Apr 9, 2020
12
0
10
Hi, my computer was working well till yesterday but since this morning I'm getting two errors when I turn on my PC.
"CPU Fan Error" and sometimes "CPU Over Temperature Error"

BIOS - updated to latest version.
Fans - All 3 fans are working well including liquid cooled fan for CPU. Cleared off the dust first.

The system doesn't proceed to "Starting Windows" screen at all. It's stuck at "CPU Fan Error".
The red LED light of TPU on MoBo stays turned on.

In BIOS CPU temperature shows 88°C.
By chance booted system successfully once and HWMonitor shows CPU temperature 98°C on idle!

How can I solve that?

Specs:

CPU: Intel Core i7-4790K
GPU - Asus GTX 970
MoBo : Asus Z97-A
Cooler : Corsair Hydro H60
 
Solution
Everything looks normal. No OC, no retarded voltages, fans normal..

Coolant level issues usually start out intermittent, multiple spikes as coolant struggles to flow. Blockages are sudden and lasting. Even paste failures aren't sudden unless physically made so by breaking the paste seal, and even then it'd take a hard physical movement to affect the entire cpu. So from the looks of things, it's a blockage in the pump.

It's not unheard of for core sensors to fail and give false readings, but the chances of all the cores failing simultaneously is astronomically slim.

Time to call Corsair.

Barty1884

Retired Moderator
The CPU fan is working.... but is the pump?

An AIO liquid cooler has the additional point of failure in the pump. The fan can be moving air through the radiator, but if the pump isn't moving liquid from the block (CPU) to the radiator, it's not dissipating heat.
 
Apr 9, 2020
12
0
10
The CPU fan is working.... but is the pump?

An AIO liquid cooler has the additional point of failure in the pump. The fan can be moving air through the radiator, but if the pump isn't moving liquid from the block (CPU) to the radiator, it's not dissipating heat.
How can I check if pump is working or not?
I just pointed infrared thermometer on pump and noticed temperature rising from 42°C to 56°C within 2 minutes and then I switched off my PSU.
 

Paperdoc

Polypheme
Ambassador
Open your case so that you can reach in and feel the two tubes from the pump to the radiator. Ensure your system has been off for a while so it is cool. Turn it on, and feel both hoses. If the pump is working, one hose will feel warmer than the other, but not a huge difference. If they are the same it is likely the pump is not moving any warm fluid out to the rad and has failed. As before, don't let this run too long - 2 min is plenty long enough.
 
Apr 9, 2020
12
0
10
Open your case so that you can reach in and feel the two tubes from the pump to the radiator. Ensure your system has been off for a while so it is cool. Turn it on, and feel both hoses. If the pump is working, one hose will feel warmer than the other, but not a huge difference. If they are the same it is likely the pump is not moving any warm fluid out to the rad and has failed. As before, don't let this run too long - 2 min is plenty long enough.
I could feel the liquid moving though both pipes but I couldn't notice any temperature difference. You said the difference won't be huge so I was extra careful to notice the difference in temperature but couldn't find any.

While I was checking that the temperature rose to 72°C within 30-45 seconds of checking the pump.

Shall I replace my pump or is there something else?
 

Paperdoc

Polypheme
Ambassador
Call Tech Support at Corsair and ask for their help. A comment, though: you say you can feel the fluid flowing in the tubes, but no temp difference. It can be hard to know whether fluid is flowing - normal operation would produce very little vibration caused by the flow. However, one thing that CAN cause lots more vibrations for those tubes is if PART of the fluid has been lost from the system so that the fluid flows in surges, but not smoothly enough to actually carry heat up to the rad for removal. I don't know if that is what is happening to your system. I do know you cannot fix that yourself, if it is. The liquid loop system basically is filled and sealed at manufacturing time.
 
Apr 9, 2020
12
0
10
Call Tech Support at Corsair and ask for their help. A comment, though: you say you can feel the fluid flowing in the tubes, but no temp difference. It can be hard to know whether fluid is flowing - normal operation would produce very little vibration caused by the flow. However, one thing that CAN cause lots more vibrations for those tubes is if PART of the fluid has been lost from the system so that the fluid flows in surges, but not smoothly enough to actually carry heat up to the rad for removal. I don't know if that is what is happening to your system. I do know you cannot fix that yourself, if it is. The liquid loop system basically is filled and sealed at manufacturing time.
After reading your message I checked again to feel vibration. I DO feel vibration a little but as you said it's hard to know if fluid is flowing through or not, am I feeling vibration because those pipes are attached to radiator/fan? Fan is working fine.

This time check temperature of both pipes individually I used an infrared thermometer. Before turning on computer the temperature of both of them was 31.6 °C - 31.8 °C

After 1 minute turning on my computer I check temp of both pipes - one pipe's temp didn't change and stayed 31.7°C -31.8°C while other pipe's temp rose to 34.5 °C

Is that normal?

Note - I built this system 5 years ago and applied thermal paste only once yet. Never reapplied.

I still have a few months warranty left, shall I replace my cooler or try reapplying thermal paste?

Or the problem is something else? The TPU red LED light on motherboard stays lit.
 

Paperdoc

Polypheme
Ambassador
Do not worry about the LED marked TPU. See your mobo manual p. 1-41. There is a user- operated switch next to that you can use to let the mobo do a type of automatic overclocking for improved performance. You can Disable this or use either of two levels of improvement. The LED merely tells you that you have set the switch for one of those automatic improvements. It is NOT a warning of a problem.
 
Apr 9, 2020
12
0
10
Do not worry about the LED marked TPU. See your mobo manual p. 1-41. There is a user- operated switch next to that you can use to let the mobo do a type of automatic overclocking for improved performance. You can Disable this or use either of two levels of improvement. The LED merely tells you that you have set the switch for one of those automatic improvements. It is NOT a warning of a problem.
Relieved to know that.
So now I have two options I guess
  1. Reapply thermal paste and recheck temperature
  2. If problem persists then I replace my liquid cooler.

Will be able to do it only after lockdown ends.
I will give update after that.
 

Karadjgne

Titan
Ambassador
Aios have a 3rd point of failure most don't consider, part of which was mentioned earlier, and thats the coolant itself.

Coolant failure happens in 2 ways either the level is insufficient (thats the biggest cause of failure after the 5 year (ish) mark. Even though the tubing is a special low evaporation loss compound, evaporation still happens as the chemicals start breaking down into their basic elements, and oxygen is a specifically small element and will slowly pass through the hose. It's the same practice as tires that slowly go flat, a process stopped by using nitrogen in the tires vrs regular air.

The second, and can happen anytime but more commonly on old units, is gunk. Get a bad batch of coolant with insufficient biocide and you get algae buildup in the coolant. When that happens, it'll travel around until it hits the microfins in the pump and gets stuck, blocking coolant flow through the fins.

Pump works, fans work, but there's very little flow and very high cpu temps as a consequence.

Normally, the output tube from the pump will be 2-4°C higher than the input tube, so those temps look normal. I'd say you have at least some flow, but it's maybe not enough. Your rad will only be a little warm, you are changing the heat output from something 1cm x 2.5cm and exchanging it on something that's 12cm x 24cm.

Bios temps should be closer to 45°C ish.

You have a cpu that starts out at @ 100w, and due to OC, usage, gaming etc, can output 150w easily. On a 140w aio that's intrinsically exactly the same as a CM hyper212 evo. Not the best combination for cooling an i7.

Check your bios, see what voltages the cpu is at. If something has changed bios settings and vcore is dangerously high for some odd (TPU?) reason, it'd explain the temps easily on such a small cooler on a big cpu.
 
Apr 9, 2020
12
0
10
Aios have a 3rd point of failure most don't consider, part of which was mentioned earlier, and thats the coolant itself.

Coolant failure happens in 2 ways either the level is insufficient (thats the biggest cause of failure after the 5 year (ish) mark. Even though the tubing is a special low evaporation loss compound, evaporation still happens as the chemicals start breaking down into their basic elements, and oxygen is a specifically small element and will slowly pass through the hose. It's the same practice as tires that slowly go flat, a process stopped by using nitrogen in the tires vrs regular air.

The second, and can happen anytime but more commonly on old units, is gunk. Get a bad batch of coolant with insufficient biocide and you get algae buildup in the coolant. When that happens, it'll travel around until it hits the microfins in the pump and gets stuck, blocking coolant flow through the fins.

Pump works, fans work, but there's very little flow and very high cpu temps as a consequence.

Normally, the output tube from the pump will be 2-4°C higher than the input tube, so those temps look normal. I'd say you have at least some flow, but it's maybe not enough. Your rad will only be a little warm, you are changing the heat output from something 1cm x 2.5cm and exchanging it on something that's 12cm x 24cm.

Bios temps should be closer to 45°C ish.

You have a cpu that starts out at @ 100w, and due to OC, usage, gaming etc, can output 150w easily. On a 140w aio that's intrinsically exactly the same as a CM hyper212 evo. Not the best combination for cooling an i7.

Check your bios, see what voltages the cpu is at. If something has changed bios settings and vcore is dangerously high for some odd (TPU?) reason, it'd explain the temps easily on such a small cooler on a big cpu.
When I open BIOS and leave with "Discard changes and exit" then I can log in my system.
I took pictures for you to analyse so please take a look.

6 pics taken of BIOS and HWmonitor.

View: https://imgur.com/a/W36786B
 

Karadjgne

Titan
Ambassador
Everything looks normal. No OC, no retarded voltages, fans normal..

Coolant level issues usually start out intermittent, multiple spikes as coolant struggles to flow. Blockages are sudden and lasting. Even paste failures aren't sudden unless physically made so by breaking the paste seal, and even then it'd take a hard physical movement to affect the entire cpu. So from the looks of things, it's a blockage in the pump.

It's not unheard of for core sensors to fail and give false readings, but the chances of all the cores failing simultaneously is astronomically slim.

Time to call Corsair.
 
Solution
Apr 9, 2020
12
0
10
Everything looks normal. No OC, no retarded voltages, fans normal..

Coolant level issues usually start out intermittent, multiple spikes as coolant struggles to flow. Blockages are sudden and lasting. Even paste failures aren't sudden unless physically made so by breaking the paste seal, and even then it'd take a hard physical movement to affect the entire cpu. So from the looks of things, it's a blockage in the pump.

It's not unheard of for core sensors to fail and give false readings, but the chances of all the cores failing simultaneously is astronomically slim.

Time to call Corsair.
Thanks for narrowing down the main problem. I will contact Corsair to replace my cooler as it is under warranty.

I will update this thread after getting callback from them and solution.

I'm grateful to you for your help and others' as well.
 
Apr 9, 2020
12
0
10
Everything looks normal. No OC, no retarded voltages, fans normal..

Coolant level issues usually start out intermittent, multiple spikes as coolant struggles to flow. Blockages are sudden and lasting. Even paste failures aren't sudden unless physically made so by breaking the paste seal, and even then it'd take a hard physical movement to affect the entire cpu. So from the looks of things, it's a blockage in the pump.

It's not unheard of for core sensors to fail and give false readings, but the chances of all the cores failing simultaneously is astronomically slim.

Time to call Corsair.
Update:

Finally after two months and end of lockdown Corsair replaced my cooler. I got new version of H60 with SATA cable attached to pump.

I successfully connected my cooler and now I don't get errors:
"CPU Fan Error" or "CPU Over Temperature Error"

and my computer boots easily.
However CPU temperature is still very high.

When I check in BIOS while booting my CPU temperature shows 52-55°C and goes up to 75°C and stays stable.

And when I log in my Windows and check temp with HWmonitor it shows me 70°C + on idle too. When I launched Chrome I could hear fan going loud and running super fast and temperature 98°C .

Why it happen? My cooler is new.
I want to mention that I didn't apply any thermal paste as paste was applied already. However I feel that it was very thin layer and while handling pump some paste was scrapped with my thumb.

Is that the reason for high temperature or anything else going on?

@Karadjgne @Paperdoc @Barty1884

Thank you
 

Karadjgne

Titan
Ambassador
Umm probably. There's a bunch of temp sensors inside the cph, and if you left a scratch mark on the paste, it could have left an empty 'hot spot'.

What you can do is seat the pump. Run the pc at idle for 10 minutes or so, let it warm up good. Then shutdown. Pop the side of the case, push gently on the pump and give it some wiggle left/right.

The paste will be warm, so quite malleable, and under the pressure from your push should spread quite evenly. Just don't push too hard, nor on one side, a nice even center.

It's not going to wiggle much, if it does, you aren't tight to the motherboard. That's been an issue in the past, but the fix is easy. Just use a rubber/Teflon/paper washer under the backplate, you can do it 1 screw at a time, so no need to dismount the pump.
 
Apr 9, 2020
12
0
10
Umm probably. There's a bunch of temp sensors inside the cph, and if you left a scratch mark on the paste, it could have left an empty 'hot spot'.

What you can do is seat the pump. Run the pc at idle for 10 minutes or so, let it warm up good. Then shutdown. Pop the side of the case, push gently on the pump and give it some wiggle left/right.

The paste will be warm, so quite malleable, and under the pressure from your push should spread quite evenly. Just don't push too hard, nor on one side, a nice even center.

It's not going to wiggle much, if it does, you aren't tight to the motherboard. That's been an issue in the past, but the fix is easy. Just use a rubber/Teflon/paper washer under the backplate, you can do it 1 screw at a time, so no need to dismount the pump.
I followed what you advised, I let my computer run for 10 minutes on idle. The temperature stayed varied from 50-75°C during that. CPU usage was 5-20%.

As soon as I launched Chrome it went up to 98°C again but I immediately shutdown within seconds. Then I removed side of my cabinet and pushed the pump from center a little and tried to wiggle it. The pump is sitting there fixed tight so it didn't wiggle at all.

After doing that I turned on my computer again and faced the same error "CPU Over Temperature Error" and in BIOS the temperature rose from 75°C to 88°C within seconds.

I assumed it CPU was already hot because I shut it down at 98°C. So I turned it on again after 5 minutes of rest and this time I didn't get temperature error while booting and successfully logged in but my CPU temperature stayed high 75-80°C on idle.

What should I do now?
 

Karadjgne

Titan
Ambassador
That's high, even for bios, which generally runs at @ 50% loads on the pc. Assuming the cooler is fully functional, it's possible it's actually a bad sensor in the cpu itself.

You can use ThrottleStop to bypass that sensor, allowing you to check the individual core temps in something like HWInfo. But that requires getting into windows.

Verify in bios that everything is actually correct. Vcore should be @ 1.0v ish, disable any q-fan controls to allow full pump/fan speeds, psu voltages should be correct for the rails, no crazy OC values etc.
 
Apr 9, 2020
12
0
10
That's high, even for bios, which generally runs at @ 50% loads on the pc. Assuming the cooler is fully functional, it's possible it's actually a bad sensor in the cpu itself.

You can use ThrottleStop to bypass that sensor, allowing you to check the individual core temps in something like HWInfo. But that requires getting into windows.

Verify in bios that everything is actually correct. Vcore should be @ 1.0v ish, disable any q-fan controls to allow full pump/fan speeds, psu voltages should be correct for the rails, no crazy OC values etc.
My system is not overclocked.
I clicked a picture of my BIOS for you. Let me know if you find any problem in this.

View: https://i.imgur.com/j3yF8Kn.jpg



After 8 hours I turned on my computer again. The temperature in BIOS started with 55°C and rose to 88°C within 10-15 seconds only. The fan became very fast and loud from the beginning.

How can I disable q-fans and check PSU voltages?

Would you recommend me to re-apply thermal paste? ( any basic or Arctic MX4)

Also, my apologies for taking this for too long but it's my main work computer and can't find a solution. You've been very kind and helpful. Thank you.
 
Apr 9, 2020
12
0
10
Update:

So finally I resolved my issue. I identified the problem. Actually after getting new liquid cooler I hadn't replaced the previous backplate and standoff screws. The standoff screws in newer cooler version are shorter hence my pump was sitting tight on screws but it wasn't sitting properly on my CPU.
I noticed that there was no thermal paste applied on CPU only at one side of it which might have been applied there while trying to set pump on it. But in the end my pump wasn't touching CPU at all.

I applied Noctua thermal paste and replaced mounting hardware with new. Now I get idle temperature 31°C - 35°C and then I ran HeavyLoad software and stress tested 100% CPU for 5 minutes and it reached maximum 75°C .

I believe my problem is solved now. Thank you all for giving me valuable inputs to resolve this issues. I appreciate everyone's help here.
Thank you