Nvidia fan control issue, hall event duplicator work around?

[Jasonaaronfox]

Depends on your fan, or more importantly on your old fan. It produces X pulses per second based on each complete rotation. You could open up the old fan and see how many magnets it has. Each magnet is one pulse per rotation as far as I understand it. Four magnets == 4 pulses. #pulsesPerSecond * 60 / numMagnets == RPM. 3000 rpm / 60 seconds would equal 50 rotations times 4 pulses, which would give you a square wave frequency of 200htz for full fan speed. This all changes based on how many magnets your original fan had. If you have access to an oscilloscope you could just turn the old fan on 100% and read the frequency the tach puts out. A 555 could be used to generate your square wave. There are also super cheap square wave generators on eBay. This solution would trick it into thinking all was well, but I’m not sure it would still control the fan right. You would need a pll for that.

One other option though. The old fan. Wire it up to a separate power source, molex or something. Use a pwm splitter to send the pwm signal to BOTH fans. Cut the blades off the old fan. Feed the tach signal from your now bladeless old fan back to the gpu. Instant solution. The old fan will buzz along somost silent with no blades. Just mount it somewhere.

 

[pqwoerituytruei]

I would guess it has 2 magnets as i think that is standard and the software reports it's rpm accurately
so if i use a 555 timer what else would i in theory need to use
the sense pin 0.55v on the stock fan at 100%
stock fan gives me 59.9 Hz at full speed
fan was given +12v via wall wort
fan seance was put on a DMM and the DMMs ground lead was on the wall worts +0v line
getting 60 hrz seems a bit low, i wonder if i somehow was reading mains power frequency?

 

[Jasonaaronfox]

Might not have had the resolution to see the frequency correctly. Really need an oscilloscope. 2 magnets and 3000 rpm should have been 100htz.

I posted about a cheap eBay square wave generator I bought. It’s adjustable from 1htz to 150khtz and less than $10.

https://www.overclockers.com/forums...ng-PWM-Controller-for-4-wires-PWM-fan/page115

You could just grab that and play around with it until you find the right frequency.

 

[pqwoerituytruei]

I did find this:
https://pdfserv.maximintegrated.com/en/an/AN3327.pdf
Assuming i found something useful do you think you could find parts that are stocked that can do that?
this seems to be geared toward much higher frequencies than this would involve
i would very much like to avoid smt parts, i would much rather use through hole parts
edit: this seems to be the TLL chip used in the file: MAX9010EXT+T

 

[Jasonaaronfox]

My concern with that circuit is that I *think* it is for a fixed frequency doubler. Yours would be dynamic, having a frequency range instead of a set number of beats. I could be wrong though, PLL and all this are new to me. I'm more familiar with microcontrollers.

 

[pqwoerituytruei]

for the meantime i am using a 0/5/7/12v switch for the fan
on another topic, but pll related, would you know of a cheap circuit for a fixed 1kHz to 25 kHz, id be using 1% increment duty cycles if that matters

 

[Jasonaaronfox]

It’d hard to hear the price on eBay for that square wave generator I linked to in the 4 wire pwm thread above. Less than $5 and you can choose frequency and duty cycle.

Search eBay for https://rover.ebay.com/rover/0/0/0?mpre=https://www.ebay.com/ulk/itm/273645141665

 

[Brando]

what about the quadro fan controller? it was recommended to me but i ended up finding another way. if you don't have enough pwm headers to use with argus monitor you could hook up a quadro and use it to control your whole case including video (i think) https://www.aquatuning.us/air-cooli...uacomputer-quadro-fan-controller-for-pwm-fans