Question about a fan.

[airplaneman]

Hey, I have a fan that uses a temperature sensor to control the speed but I don't really know where to put it so that the fan will actually spin at a decent speed. Is there any way to use a program to control the fan? Speedfan says I have a system fan spinning at around 900 rpm but I don't know if that is my side panel fan or the temp sensor fan. If someone could help me out it would be great.

 

[Billb2]

Disconnect the fans one by one and see which reading goes away.

And it's not a good idea to use a temperature controlled fan on a fan header where the speed of the fan on that header is comtrolled by the motherboard. You would have two things trying to control the fan at the same time.

 

[airplaneman]

Yeah, figured out which fan it is thanks.

I don't really have much of a choice as to where I put the fan...but I don't want it being controlled by the temperature sensor, but I don't want to cut the sensor off unless I have a sure-fire way of getting something else to control the fan speed..

 

[mindless1]

Generally a fan with integral temp sensor is at max RPM when it's sensor is at max conduction. It's at minimum RPM when at least conduction. You can effect a change in the fan's behavior regardless of the motherboard circuit by cutting off or desoldering the sensor for lowest RPM (probably way too low to be used except straight from 12V and it may pulse from such low RPM), bridging the sensor leads, or putting a experimentally determined value of resistor in parallel with the thermal sensor. Experiment can mean trying a few or it can mean measuring changes in the sensor as you change temps and mathematically arriving at the paralleled resistor value you want to increase the RPM such that the motherboard fan control circuit has more control over fan RPM than the fan's temp sensor does.

I don't understand what you mean by having a sure-fire way of getting something else to control the fan speed. Is the motherboard or a separate fan controller not now effecting any change in fan operation?

If all else fails, fans are about $5 each. You could spend an hour on this or mow somebody's lawn for a profit after buying another fan when all is said and done... some things just aren't worth the bother except as a learning exercise.

Here's another thought - The ideal some have that a fan should speed up so a system stays at some constant temperature is usually an unnecessary idealism that ignores the critical details of system function. It is fine if parts get warmer and cooler as their activity level changes, so long as they don't go above a threshold value that effects stability or acceptable lifespan. Point being, unless you have SLI'd video cards it is not necessary important to change any fan speed except that of the CPU, as 120mm case fans are more than ample for daily computing heat levels in moderate ambient conditions. Perhaps yours is not moderate but if not, more specifics of your exact goal, system specs and the condition you need to counter might be useful.

I may have drifted off on a tangent. To know where to put the fan or it's sensor if on leads, to have the right response will depend on how the fan sensor circuit is calibrated. You'll need to know it's range vs RPM to know if it is useful at all at typical PC ambient temps (many fans aren't designed to use very useful ranges, are set to require pretty high temps before they increase RPM significantly enough). Generally if the motherboard has fan control on an available fan header, and the fan is compatible (4 pin PWM control is sketchy controlling 3 pin fans for example) your best bet is to set the fan at a max steady RPM and let the motherboard or separate fan controller take over it's speed control.

You don't want to use a program to do this, programs are subject to OS or the program itself crashing, which can create a busy loop that actually causes higher temperatures than the last setting the program used to set a low fan RPM. The cooling system should be independent of the software running, if you leave a system just sitting in the bios menu for example the fan should either respond accordingly or revert to full speed if not given continual commands through the control circuit to spin any slower.

 

[airplaneman]

Wow. So much information! Haha. You make a lot of good points and I think I'm just going to let the fan run at whatever temp it wants for now, then buy another fan in a little bit. Thanks.

 

[airplaneman]

One more question, what will happen if I cut off the temp sensor, then plug the fan into a fan controller? Would that work?

 

[mindless1]

If you cut off the sensor that causes minimal RPM. At that point you would want to power it directly from the 12V rail, it would run slow, probably too slow, possibly even failing to run at all until the system temp was too high.

If you instead want to have control, consider how the control circuit works, what it's limitations are. Bridging, soldering, etc, the thermal sensor leads causes max RPM when the fan control circuit tries to cause max RPM, but if a fan is especially high RPM without any control it may be too much noise, wear, or dust buildup for your needs. Therefore, unless you hand-pick an optimal fan (max speed per model), the ideal to reuse one you have is to put a resistor in parallel with the existing thermal sensor. Generally, the value of resistor needed is inbetween 1K and 10K, but that's just what is common, it's not a rule or anything. The thermal sensor feedback is a low current circuit, instead of a fixed resistor value you could make it adjustable by using a 20K POT instead, though depending on the thermal sensor used it might be more precise with a lower value POT.

In other words, the best course of action is to measure the thermal sensor resistance value at typical system heat levels (or open air will be close enough) and pick a POT that is about twice the value of the thermal sensor reading.

On the other hand you could just say screw it and run the fan at an RPM a little higher than what the system needs at it's lowest processing/heat levels, and accept it will get a few degrees warmer under load, so long as it doesn't get hot enough to be instable or effect lifespan beyond what you deem acceptable. The latter is what OEMs usually [Edit:] did, before they switched to single rear fans for CPU heatsink and exhaust controlled by the motherboard thermal sensor fan control circuit [/Edit], and manage to do so while being fairly quiet, but they have less cooling subsystem redundancy, less ability to o'c as far, upgrade as much, and they don't care what happens after the warranty ends.

 

[airplaneman]

I guess there is really no sense in me buying a resistor then, just to fix one fan haha. Thanks for all your help man.

 

[Zepper]

Generally the temp sensors on most fans that come with them are NTC (negative thermal coefficient) thermistors (as the temp of the thermistor rises, the resistance drops) in series with the fan voltage. I have on occasion jumped across the thermistor with a short piece of wire (tack soldered) that, in effect, removes the thermistor from the circuit . Then you can control the fan with any fan controller and it is easy to remove the jumper if you ever need the fan to control itself again. Cutting the thermistor leaves the fan open and it won't run, so don't do that until after you've soldered the jumper and are sure you'll never need the thermistor function down the road.

.bh.