CPU Fan issue

[kmadams23]

Hey,

I built a new system late last year (maybe December) and have had few problems with it except for one thing. It's been livable, so I've dallied about fixing it, but have some free time coming up so I thought I'd tinker with it. I actually believe I found the solution in a thread from a couple months ago (http://www.tomshardware.com/forum/280464-31-spinning).

Same motherboard, Intel i7 860 instead, and a Zalman fan (the 9900A LED). This fan requires BIOS be set to PWM - I didn't do that at first and it would tick and attempt to spin, but hardly move. I set to PWM, and now it just does nothing. Plugged in right to CPU_fan, so what gives? After no real helpful advice from either the mobo manufacturer or Zalman I decided to look elsewhere. Temperatures are still within a reasonable range (according to speedfan) but a bit on the high side and sometimes some quite hot spikes here and there as I game, so I'd rather the fan be spinning than not and keep things just a bit cooler.

Well, I stumbled upon that thread, tried the sys_fan, sure enough the fan spins happily away. So here's my question - should I leave it set up like that? Or should I set it back to the cpu fan and try to tinker with settings somehow? Using sysfan it seems to be on a pretty high speed setting, but maybe that's okay, just seems like an unnecessary use of power and noise contribution.

 

[Deleted member 362816]

If it works how you have it don't fix it.

 

[Paperdoc]

Am I correct in assuming that the fan actually is a 4-pin fan, and the CPU_FAN pinout on the mobo is 4-pin also? Further, I'm hoping the plug is put on correctly - it's hard to do it backwards!

Now, I am intrigued that you have a CPU cooling fan that is not running and it is not annoying you. Many mobo systems have a special setting in the CPU cooling fan control system that monitors the fan's speed looking for a complete fan failure (not turning, as you have). In that event it normally beeps a big warning and, in short time, shuts everything off to prevent an anticipated CPU melt-down. But just in case you have a reason for it, the BIOS allows you to tell the system to Ignore the CPU Fan speed and not bother with this protection system. Did you do that? Have you told your BIOS to ignore the non-functioning fan so it will let your machine run? (Actually, right now there is NO fan connected to the CPU_FAN pinout, working or not, so the mobo will see a zero speed signal on that port.)

IF you told it to ignore the fan, maybe that's why it is not making any attempt to feed it power. Check that the fan speed control is Enabled and Automatic, and of the 4-pin PWM variety. Then plug your CPU fan back into its proper mobo pinout and see if that works.

 

[kmadams23]

Yes, 4 pin to 4 pin header, and plugged in correctly - I don't think I'm even strong enough to break the tab and do it backwards!

No, I never told it to do that, but I think the shut off failsafe action is only for CPU temp, and the CPU never really reached "dangerous" temperatures, just ones which could cause a shorter CPU lifespan. The heatsink itself is HUGE, so that much copper seems to be taking enough heat of the CPU that it won't shut the system down.

I can't set bios to Automatic fan control - that was the default setting and does not work with this fan. I don't know what method the automatic was using to control the fan, but it would only move a centimeter and make a clicking noise and stop. I reread the manual for the part and it said I must use PWM setting for this fan, so I changed it to that - no longer failed attempts to spin, just no attempts.

As cia24 points out, it is working in a sysfan header so it doesn't really need fixed, I'm more just curious what could be causing this and inclined to tinker around to see if I can get this to work more "properly".

 

[Paperdoc]

Almost all automatic CPU cooling system controls in BIOS include functions to vary fan speed according to measured CPU temperature (a classic feedback control loop), including safety procedures if the CPU temp exceeds limit(s). For this, the temperature measurement sensor is built into the CPU case / chip and fed to the mobo thorough one of the chip pins. Some max temperature systems simply shut down the system when the measured CPU temp exceeds the limit; others have 2 limits - the lower one causes substantial clock speed reduction on the CPU to reduce its power consumption but keep running, and the upper one is shut down time. But in addition to those systems, SOME (not all) BIOS's also do a separate check on the speed signal coming from the CPU fan. If it detects that the fan is NOT turning at all, it does not wait for the CPU to overheat as a result of the loss of cooling. It simply puts out a loud warning signal and then shuts everything down immediately. Of course, this cannot work if there is no fan connected to the mobo CPU_FAN pinout header, or if there is something faulty in the fan speed signal system. So the BIOS's that have this additional feature allow you to turn it off if you choose. It sounds like your system does not have this - you do not report any wild alarm signals or any notice that the BIOS has options for such a feature.

It does sound like your CPU_FAN pinout is NOT preforming properly. On a 4-pin fan with its pinout configured to provide PWM control, the wires to the fan are Ground and + 12 VDC power supply, a fan speed pulse signal (2 pulses per revolution) coming back from the fan motor, and the Pulse Width Modulation signal. This latter signal is a type of "square wave" except that it is not a plain "square". It has variable "% time on". It is used by a small control circuit inside the fan to control how much power (based on time) from the +12 VDC supply is actually fed to the motor, and this controls fan speed. A motor designed for this system cannot operate solely on the +12 VDC supply without the PWM control signal. If you have connected an older 3-pin fan to the 4-pin port, you tell the BIOS that you have a 3-pin fan there and it arranges to ignore the PWM signal on the unused pin and supply a varying +VDC to set fan speed. If you have the BIOS set to 4-pin PWM mode instead, it sends the full +12 VDC out on that line, plus the PWM control signal on its fourth line so the fan can do the final speed control. That is why you must set the BIOS to PWM mode to use a 4-pin PWM fan on the port. But apparently that does NOT work on your mobo, even though the fan performs well if connected to another 4-pin SYS_FAN pinout. That seems to indicate the fan is OK, but the CPU_FAN output port is not.

The only problem you have here with the substitute setup you have been forced to use is that the automatic fan speed control is now being driven by a temperature measured somewhere on the mobo with a sensor, and NOT by the temperature measured by the sensor built into the CPU case itself. Although there may be a correlation between these two temperatures under normal circumstances, you do NOT have the full protection systems intended. The CPU fan speed will be changed as the whole case interior heats up, but the response to quick changes in actual CPU temperature will not happen. However, I expect the CPU max temp safety limit system is still operating - that has nothing to do with fan operation.