[SOLVED] Cpu cooler dead? Im not sure what happened!

[Dman99]

Hello and thank you to anyone who can offer me some help with this. I'll list my system specs to get that out of the way, its 5ish or so years old now, some parts are newer or older.

i7 6700K
Asus Maximus Formula VIII Mobo
16gb Trident Z RGB
Gtx Msi 980ti
H100i V2 (an older model by now for sure)
NZXT H440 loaded up with Corsair Mag Lev fans. 6 of them, 3 intake (hindered by a bunch of drives and stuff) and 3 exhaust (2 on Radiator)

My system has ran well for a very long time and I have a decent amount of experience building them. I was playing games all day today with perfectly normal temperatures. 60c or so on more demanding games like Metro Exodus. Played it for a few hours straight without stopping with no issues. I go to restart my computer to switch to a different os (switch to my main Linux drive as Exodus doesn't work for me as well in Linux) and immediately get a bios error saying cpu temp error. Its not lying either, jumps quickly to 90c in just the bios.

I did nothing except restart my computer. Its physically been in the same spot as it has for a while. It feels like the pump is running, there's a soft vibration coming from it and theres no abnormal clicking or other odd audible noise to me anyways. I don't see any indication that the aio leaked either but I do know its older and water will escape over years of time.

But, I guess I don't understand how it could go from running as usual to immediately terrible temperatures. When it first brought me to the bios it said 46c, after about 1 min it was at 88c. I shut it down at that point. I've let it cool down a few times and it still is getting wicked hot really fast for just being in the bios.

Some fixes I tried was going to default bios settings and clearing CMOS. Ive tried switching fan header's from aio pump to the pump header vs having it on the cpu fan header which it been connected to since I've had this computer. I haven't checked the thermal paste application but I've not removed the pump for 2ish years and have never had temp issues. Im more than willing to, but don't want to if its unnecessary.

I also believe my aio is installed in the suggested way as to not cause air bubbles to get into the pump. Once this issue arose, I tried flipping the pc around carefully to try and get water in the pump to move around but nothing is working out for me. I also tried to look at similar forum posts for solutions but I couldn't find any that would apply or where the problem came so abruptly in a older and seemingly stable system.

My personal inclination is that something is wrong with my pump/it reached the end of its life but I'm really not sure overall. I would love to know thoughts on this problem whatever they are. Thank you immensely for any responses. If I missed explaining anything, I apologize and will try to fix it immediately!

 

[CountMike]

Hello and thank you to anyone who can offer me some help with this. I'll list my system specs to get that out of the way, its 5ish or so years old now, some parts are newer or older.

i7 6700K
Asus Maximus Formula VIII Mobo
16gb Trident Z RGB
Gtx Msi 980ti
H100i V2 (an older model by now for sure)
NZXT H440 loaded up with Corsair Mag Lev fans. 6 of them, 3 intake (hindered by a bunch of drives and stuff) and 3 exhaust (2 on Radiator)

My system has ran well for a very long time and I have a decent amount of experience building them. I was playing games all day today with perfectly normal temperatures. 60c or so on more demanding games like Metro Exodus. Played it for a few hours straight without stopping with no issues. I go to restart my computer to switch to a different os (switch to my main Linux drive as Exodus doesn't work for me as well in Linux) and immediately get a bios error saying cpu temp error. Its not lying either, jumps quickly to 90c in just the bios.

I did nothing except restart my computer. Its physically been in the same spot as it has for a while. It feels like the pump is running, there's a soft vibration coming from it and theres no abnormal clicking or other odd audible noise to me anyways. I don't see any indication that the aio leaked either but I do know its older and water will escape over years of time.

But, I guess I don't understand how it could go from running as usual to immediately terrible temperatures. When it first brought me to the bios it said 46c, after about 1 min it was at 88c. I shut it down at that point. I've let it cool down a few times and it still is getting wicked hot really fast for just being in the bios.

Some fixes I tried was going to default bios settings and clearing CMOS. Ive tried switching fan header's from aio pump to the pump header vs having it on the cpu fan header which it been connected to since I've had this computer. I haven't checked the thermal paste application but I've not removed the pump for 2ish years and have never had temp issues. Im more than willing to, but don't want to if its unnecessary.

I also believe my aio is installed in the suggested way as to not cause air bubbles to get into the pump. Once this issue arose, I tried flipping the pc around carefully to try and get water in the pump to move around but nothing is working out for me. I also tried to look at similar forum posts for solutions but I couldn't find any that would apply or where the problem came so abruptly in a older and seemingly stable system.

My personal inclination is that something is wrong with my pump/it reached the end of its life but I'm really not sure overall. I would love to know thoughts on this problem whatever they are. Thank you immensely for any responses. If I missed explaining anything, I apologize and will try to fix it immediately!

Can BIOS read pump speed ? Are you using any SW to regulate fan speeds ?

 

[MeanMachine41]

With the temps your experiencing and it happening all of a sudden,it will be either the TIM needs to be replaced or the AIO is failing.
Only way is to either swap out the cooler with a known working unit and replace your TIM with a good one.

Check all cabling and connections and also for dust buildup. Clean all fans as dust can build up a hot spot after 5 years.
Bring any Overclocks back to default then stress test again.

 

[Paperdoc]

This does sound like a failure of fluid movement. Could be either pump has failed, or there's a large air bubble in it because of progressive loss of fluid.

I'm presuming you have checked that the rad fans are running. In fact, with those rapid high temp readings, your fans should have ramped up to max speed pretty quickly.

Try this check. When running (repeat over as long a time as you dare run your system), feel the two hoses coming out of the pump. In normal operation, one will feel slightly warm, one a bit cooler. If you find that neither gets warm at first, but both get about equally warm (or maybe even very warm) after a few minutes, that would confirm that the fluid in the pump is NOT being forced out through the loop. Check also the hoses at the rad - if neither ever warms up, that's another proof fluid is not flowing.

If those tests indicate flow failure, it could be either of the causes I suggested. Now, iCue (if that's what you are using) will tell you clearly the PUMP speed and the rad fan speed. So, does it tell you the pump is running full speed (usually about 2000 RPM, maybe more)? If not, then the pump motor has failed. If it IS running, but apparently not making the fluid flow, other pump parts may have failed, or there may be a bad air bubble preventing the pump from working. In all three of these cases, about the only solution is replacement of the pump. And often thatimeans replace the entire system - it's just easier and you get a whole new AIO system.

 

[Dman99]

Thank you all for all the replies, I have been swamped today with stuff and haven't been able to look at my computer until now. I apologize for that and appreciate you all offering help. My issue has some interesting updates that have me perplexed as to what happened/is going on. I had the power completely off to my PC all night and all day today. When I turned it on, my temps were higher than usual for being in the BIOS (45c) but was not exceeding that really. When I looked at my boot drives to load an OS, my windows installation is just gone, no windows boot manager I should say, I think the drive is still listed.

My main boot/OS is Linux Pop Os 20.10 and that was still listed so I booted into that. For some reason my temps are completely normal in Pop Os. My idle temps have a package of 28c which is somehow way lower than what the bios is reading. On Linux I'm just using Psensor for temps, I believe its using the same sensor id's that the bios uses. I've restarted a few times checking the bios temps and OS temps and its been pretty consistent at those temps. As far as software control of fans/pump, I'm not using anything on Linux, but I do use iCUE on my windows boot. I also have fan curves set in the BIOS, which I just redid since I cleared CMOS and reset to default settings last night.

On the AIO, last night the fans were definitely spinning at max speed. Today since the temps are seemingly in check, I can't hear them at all but are still spinning and speeding up if needed. In the BIOS, the pump today was reading about 1600 or so rpms. I was not able to get its speed last night and I also dont have any way in Linux right now either (windows is still gone at least to boot, I can probably recover it) . To the other suggested tests, one tube does feel warmer than the other including at the radiator. One thing I'll add is that before this issue happened Windows 10 was the last OS I used and it finished an update when I booted into it, then as I said I played games with no issues for a few hours, restarted and had the issues indicated in my topic message. Doubt that's important, but I don't understand what happened here.

Also, my system is pretty clean, I try to clean it at least once a month blowing out the radiators and fans at least. I have not yet replaced TIM or removed pump, I know I put Arctic Silver 5 on it originally 4ish years ago, but I have some Noctua NT-H2 TIM I can replace it with if needed. (and before anyone asks, I removed the pre applied stuff and did a pea sized dot on the IHS) I don't have another cooler to use right now but I'm kinda thinking regardless its probably time for a new AIO or cooler in general. I'd like to build a whole new system but parts are impossible to get right now it seems.

I'm still just so confued why any of this happened. I dont understand why my computer was overheating lightning fast last night, why my temps are better today but still weird with the BIOS being hotter than in my main OS, and why Windows boot manager is no longer there so that I cannot currently boot into it. Other than having it off, I didn't do anything between my post last night and booting it up today. I would genuinly love and appreciate any help or thoughts with this because I'm at a loss. Thank you to everyone!

 

[Karadjgne]

There's 2 parts to a pump. The motor and the diaphragm. As noted, you feel the vibration, so the motor is running. But that doesn't mean the diaphragm is functional.

AIO's of certain age can also suffer from plugging. That's where algae has finally overcome the biotoxins in the coolant and built up enough where a chunk has worked itself free. This chunk gets shoved through the loop until it hits the pump. The pump would normally shove the coolant (that's the diaphragms job) through some fins and out the other hose to the rad, but if that algae chunk blocks the fin array, the coolant doesn't move and you'll get almost instant temp soaring.

That can easily explain your issue, why it worked and then quit. If the diaphragm has failed, that too can explain your issue, coolant is no longer traversing the loop.

Either way, it sounds like you have a coolant flow issue and it's generally not user fixable unless you are comfortable disassembling the pump, flushing the system and replacing the coolant.

 

[Paperdoc]

Karadjgne makes some interesting points above.

OP, your most recent post stirred other thoughts. If I understand correctly, this all started suddenly shortly after a Windows Update that appeared to be OK because you were playing a game. However, since then you cannot boot into Windows at all, so obviously Cue is not operating. Thus how your AIO system is controlled is unknown. This does, however, lead to the question: if you could not boot Windows, what were you doing when you recognized the severe overtemperature?

The fact that Windows cannot be found and loaded suggests that it was corrupted, probably in the update operation. That leads to our not knowing what it did (or could have made iCUE do) to your mobo configuration settings before it was last shut down.

The installation instructions of the H100i system, as I recall, tell you to plug the 3-pin female fan connector from the PUMP unit into the mobo CPU_FAN header, and the two rad fans into output connectors on a cable from the pump. Then you connect a cable from the pump to a mobo USB2 port and install and run iCUE. If you do it that way, then the pump unit gets power to run from the CPU_FAN header, and that header receives from it the PUMP speed, and monitors that for possible failure. So within BIOS Setup (independent of what OS), the CPU_FAN speed shown will be the PUMP. Your most recent post says that now, under apparently normal operation, that speed is around 1600 RPM. By the way, with this connection, as long as the CPU_FAN header is left in the PWM Mode option, the header can NOT change the pump speed - it always will be full speed. The only way to change the pump speed would be to change the header's configuration to Voltage Control Mode and then reduce the speed setting there. BUT if that header is left in AUTO Mode, it may do a test and find out that what is connected to it acts like a 3-pin fan, and automatically change its own configuration to Voltage Control Mode, so that then it CAN run the pump at a reduced speed. If it is running, iCUE can ensure that does not happen even in AUTO Mode, but without iCUE it might.

Regarding your CPU temperature reading under Linux using a third-party tool, many such tools are known to give incorrect readings until you calibrate them. Although they may use the same raw data reading from the CPU internal sensor, the way they process that into a Celsius reading depends on calibration parameters that may not be the same as the mobo is using. I would trust what BIOS Setup shows you.

Back to the H100i system. When connected as its instructions say, the rad fans can NOT be controlled by the mobo - there is no connection between them. This also means the BIOS Setup screens cannot tell you their speeds. Instead, iCUE takes control of those fans via its USB2 communication link to the pump unit. iCUE does use the CPU internal temperature reading it gets from the mobo and sets the rad fans according to that. It can show you the rad fan speed. If iCUE fails, or does not run, or cannot communicate with the pump, the pump has a default rad fan setting that is fairly fast (not clear whether full speed or not) to ensure the CPU gets pretty good cooling, even if not max cooling. That is a fixed speed. Of course, all of this only happens when iCUE is running under Windows. No Windows, no iCUE, and the H100i system will fall back to default operations.

So, when you run under a Linux OS, there is no real mobo control of the H100i system at all. The pump is the only item connected to the mobo, and it runs full speed. The fans have no mobo connection, so they operate at the H100i system's default fast fixed setting. And THAT makes me wonder: you say that now, under Linux, you have seen those fans on the rad CHANGE speeds. The only way I can understand that happening is if the rad fans are NOT connected to the pump outputs, but are plugged into some mobo fan header. Is that how you have them?

Come to think of it, there MIGHT be another way those rad fan speeds can be changed by the mobo in response to CPU temperature. Under the connection scheme according to the installation instructions, the rad fans are not under direct control by the CPU_FAN header. BUT their power IS taken from the CPU_FAN header via the 3-pin cable from it to the pump. Normally with that header configured to PWM Mode, that supply voltage from Pin #2 will be 12 VDC always, so the pump can run full speed and the fans at some fixed speed, according to the pump's default fan setting. BUT if the CPU_FAN header is set to Voltage Control Mode, OR if it is in Auto Mode and has changed itself to Voltage Control, then the power supply on Pin #2 WILL be reduced from 12 VDC when the CPU is at moderate temperatures, and BOTH the pump and the rad fans will be slowed down, AND may change their speeds as that voltage is altered by the mobo in response to CPU temp changes. So, check the configuration details in BIOS Setup for the CPU_FAN header.

A couple further thoughts. You were using a utility, Pop Os, under Linux to read temperatures. Any chance it also makes changes in BIOS configuration settings that affect the CPU_FAN header? Also, I note your report that the tube temperartures at both pump and rad ends show a typical small difference, which would indicate that, at least right now, there IS some reasonable flow of liquid around the loop. So for now at least, the pump is running and there is no bad blockage.

 

[Karadjgne]

The pump and fans are not controlled until windows starts. But that doesn't mean they are unpowered. They run just like everything else at POST, at @ 50ish % load. Starting any liquid cooled pc that happens, fans come on immediately under bios instructions, control happens like a case fan and ramps down with windows and iCue startup.

 

[Paperdoc]

I disagree. Fan headers and what they power and control are all done by the mobo in common situations. But when using the Corsair system, iCUE takes over for the CPU_FAN header only, and that means when Windows loads and then iCUE, control sudddenly changes for that particular set of hardware. IF you connect as they direct, the rad fans are NOT connected to any mobo header, and so are not handled that way any more. As soon as the POST proccess begins on the mobo the rad fans will start up according to the H100i system's default power output (from the pump) to its rad fans, and then change as iCUE takes over. But if Windows never loads, the pump and fans will continue in their default state to ensure CPU cooling with no actual temperture-dependent control. Similarly, if you load another OS instead (like a Linux) that does NOT load and run iCUE, then the H100i system will continue to operate in its default "no iCUE no control" state. HOWEVER, all the power for the H100i system is coming from the power pins of the CPU_FAN header. In the absence of iCUE takeover, IF that header is operating in the older Voltage Control Mode, then that power source is NOT kept at a constant 12 VDC. It is reduced BY THE MOBO according to its read of the CPU temperature, which would reduce the speeds of both the pump and it rad fans just as the speed of a 3-pin fan cooling the CPU would do.

 

[Karadjgne]

That's what I said.

The pump and fans are not controlled until windows starts. But that doesn't mean they are unpowered. They run just like everything else at POST, at @ 50ish % load

That's default. @ 50% fans. Pump is 100%.

control happens like a case fan and ramps down with windows and iCue startup.

After windows loads, the motherboard software takes over control of any fan headers, but until then they operate under bios instruction fan curve settings. Which at 40-50°C in bios and a 70°C default max, puts fans at roughly a 50%-60% rpm as well.