[ControlAlternateDelete]

I have a 6-year-old, self-built, AIO water cooled PC (specs below) that for various reasons has only been in regular use for the past 3 years. For those last 3 years, there has been no change in configuration and things have been running smoothly (successful stress test done a year ago). Two days ago, I received a "CPU Fan Error" and “CPU Over Temperature Error” on bootup and on subsequent examination, the CPU was found to be running consistently at 99-100 C at idle (rapid linear increase).

SPECS:

Cooling (liquid): Corsair H100i v1
CPU: Intel i7-4770K
MoBo: Asus Maximus VI Formula
PSU: Corsair HX1000
Case: Corsair Carbide Air 540


From my reading, the possible reasons for this "CPU fan error and overheat" could be:
a) Cable connection issues (to the PSU and/or motherboard)
b) Thermal paste has dried up
c) Poor seating of CPU cooler and/or dust clogged exhaust fans
d) Failure of the AIO water cooler


What I did:

A) Cable Connection Issues: Since the configuration has not changed and the computer has been running smoothly till now, I assume that this cannot be the issue. In any case, all connections were found to be secure (including the three pin "CPU_fan") and I even changed the SATA connection to the PSU with no improvement.

B) Thermal Paste Drying: Old thermal paste was removed and replaced with fresh paste

C) Poor Seating or Clogged Fans: Seating was confirmed to be solid; exhaust fans (horizontal, located at the top of the case) are visibly working and dust free. Mild vibration in pipes is felt.

D) Failure of AIO Water-cooling: When the status of the H100i was checked using Corsair's Link application (screenshot included), it showed coolant temperature at 70-80 C (on average), pump fan speed at around 2500 rpm and exhaust fan speed at around 2700 rpm.



After all this, I still am receiving the same errors.

So, my question is how should I proceed? Everything seems to be working! Despite this, the CPU is not being cooled.

A) Cables seem secure.

B) Thermal paste has been replaced.

C) Despite the CPU fan error thrown up, the Link software shows both the pump fan and exhaust fans are running. In addition, the high temperature of the coolant would seem to indicate that it is taking up the heat from the CPU. Is it sufficient to say the water cooler is working because the fans are working? Could the fans be working and the water cooler still not doing its job? Any way to check that?

D) Additional Question: The CPU is currently running at 100 C at idle/minimum load, as soon as Windows loads. Is that solely because of the lack of heat removal or could the CPU itself be at fault? Shouldn't a CPU at idle run at a much lower temperature even in the absence of cooling?

Finally, where can I find the fan settings in BIOS? I cannot seem to locate them.

I do not want to buy a new CPU cooler only to find out the problem is something else since the refund/return policy where I am is non-existent. Any help is appreciated. If any other information is required, I will try my best to provide it. Thanks for reading.

 

[Corday]

Softwares | CPUID Download HW Monitor (free version) and report temps.

 

[ControlAlternateDelete]

Softwares | CPUID Download HW Monitor (free version) and report temps.

HW Monitor Screenshots ~30 secs after bootup:



CPU temps are 90 to 100 C (increases to 99 to 101 C with time) and Coolant temps are 60 to 70 C (increases to 80 C with time).

 

[johnwill]

What does task manager say about the processor load? Is it loafing or is something pinning the CPU to the rails?

If you see excessive CPU use on this graph, go to the Processes tab and sort on CPU usage and see who is using all the processor.

 

[ControlAlternateDelete]

What does task manager say about the processor load? Is it loafing or is something pinning the CPU to the rails?

View attachment 331427

If you see excessive CPU use on this graph, go to the Processes tab and sort on CPU usage and see who is using all the processor.

View attachment 331428

Thanks for highlighting this point, johnwill!

The CPU under virtually no load (1-3% CPU usage only): Essentially at idle.



This brings me to an additional question to make things clearer (I've will add it to the original post as well, if I can):

The CPU is currently running at 100 C at idle/minimum load, as soon as Windows loads. Is that solely because of the lack of heat removal or could the CPU itself be at fault? Shouldn't a CPU at idle run at a much lower temperature even in the absence of cooling?

 

[johnwill]

I think you need to look at the cooling and how the heatsink is mounted. It sounds like you have a lousy bond between the heatsink and the processor.

As a comparison test to your situation, I just fired up CPU-Z for a stress test on my processor, after a few minutes I took these readings. The Processor is running in the low 60's, about where I'd expect maxed out. I have seen it hit 70C, the highest I've ever seen for a temperature. That was processing a large video file, I expected it to be getting hot under the collar.

Task manager assures me we're using all the processor.

As soon as I stopped the stress test, temperatures dropped into the normal range.

 

[ControlAlternateDelete]

As a comparison test to your situation, I just fired up CPU-Z for a stress test on my processor, after a few minutes I took these readings. The Processor is running in the low 60's, about where I'd expect maxed out. I have seen it hit 70C, the highest I've ever seen for a temperature. That was processing a large video file, I expected it to be getting hot under the collar.
...
As soon as I stopped the stress test, temperatures dropped into the normal range.

johnwill, thanks for the idea. I actually did a full system stress test about a year ago (using Prime95 for the CPU) and the desktop behaved exactly as you described. So I know there was no problem then.

I think you need to look at the cooling and how the heatsink is mounted. It sounds like you have a lousy bond between the heatsink and the processor.

I thought that might be the case so I changed my thermal paste two days ago and checked the heatsink seating before and after: both times it was firm.

It seems more and more likely that the issue is with the AIO water cooler: Broken impeller? Coolant loss or block?

If I need to get a good CPU air-cooler, what do you suggest (the Noctua NH-D15 and DEEPCOOL Assassin III are too large for my current setup)?

 

[johnwill]

Yep, the symptoms certainly scream out cooling issues. You should be able to determine if the cooler is pulling heat off the processor. If the radiator isn't pushing out significant heat, there's your problem. If the processor is at 100C, and the water cooler was working, you should be getting some really hot air out!

I think the first step would be to see if the water cooler is doing something, and maybe see if you can simply fix it before starting over with a new system.

 

[ControlAlternateDelete]

Yep, the symptoms certainly scream out cooling issues. You should be able to determine if the cooler is pulling heat off the processor. If the radiator isn't pushing out significant heat, there's your problem. If the processor is at 100C, and the water cooler was working, you should be getting some really hot air out!

I think the first step would be to see if the water cooler is doing something, and maybe see if you can simply fix it before starting over with a new system.

Radiator is not even warm . I think the water cooler must be the issue then. May be an impeller or coolant block/loss problem.

 

[ControlAlternateDelete]

Parallel thread for replacement air-cooler recommendations:

CPU Cooler Recommendation for Limited Vertical Space Setup

Your opinions would be welcome!

 

[johnwill]

Radiator is not even warm . I think the water cooler must be the issue then. May be an impeller or coolant block/loss problem.

Clearly, that's the issue. I've only used one water cooler, and I haven't built a PC in a few years, so I don't have any current opinions on replacements. However, now that you have defined the problem, I suspect a solution will come to mind. I'd check and see if it's something simple with the current setup like maybe the impeller motor isn't running...

 

[ControlAlternateDelete]

Clearly, that's the issue. I've only used one water cooler, and I haven't built a PC in a few years, so I don't have any current opinions on replacements. However, now that you have defined the problem, I suspect a solution will come to mind. I'd check and see if it's something simple with the current setup like maybe the impeller motor isn't running...

Much obliged for the help, johnwill.

 

[Sophus]

According to a video, premature AIO failures can be caused by non-optimum mounting of the components relative to each other...and many people install their AIOs improperly (many case and AIO designs don't help by restricting how you can mount the items).
Here's a video that provides information that could be of help in improving the lifespan. My current AIO install hasn't been in place long enough to personally confirm this but it does appear to be based on reasonable observations.
HowToKillYourCPUcooler(LiquidAIO.MountingOrientation)
youtube.com/watch?v=BbGomv195sk

 

[ControlAlternateDelete]

Here's a video that provides information that could be of help in improving the lifespan. My current AIO install hasn't been in place long enough to personally confirm this but it does appear to be based on reasonable observations.
HowToKillYourCPUcooler(LiquidAIO.MountingOrientation)
youtube.com/watch?v=BbGomv195sk

Thanks for the link, Sophus. Fortunately, according to that video, my cooler was installed properly: radiator (mounted horizontally) above pump. So whatever caused the error, it was not the configuration.

 

[ControlAlternateDelete]

Hyper 212 RGB Black Edition.

Installed and working great!




Marking this thread solved, repping and closing (if possible and allowed).

Thank you all for reading and esp. johnwill and Sophus for your help. You made a complex issue, straightforward.

 

[johnwill]

Glad to see a happy ending.