Question Weird fan behaviour,need advice

[yallowood]

Hello everyone, i have MSI h110m pro vd plus motherboard with 2 fans plugged in it, cpu fan (presumably DEEPCOOL CK-11508) and sys(case) fan (presumably Thermaltake RIING 12 120mm default model) both are 3 pins. At varying intervals my cpu fan suddenly goes nuts, i dont know what causes it, but after approximately two weeks of silent (tolerable) work it starts making loud noice at random moment (playing game,wathing movie, doesnt matter) and randomly changing between 800-1600 rpm, i mean literally, changing speed in this interval every 20 minutes and sounding like plane taking off, reloading pc doest help. This continues about 1-2 days, then it returns to normal 1800-2000 rpm with tolerable noice. It has been like that for half a year now. I tried controlling fan speed manually, but no matter which tool i use, bios(msi click bios with limited control options, i mean there are no pwm dc options etc) or any program for fan controlling, i cant control cpu fan at all. Meanwhile sys fan can be entirely controlled through all means, despite being plugged the same way as cpu fan. I really dont understand whats the problem, it just works as intended for two weeks and then becomes unbearable for couple of days and then again returns to normal, cycle repeats. I have been using this setup for 5 years so yes its quite old, but despite this i dont have any problems besides this strange behavior, also can confirm that there is no notable dust on either fan. Any thoughts on what might cause this?

 

[Paperdoc]

Considering the age of your system, you may be experiencing the early stages of worn bearings on the CPU cooling fan. I'll describe the symptoms. HOWEVER, what you describe is actually the reverse of what may worn fans do.

After several years of normal use, the bearings of a fan wear somewhat. This means the clearance of the motor shaft inside the bearing sleeve gets a bit larger and looser.This clearance is affected by temperature: as the fan is used and its bearings warm up, the sleeve expands more than the shaft, so the clearance gets looser when warm. The first noticeable symptoms of this process is that the system starts and runs smoothly as normal when cold, but after running for a while it starts to make a lot of noise because the fan shaft is "rattling" in the loosened bearing sleeve. This may also result in the fan's speed running slower, and maybe a bit erratic. This behaviour will continue as long as you run, and it will start again immediately if you just reboot. BUT if you shut down for an hour or more, the fan will cool off and run smoothly when re-started, until it heats up again.

Over the space of several months the bearing wear gets worse, so the initial silent operation period gets shorter until it is always noisy. If you continue to use the fan, eventually it will seize up completely and simply stop running, which means you get NO CPU cooling. So once it is almost permanently noisy you must replace that fan before such total failure.

Now, many fans undergoing this process do the reverse. That is, they are noisy at start-up, then quiet down when warmed up. Over time their initial noisy period becomes longer, etc. The difference between that and your experience is in the metal used for the bearing sleeve. If that metal sleeve expands on heating more than the shaft does, the clearance becomes larger and looser as it runs, producing noise when warmed up. If the bearing metal expands less than the shaft on heating, the clearance is loose and noisy at start-up but gets less and goes quiet when warmed. But it's the same process, and the same ultimate solution.

 

[yallowood]

But why in my case it happens so sporadically? For example it works without problems(no noice at startup or whatsoever) for a long amount of time, and then absolutely randomly during any activity which doesnt include temperature change instantly drops from 2000 rpm to 1200 rpm with noice no matter what i do and stays like that for one day , and then the next day it works normally at 2000 again? I just cant understand how to apply what you said about temperature to this pattern? And laso may be you know the reason why i cant control cpu fan at all? Because i can control sys fan despite being 3 pin and being connected in absolutely same way.

Considering the age of your system, you may be experiencing the early stages of worn bearings on the CPU cooling fan. I'll describe the symptoms. HOWEVER, what you describe is actually the reverse of what may worn fans do.

After several years of normal use, the bearings of a fan wear somewhat. This means the clearance of the motor shaft inside the bearing sleeve gets a bit larger and looser.This clearance is affected by temperature: as the fan is used and its bearings warm up, the sleeve expands more than the shaft, so the clearance gets looser when warm. The first noticeable symptoms of this process is that the system starts and runs smoothly as normal when cold, but after running for a while it starts to make a lot of noise because the fan shaft is "rattling" in the loosened bearing sleeve. This may also result in the fan's speed running slower, and maybe a bit erratic. This behaviour will continue as long as you run, and it will start again immediately if you just reboot. BUT if you shut down for an hour or more, the fan will cool off and run smoothly when re-started, until it heats up again.

Over the space of several months the bearing wear gets worse, so the initial silent operation period gets shorter until it is always noisy. If you continue to use the fan, eventually it will seize up completely and simply stop running, which means you get NO CPU cooling. So once it is almost permanently noisy you must replace that fan before such total failure.

Now, many fans undergoing this process do the reverse. That is, they are noisy at start-up, then quiet down when warmed up. Over time their initial noisy period becomes longer, etc. The difference between that and your experience is in the metal used for the bearing sleeve. If that metal sleeve expands on heating more than the shaft does, the clearance becomes larger and looser as it runs, producing noise when warmed up. If the bearing metal expands less than the shaft on heating, the clearance is loose and noisy at start-up but gets less and goes quiet when warmed. But it's the same process, and the same ultimate solution.

 

[Paperdoc]

I suspect the reason for ability to control is in a quirk of many MSI mobos. Like most boards these days, its fan headers all can be set to use either the older 3-pin Voltage Control Mode (aka DC Mode), or the newer 4-pin PWM Mode. Unlike many, MSI often sets the CPU_FAN header to use PWM Mode by default, and the SYS_FAN header to use DC Mode. So if you made no attempt to change those, that is how they are now. For 3-pin fans, the ONLY way to control their speed is to use the older DC Mode, which is what the SYS_FAN header is doing. But with the CPU_FAN header set to the new PWM Mode, it cannot control a 3-pin fans' speed. Change that option for the CPU_FAN header to DC Mode (or Voltage Mode).

What you are experiencing now appears to be random. That is normal for the early stages of fan bearing wear. It will settle into a more predictable pattern later. For now you do not need to rush to replace the fan - it likely will last months or even over a year before getting really bad. But if you WANT to replace it now to get away from this issue, you can.