU shoots up to >85c (the stress test shuts down). (OCCT linpack with AVX)

I have reseated the CPU many a times already and made sure there is proper contact between the CPU and Block, yet I cannot figure out why the CPU temps would be so high, especially considering the voltage I am running on. My best guess would be that the block is either broken/malfunctioning, or that the CPU/block contact is somehow messed up despite my attempts to make proper contact (CPU could also be broken, as the CPU really shouldn't be running at a much higher temp than the GPU considering the GPU uses up a lot more W than the CPU (almost three times as much)). Or maybe the 120mm isn't enough to cool the temps coming out of the GPU before flowing into the CPU? Should I invest in a 240mm rad?

Any ideas? :confused:
 
Last edited:
if its shooting up in a matter of seconds it sounds to me like either block incorrectly seated or built, thermal paste incorrectly applied, or a blockage of some kind.

Is the thermal paste spreading correctly once the cpu block is screwed down? De-lidded correctly and paste contacting die and heat spreader correctly?

What pump speed are you using? Fluid flow correct? As i know EK supremacy blocks have an IN port on the left and OUT on the right.

Are these stress test temps after gaming or on a cold boot?
 
tyler_jrb said:
if its shooting up in a matter of seconds it sounds to me like either block incorrectly seated or built, thermal paste incorrectly applied, or a blockage of some kind.

Is the thermal paste spreading correctly once the cpu block is screwed down? De-lidded correctly and paste contacting die and heat spreader correctly?

What pump speed are you using? Fluid flow correct? As i know EK supremacy blocks have an IN port on the left and OUT on the right.

Are these stress test temps after gaming or on a cold boot?
Click to expand...

Everything is making contact (took it apart a couple of minutes ago and it is touching the CPU), but now I am wondering if it maybe the 120mm that is giving me trouble, as the Vardar 120mm fan is mounted on the outside of the case which leaves a couple of mm space between it and the radiator. Tested again and now the max temp in the OCCT Linpack bench is 77-79c which is still too high considering I am running lower than stock voltage and rads with a larger surface area than a 360 rad (at least I now know the cpu is making contact).

Maybe I should try mounting the 120mm rad on top to make sure the fan makes good contact with the rad (that way I could probably get two 120mm Vardars cooling it instead of just one)? Also thinking about getting a Silverstone AP182 to replace my current Bitfenix Spectre Pro 230mm.

EDIT: Mounted the 120mm fan on the inside with the 120mm rad to make sure it is making proper contact, and while the noise definitely increased the temps didn't. Also I measured the space between the gpu block and top of the case and concluded that I only have space for a dual Alphacool ST30 with two vardars (or Coolstream PE 240 if I mount the vardars on the outside). Dual 240mm would only be possible if I mounted the fans on the outside.
 
Last edited:
It will probably be not enough pressure from the CPU-CPU block i had the same problem when i directly cooled the CPU, i went back to using the IHS and it solved it as it puts plenty of pressure on.

I've also found the first time i applied liquid pro to a CPU i got a tiny strand of cotton from the cotton bud on there and it caused awful temps, i couldnt even see it cause it was so tiny but i ran my finger across the die and it pulled it out then my temps shot down.

That fact that your GPU runs cool shows your loop is ok, so the only things that could be causing it really is bad contact or a incorrectly configured CPU block, but even if ya using the wrong jet or orientation in your block the temps wouldnt be that bad, they wernt with me at least. Just to be sure you could check your using the correct in/out but its hard to get that wrong if ya followed the instructions building it.
 
Last edited:
BluD said:
It will probably be not enough pressure from the CPU-CPU block i had the same problem when i directly cooled the CPU, i went back to using the IHS and it solved it as it puts plenty of pressure on.

I've also found the first time i applied liquid pro to a CPU i got a tiny strand of cotton from the cotton bud on there and it caused awful temps, i couldnt even see it cause it was so tiny but i ran my finger across the die and it pulled it out then my temps shot down.

That fact that your GPU runs cool shows your loop is ok, so the only things that could be causing it really is bad contact or a incorrectly configured CPU block, but even if ya using the wrong jet or orientation in your block the temps wouldnt be that bad, they wernt with me at least. Just to be sure you could check your using the correct in/out but its hard to get that wrong if ya followed the instructions building it.
Click to expand...

I'll try to disassemble the loop tomorrow and lap the heatspreader and try adding some liquid pro between the cpu and it to see if that would solve the issue. I'm starting to think my CPU is somehow just broken; it's been like this from the beginning, but I thought it was just due to some shoddy tim between the core and heatspreader, but I haven't been able to drop the temps even after delidding.
 
Some can be. Depends how far the sensors are out in the die. More than likely its incorrect seating of the die-IHS. Or IHS-block.

Ensure the IHS itself makes good contact with the die when you strip it down tomorrow.

Id also use ROG realbench aswell for testing. XTU has been known to overstress.
 
Last edited:
Yeah i would just do like ya said, clean the DIE, block and IHS thoroughly and put them back together using normal thermal paste first, one youve found out if it solves it then go to liquid pro and lapping.

You can tell if youve got good contact by the way the thermal paste is all pushed out around the contact area, the top of my IHS or DIE/block just look like someone has breathed on it slightly theres so little paste left in there.

Liquid pro tend to dry hard and can be annoying to remove, i have to relap mine every time i put it back together so i would make sure its all working good with normal paste first.
 
BluD said:
Yeah i would just do like ya said, clean the DIE, block and IHS thoroughly and put them back together using normal thermal paste first, one youve found out if it solves it then go to liquid pro and lapping.

You can tell if youve got good contact by the way the thermal paste is all pushed out around the contact area, the top of my IHS or DIE/block just look like someone has breathed on it slightly theres so little paste left in there.

Liquid pro tend to dry hard and can be annoying to remove, i have to relap mine every time i put it back together so i would make sure its all working good with normal paste first.
Click to expand...

Lapped the underside of the IHS (the top side was actually really really flat), and put some liquid ultra between the core and the IHS - temps shot up to 94c at boot, but I quickly realized the IHS wasn't touching the core, so had to take it all apart once again, and it turns out that my ultra thin layer of electrical tape was keeping the IHS from touching the core. :p Well off with the tape and replace the liquid ultra and nh-1, and voilà temps now max out at 60c on 1.19v (still quite high, but this is with linpack AVX so I believe it is pretty good fore what I have currently at least especially considering the ambient temps are around 30c (PC is located just below the ceiling on a shelf and there's no AC)).

Hoping the dual rad (EK coolstream-pe 240) and Silverstone ap182 will make the temps even better! Thanks for all of your help! :cool:
 
Cool, just replace the Electrical tape with some plain nail vanish, it adds no obstruction but nicely insulates the caps which i assume is what you were doing. I switched to nail vanish last time i took mine out after someone else suggested it.
 
So I just received my 240mm EK-CoolStream PE and was thinking of keeping all three radiators in the case(going to need some modding). Because the PC is sitting on a shelf near the ceiling the temps get pretty high as the ambient temps rise, and was thinking of maybe cutting a hole in the side of the case and have my 230mm Spectre Pro blow cool air into the case, and all other radiator fans pushing air outwards from within the case (or would pull be better? Guess I have to try both to see which is better). That way the inside of the PC should stay relatively cool.

Like this:
casetmsku.png
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom