Corsair H80i CPU Cooling Issues - Troubleshooting Sanity Check

[s_b]

Read through several different posts in the forums about H80i, and have a feeling that my H80i is at its ‘end of life’ but wanted a sanity check and see if there is anything else I can do to troubleshoot/fix.

 

System (2013 Build):

 

ASUS X79 DELUXE

Intel i7-4930K CPU @ 3.40GHz

CORSAIR Vengeance Pro 32GB DDR3 1866 

PNY GeForce GTX 770

CORSAIR Hydro Series H80i (Bought Nov 2013)

 

system has no overclocking

turned off XMP for RAM

Q Fan control for cpu_fan header disabled in BIOS

 

radiator hoses installed down (as recommended in the H80i installation manual)

 

motherboard/internal temp generally between 32C - 36C

 

Issue:

 

About 2-3 weeks ago started seeing thermal throttling (@> 80C) often on my pc.

 

Currently, the pc idle temp is ~60C. That's at ~15% cpu load.

 

During any short cpu spike ~20 sec (something as simple as opening chrome with like 5 tabs) then back to idle, the temp will go from ~60C to at/over ~80C in less than 10 seconds. temp will creep back down to near idle slowly after.

 

Haven’t tried intensive/long duration tasks as they just seem to peg the temp, then cpu throttling begins.

 

Troubleshooting/Diagnostic Steps Performed (none fix the issue):

 

First, I pulled the fans/radiator to give a good cleaning. There was some dust accumulation, but IMHO, not enough to be the sole issue.

 

Second, pulled the block and reapplied new thermal paste.

 

Third, installed Corsair Link.

 

CL shows cpu temp consistent with the cpu temp in the little ASUS program that comes with the MB , both H80i fans are running ~2200-2400 rpm and pump reports ~2240 rpm.

 

In Corsair Link, pump rpm does not change speed when toggle profile between performance, balanced, quiet. always stays between 2235-2250 rpm

 

My conclusions:

 

The fact that the cpu is running so hot at idle, and the temp spikes so much so quickly makes me think that while the fans/pump are running, there isn’t enough fluid in the loop for meaningful cooling? Or perhaps there is a blockage, but if that were the case, would anticipate fluctuations in the pump rpm? (or that there would be no rpm b/c it burnt itself out?)

 

Know the product is now at like 2x design life, so my instinct is to look for another cpu cooling solution, but would like to hear any other thoughts on how to diagnose/troubleshoot/fix this issue if possible.

 

Thanks!

[c-attack]

There might be some evaporation but most likely you’ve got a pretty decent blockage going on. On a cold boot you might get 5-10 min of normal temperature but since the fluid can’t get around the loop fast enough, the temp will continue to tick upwards a degree until it gets back to that 55-60C liquid temp.  The liquid temp is the minimum possible cpu temp with zero volts, so any time the Vcore goes up you’ll see cpu temps start from there. Ultimately this is how most “no maintenance” AIOs will end. There isn’t a user side fix, so it’s probably time to start looking at replacement options. 

[s_b]

@c-attack

Appreciate the prompt response.

Have done a few tests based on your comments and yeah, any +Vcore (from idle) = +15/20 C cpu temp spikes. -Vcore, the temp drops slowly back to idle-ish temp.

Last question, and mostly with my inexperience with AIOs - wouldn’t one expect odd/variable pump rpms if there was a blockage? Seems like the pump would have to work harder to get through that?

All said, thank you for the information, and my instinct was that this cooler was end of life.

Ordered three different coolers earlier today, concerned mostly about the fit b/c of the old form factor. Hopefully will be back up and running by the weekend.

Thanks again.

[c-attack]

If the blockage is severe enough to overcome the pump pressure, then you get a couple of things. 1) Really weird sounds coming from the pump motor as the blockage pressure pushes back. That likely would create some random rpm shifting, although that can be hard to distinguish from normal.  2) Instead of noticing temp seem warm, you’d have about 30 seconds after boot to figure it out before the motherboard pulled the fire alarm and shuts you down. At that stage the fluid in the micro fins isn’t moving anymore and it simply heats until shutdown, unable to transfer the heat anywhere. Users approaching that stage will often report one very warm to hot tube and the other is cool. At full stop it will get on both tubes coming up from the pump as the water sits there. Tubes should always be identical to touch with only an average 1C drop per pass through a radiator. 
 

Right now your flow rate is restricted and it’s slowing the water pickup and transport rate. Think of it as running your pump at 600 rpm instead of 2000. So while there is some trade off in flow rate for the amount of heat removed as it passes through a radiator (more time in channels, more heat removed), you don’t want the water to linger in the cpu or gpu block. That heat is transferred right back to the cpu. Your liquid in the block is hanging around in that 55-60C range and thus it heats your cold cpu to that temp too. Heat goes both ways across the cold plate. Then the cpu actually does something and you get a normal “+30C increase” from voltage, but on top of this new 60C baseline instead of 30C. That’s the problem and why liquid temp is what you are trying to control with the AIO. If I put your mb and cpu on a 4x480mm radiator system, your cpu will still go +30C when it loads. What changes is the ability to keep the baseline lower than with a smaller radiator. 50-60C is out of bounds for any AIO.