Sign-in / Sign-up
Your question

Suddenly having overheating problem

Tags:
  • Homebuilt
  • Systems
Last response: in Systems
a b B Homebuilt system
October 10, 2010 9:58:44 AM

Am currently getting idle temperatures in the mid/high 40s and up to 62-65C during even moderate gaming (Sims 3, Starcraft 2). This was not happening until about a week ago, which I will explain below. System is a q9550 with a standard mid tower case, 2x80mm fans on the side (Thermaltake Smartfan, adjustable up to 70CFM apiece), 1x80mm each on the front (up to 50CFM, set at 35CFM) and back (35CFM). CPU fan is a Thermaltake Silent 775. Normally, it idles in the low 30s with the fans on low and gets up to the mid-40s in heavy gaming; highest temp I think I ever saw until this month was 51C, and that was on a hot day.

About a week and a half ago, one of my side fans stopped working entirely, which I figured out was because my manual fan controller on the front panel was damaged (the baby had gone up and twisted one of the knobs too hard and ruined it). I plugged that fan into a different port on the same controller, and both work fine now. However, at the same time, the CPU fan began constantly spinning at max RPMs no matter what the temperature was, making a hell of a lot of noise. The system would still idle in the high 30s or low 40s, but the CPU fan was intolerably loud and the motherboard had apparently forgotten how to adjust the speed according to temperature.. Tried Speedfan and a couple of other software fixes to control the CPU fan, but nothing -- it continued to spin. Eventually did some research on my motherboard (Asus P5E3 Pro) and found there was a BIOS setting called "Q-Fan" that lets you set the mobo's CPU fan control preferences. No help at all -- if Q-Fan is enabled, the CPU fan spins at the minimum 680-700 rpm no matter what, and the temperature skyrockets during gaming. Turning both side fans to max (for a total 140CFM and the noise level of a jet engine) does not help the temp by more than 1-2 degrees.

Last time I applied new thermal paste was less than a year ago. Last time I blew out the dust from the case was about a week ago. All the connections are solid. My fear is that the CPU fan control on the motherboard has crapped out and stopped working, or else the speed control element in the CPU fan itself is screwed up. One fix I have considered is getting a new front panel fan controller that can deal with the CPU fan ... but no guarantee that will work if the fan's control element is on the fritz. Otherwise, the options are replacing the CPU fan (a huge pain in the ass to do in my case) or replacing the motherboard. My first guess is that the mobo fan control is bonkers, not the fan's element, given that the fan can still run at more than one speed. I am not very happy about this, and if anyone has experienced a similar problem, I am all ears.

More about : suddenly overheating problem

a b B Homebuilt system
October 10, 2010 4:14:50 PM

Yeah ... this HSF came with those crappy push pins, but they were so terrible that I immediately removed them and fastened it with nylon screws and nuts. It's on there tight. I also ditched the stock thermal paste and went to Arctic Silver since I first built the machine. My main concern is that the fan seems to ONLY spin at minimum rpm (if the dynamic fan control feature is turned on) or max rpm (if dynamic fan control is turned off).

Oh yeah -- just to clarify, the CPU fan and front/rear fans are not on the manual fan controller, just the side fans. CPU and front/rear are plugged directly into the motherboard.
m
0
l
Related resources
a b B Homebuilt system
October 13, 2010 6:16:16 AM

I'm just going to bump this one more time in case anyone has an opinion.

Basically, it looks like I'm going to be spending about $30 if I want to fix this problem, and looking for guesses as to whether I'd be better off spending it on a new HSF, or on a fan controller that can automatically control the CPU fan.

Seems like either the mobo's fan controller is on the fritz, or else the control element in the HSF itself. Maybe I should flip a coin.
m
0
l
a b B Homebuilt system
October 14, 2010 3:46:56 PM

The symptoms you describe amount to: if you allow the CPU fan to run at full speed all the time, it can keep the CPU temperature in an acceptable range, and yet the CPU temp is still higher than "normal". This is NOT a failure of the CPU fan control system. It is a failure to remove heat. Despite full fan work, it is unable to remove heat as quickly as it should, and so the CPU temp is higher than desired.

I am confused by your story a bit. You describe the current situation as abnormal, and expect that the CPU will run at acceptable temperatures with low CPU fan speed. Then you say you only recently discovered the Q-Fan system in BIOS. Well, how was your CPU fan speed NOT running full speed all the time if you never had Q-Fan in operation?

Look for more details of the fan operation. In some systems the BIOS Setup screens include in the System Health area current details like actual measured CPU Temp, CPU fan speed, and maybe some settings for the CPU Temp control loop. In other systems the BIOS does not do this, and you have to install and use a software application that came on the CD with your mobo. I suspect you have much of this already since you can quote some CPU temps and fan speeds. What I don't see, though, is how the temperature controller is set. On my ASUS mobo (older and different) for both the CPU Fan and the SYS Fan control loops I can set the target measured temperature, the low temperature at which the fan first turns on, the high temperature at which the fan should reach full speed, the warning point at which the CPU should slow down because it is too warm, and the limit temp at which the CPU will shut down completely to prevent burning up at a dangerous temperature. You may not have all those parameters. But what you describe suggests that your Q-Fan system for the CPU is set to maintain too high a CPU temperature. Or maybe it is set for temperatures that are "typical" for your CPU, but you have just been accustomed to seeing much better. Either way, look for a way to specify the target CPU temp and the higher temp at which the fans reach full speed.

While you are at it, consider the possibility that the CPU cooling fan itself is wearing out. A fan's bearing can get worn with the result that it makes much more noise than when new, and may run a bit slower for the same supply voltage. Any chance the high noise level you hear now is being generated by the fan running at the same speed as always, but with worn bearings? Any chance the "full speed" of the fan now is slower than it should be?

You need to examine why heat removal is degraded. If the fan itself is OK, look closely for loose heatsink mountings (I know you say they seem OK). If necessary, remove the heatsink, clean it off and replace the Arctic Silver again. Be SURE to read AS's website instructions on how to install it. Too much is just as bad as too little.
m
0
l
a b B Homebuilt system
October 15, 2010 6:52:16 AM

Paperdoc said:
The symptoms you describe amount to: if you allow the CPU fan to run at full speed all the time, it can keep the CPU temperature in an acceptable range, and yet the CPU temp is still higher than "normal". This is NOT a failure of the CPU fan control system. It is a failure to remove heat. Despite full fan work, it is unable to remove heat as quickly as it should, and so the CPU temp is higher than desired.

I am confused by your story a bit. You describe the current situation as abnormal, and expect that the CPU will run at acceptable temperatures with low CPU fan speed. Then you say you only recently discovered the Q-Fan system in BIOS. Well, how was your CPU fan speed NOT running full speed all the time if you never had Q-Fan in operation?


From what I can gather, Q-fan is basically the same idea as temperature targets, only they substitute "performance levels." The setting right next to Q-Fan is chassis fan temp targets, for what it's worth. But with Q-Fan disabled, the board will still adjust CPU fan speed according to some built-in default standards. That's exactly what it was doing originally.

In other words, with Q-Fan disabled, it was spinning at low rpm normally and ramping up when needed, like it's supposed to. Now, it either spins at max no matter what and stays cool (with Q-Fan off), or minimum no matter what and gets hot (with Q-fan on any setting). I don't expect it to stay cool when the CPU fan is spinning at minimum rpms -- what I do expect is that it should be able to adjust the RPMs under one option or the other, which it ain't doing.

To top it off, nothing at all changed in terms of things that should affect heat or airflow. I didn't add any components, change any settings, or move anything around inside the case.

Before:
- Q-Fan disabled
- CPU fan running at 700 rpm idle, ramping up to 2600 rpm in bursts (like it's supposed to)
- CPU idling in 30s, maxing out in 40s

Now:
- Q-Fan disabled
- CPU fan spinning at 2600 rpm constantly
- CPU idling in 30s, maxing out in 40s

OR

- Q-Fan enabled
- CPU fan spins at 600 rpm constantly (regardless of Q-fan control setting)
- CPU idles in 40s, maxes out ~60C


You know, the more I think about it, the more I'm convinced that the motherboard has simply lost its ability to control the CPU fan correctly. The fact that this happened at the same exact time that a case fan offed itself (and burned out a switch on the manual fan controller), makes me suspicious that there was some incident (power-related?) that affected both. I am now getting a reading of 11.41V on my PSU's 12V rail, which is right on the edge of suspect -- though that's only with software. I suppose it's time to bust out the multimeter for that one. And maybe find a fan controller that can handle a CPU fan, too, if the mobo's controller is fried.
m
0
l