hack to fool an RPM sensor?

jamesavery22 · Post by **jamesavery22** » Thu May 17, 2007 7:18 am

I used to have a vague idea on how the RPM sensor read the lead from a 3connector fan but now I have no idea...

I built a bottom budget PC around a lenovo 87H4658 motherboard. While Lenovo/IBM is putting updates out for the BIOS there still are no options to turn off the CPU fan sensor. I have two fans in the case, one is in the PSU and the other is an old 92mm panaflow. At 12v its quieter then a nexus. I bought these years ago. Great fans and very quiet even though they are ball bearing. Anyways... Thats the only fan I have in the case and it doesnt have a sensor lead. Well it does but its fake, internally(to the fan) its connected to nothing.

Can I just ground that out to make the bios think its spinning at 10k rpm or something? I just want to shut up the damn bios warning that pops up.

ToasterIQ2000 · Post by **ToasterIQ2000** » Sun May 20, 2007 8:56 am

The RPM signal on the third fan wire is a series of pulses. The frequency of the pulses ( peaks of 5-12 volts at ~ 50 mA ) indicates the fan's RPM. I recall 2 pulses per revolution is a common conversion factor.

Grounding or hotwiring the dead end third wire would not generate pulses. Lol, prolly a bunch of exposed circuits all over a PC that coul be sucked for such a signal ... maybe even without the voltage / amperage drain sabatoging at the tap ...

Since the Panaflow in your case has an unused third wire and your PSU has a fan, you might try pulling the fan wire from the Panaflow and poking it into the back of the plug on the PSU fan sensor wire and securing it there -- presuming that the PSU fan is sending an RPM signal itself. Some fans that are two wire have the third RPM contact inside them without a wire on it too; fairly delicate soldering to use that though.

BE VERY CAREFUL HERE: aside from voiding the warranty on the PSU by poping it's cover and exposing big PSU heatsinks that can be dangerously electrically live, if you connect the Panaflow's 50 mA sensor lead to the PSU's 12v 1.? amp hot leg you would most certainly fry something in the motherboard's fan header -- like litle puff of smoke, black soot spot on toasted motherbard resistor or diode, unknown cascade of dammage in mobo ... !

If you are familiar with this type of circuitry and have a multimeter that can detect 5-12v pulsed frequeny logic signals to verify your connections first, you can see if splitting the PSU RPM signal would work for you. If this is far from what you are familiar and confident with please do not risk voiding warranties. shocking yourself, or or blowing your mobo on my account

For the price / risk / time involved: get a RPM enabled case fan.

BrianE · Post by **BrianE** » Sun May 20, 2007 11:17 am

I don't suppose you could turn off or disable that fan header in the BIOS? Then you could run the fan off a 4-pin connector. (I know, that's probably too easy...)

I did a quick search but not much came up. A company once sold a gadget to double the RPM signal, but they're out of business now....

You could try reading through this older thread and see if there is anything you can use or modify to double the output:
viewtopic.php?t=3565

cpemma · Post by **cpemma** » Mon May 21, 2007 2:09 pm

jamesavery22 wrote:... there are no options to turn off the CPU fan sensor ( in BIOS)

So that eliminates that.

I have two fans in the case, one is in the PSU and the other is an old 92mm Panaflo.

Is the Panaflo on the heatsink? That's the one causing the BIOS warning?

I just want to shut up the damn bios warning that pops up.

The easy way is to swap the fan to one with a speed wire, though with large slow fans some older BIOSes will still think the fan has stopped. My ASUS BIOS, for example, doesn't "see" cpu fans under 1000RPM.

For the hard way you'd need to build an oscillator (555 chip or similar) running from a molex 5V or 12V, switching an open-collector NPN transistor to send a fake signal to the motherboard. The signal is actually 2 'low' pulses every rev, so an oscillator frequency of 50Hz is equivalent to 1500RPM.

Bluefront · Post by **Bluefront** » Mon May 21, 2007 3:23 pm

Yeah those Asus boards need to see at least 1000rpm, to avoid the pop-up. Here's how I fixed mine. I wanted to run one fan.....a Yate Loon that booted up much less than 1000rpm. When I couldn't get rid of the pop-up, I tricked the thing by attaching a 60mm fan, under-volted to 5V using a Fanmate, to the CPU header. The 60mm runs about 1200rpms, and located inside the case, is inaudible. I attached the Yate Loon to the other MB header. Problem solved.

FWIW.....I've got one of those "RPM doublers". But I never tried it and want to hang on to it. I think it was made by the same company that makes the NoiseMagic controllers.....also hard to find in the USA, but available in Europe (I think).

cpemma · Post by **cpemma** » Mon May 21, 2007 4:21 pm

Bluefront wrote:Yeah those Asus boards need to see at least 1000rpm, to avoid the pop-up.

My A7N8X 'talks' to me ("CPU Fan Failure!") so I changed the sound file, now it says "Hello!" every boot instead.

wildlaurel · Post by **wildlaurel** » Sat May 26, 2007 2:54 am

i tried to measure the rpms from the yellow-common wire combination of the fan but it just shows me 4V on dc scale, while zero on frequency count.. does anybody got idea why the fan wouldnt give a frequency count on rpm sensor.. m using Fluke 189 DMM to measure the parameters..

Mr Evil · Post by **Mr Evil** » Sat May 26, 2007 3:16 am

wildlaurel wrote:i tried to measure the rpms from the yellow-common wire combination of the fan but it just shows me 4V on dc scale, while zero on frequency count.. does anybody got idea why the fan wouldnt give a frequency count on rpm sensor.. m using Fluke 189 DMM to measure the parameters..

The RPM output needs a pull-up resistor on it. You can either connect it to the motherboard and measure the signal from there, or connect your own pullup resistor to it. I don't know what sort of value fans expect, but 10K will probably do, connected between the RPM wire and +5V.

cpemma · Post by **cpemma** » Sat May 26, 2007 4:43 am

Mr Evil wrote:...I don't know what sort of value fans expect, but 10K will probably do, connected between the RPM wire and +5V.

As the meter probably has an input resistance of several megohms, any pull-up in the range 1k-100k will do the trick and it can be pulled-up to 12V or any other lesser level if you're doing this away from the motherboard. The 12mA @12V with a 1k pull-up won't hurt the fan's open-collector transistor sending the signal.

With a simple DC voltage reading the number you get is just the pull-up supply multiplied by the duty-cycle of the pulse train. Nothing to do with RPM. You'll need a suitable frequency scale and may need to adjust the input voltage range to suit the meter - I think mine is 1V max on frequency. Then remember there are 2 pulses per rev, so halve the result.

wildlaurel · Post by **wildlaurel** » Sat May 26, 2007 7:07 am

i thought the DMM would pick up the slightest of the signals.. but i think you ppl are right..

so ya.. i added a pull up res of 30k from a 24V supply (the same 24V fan supply) to the sensor terminal.. previously it gave 4.3Vdc, now it gives 4.8.. but a null on the frequency.. i have two similar fans and they both give the same readings..

ToasterIQ2000 · Post by **ToasterIQ2000** » Sat May 26, 2007 9:18 am

wildlaurel wrote:i tried to measure the rpms from the yellow-common wire combination of the fan but it just shows me 4V on dc scale, while zero on frequency count.. does anybody got idea why the fan wouldnt give a frequency count on rpm sensor.. m using Fluke 189 DMM to measure the parameters..

I measure pulses with a Fluke 83. I do it rarely enough that I check the meter's manual before doing it--plug the leads into wich terminals? Set the dial on which measure? Cycle to which measure and wich range?. I recall Measuring with the fluke did not require an additional pull-up resistor. I'm away from that toolbox atm.

I'm still tickled by the idea of taking an old PC waititg for a trip to electronics disposal / recycling and probing for any exposed contact on the motherboard that might be tapped for a fake fan signal.

cpemma · Post by **cpemma** » Sat May 26, 2007 11:20 am

ToasterIQ2000 wrote:I recall Measuring with the fluke did not require an additional pull-up resistor.

Some fans have a built-in pull-up, some don't. YMMV.

wildlaurel · Post by **wildlaurel** » Sat May 26, 2007 10:51 pm

i tried with both dc and ac frequency counts.. didnt work. i have measured freq (power line and rf) with my meter before so i know theres no problem with my instrument..

hack to fool an RPM sensor?

hack to fool an RPM sensor?

Re: hack to fool an RPM sensor?