Nvidia fan control issue, hall event duplicator work around?

I swapped out my cooler and the vbios is confused as how to handle the new fan, even when using a custom fan curve it seems to panic during idle at times and ramp the fan speed up
it seems the vBios controls the fan PWM % based on the fans current RPM relative to what it expects instead of doing the sensible thing and setting the PWM % based on temperature curve
well the stock fan was a 3000 RPM fan and my new cooler uses a 1500 RPM fan, so if there is some way i can generate 2 pules for each 1 pulse the vbios will see a RPM in area it expects to see and hopefully not panic.
any ideas on how i can make trick the card is to thinking a fan is spinning at 2x speed?
this would be so easy if you could just alter the bios on 10 series cards...

That's a tough one. Are you saying there no modded bioses available for that card or are you saying you just don't have the expertise to flash one?

I have flashed NVIDIA bios before, but as far as i am aware there are no public tools for altering 10 series bios roms, i have my old cards bios before, just to set a custom 'factory OC' case at the time you could not OC nvidia on linux
where would i even check to see if there is a public modded BIOS?

the way it is now i can only set the software from 37 to 52% speed which at the hardware level that is 73% to 100% and to top it off it will randomly decided there is a problem at jump to 100%

this is my card: GTX 1060 3GB ZT-P10610A-10L i would expect the bios to be the same as you would get on a FE card, but i can upload it here if need be (max power target is 116%, the card refuses to use more than 121W when benching cause of vRel)
this card uses a 4+1 phase VRM controlled by a UP9511
edit:
card BIOS: https://www.techpowerup.com/vgabios/193332/zotac-gtx1060-3072-170606
UUID: GPU-f1e3f4c3-5c37-f829-ade1-7049e6321bec

You're misunderstanding how PWM works. The pulse width likely has a resolution of 256 bits, it's a series of off and on states that tells the dc motor in your foan to turn on and off at a rapid rate. Always full on, but switched at a frequency of 16-28khz per second depending on the fan. The PWM signal your card is generating is likely fine. What is giving you a problem is the RPM reporting back. Your graphics card is expecting more or less pulses per revolution. That you would need to alter. You could connect inline with the RPM lead from your cable a simple IC that doubles the frequency it receives. You could likely do something like that with a PLL.

https://www.digikey.com/product-det...MIvaeu09DG4AIVk4jICh1d8ATTEAQYASABEgIujPD_BwE

One of these could be used to generate whatever multiple you wanted from the RPM signal.

I am not misunderstanding how it works, it may have been a poor choice of words, what you are suggesting is what i was asking for, some way to make the card think the fan is spinning at double the pulses so the cards thinks the RPMs are the same as the stock fan

Okay, my apologies then. I updated the post with a link.

The device isn't that difficult to set up from what the datasheet shows. You could easily breadboard it for testing and then use one of those hobby boards that you cut traces on to finish the circuit.

I would recommend you put two fan headers on it. Use the fan header that goes to the motherboard to source your power and ground, run the leads between the two headers. Have the pwm signal go from one header to the other as well. The only signal that gets routed out is the fan signal, to the IC, then from the IC output to the new fan header.

If you're not real good soldering, do yourself a favor and solder a DIP socket onto your board for the IC. That way it's just plug and play, no risk of burning the chip.

This video explains the concept pretty well.

I am apparently not that bad at soldering, but i would prefer to use through hole parts, the tools i have are not good enough for smt stuff
i have most of the parts i needs aide from the chip you mentioned, though i did look up a 555 timer and i could build it from my resistor and transistors, but the sake of it low profile better not...
hopefully i can figure out the circuit i need to wire

If you have a time a circuit diagram would be very appreciated

I was wrong in my original post. 555 can't do the job. The IC I linked is cheap and PDIP so it's through hole. I still recommend you solder a dip socket to the pcb. You could coat the PCB in Type 2 non acidic silicon afterward if you were worried about short circuits. The datasheet is a little complicated, but dig down and read it and it tells you everything about how to set up the chip.

but that IC can do the job, i will try to read the sheet
too bad it is just a scan or the original, just makes it hard to look at

Yeah, it was awful to read. That is an older ic, but as far as I can read it handles lower frequencies. Worst case scenario you’re just out a dollar if it doesn’t work on your breadboard.

pqwoerituytruei said:

but that IC can do the job, i will try to read the sheet
too bad it is just a scan or the original, just makes it hard to look at

Click to expand...

Check this out

https://pdfserv.maximintegrated.com/en/an/AN3327.pdf

Much simpler.

—— correction, simpler but won’t work as it is for fixed signal. Sorry. You’re stuck using a PLL or a micro controller.

https://components101.com/ics/lm565-pll-pinout-datasheet

This one looks easier to use.

to be honest i am not sure what i am doing when i wire that up, the power pins are obvious, it has 2 inputs, i assume one would be the fan sense and the other is NC and a i guess reference output goes to the fan sens on the gpu
no idea what type of cap or what size i would need to use for the timing cap nor do i know what size resistor to use on for the timing resistor and rather it needs to be pull up or pull down, as for pins 5 and 7 do i even need them?

- - - Auto-Merged Double Post - - -

Would anyone happen to know what the card would do if there was no fan sense input? i would guess it sees 0 rpm and runs at 100%, refuses to allow the card to run, or sets the pwm duty cycle to the desired fan speed %

The new one I linked is easier to understand. The link to the data sheet at the bottom is where you should go. On page 7 of the data sheet is an example circuit for 10x multiplication. I would assume it would be easy to adjust that. I *think* it would be as easy as switching from QD to QA pins on the decade counter so it reads a pulse every other input. I’ve never used one before though. Once again, someone needs to breadboard this to test.

I thought i posed this yesterday...
i think i understand how to use those 2 parts, but it seems they are out of stock on digikey and mouser as well as obsolete...

I’ve been wracking my brain on this one and searching the net.

Maybe one of this guys sinpler circuits would suffice. They might not be super accurate, but they would likely auffice for your purposes.

My last thought would be to use an RC filter to pass your hal sensor data to a gpio adv pin on a micro controller then compare that to the known max rom of your fan, and adjust the pwm output module to match. It could be updated at a frequency high enough that the motherboard never sees lag just using the internal oscillator of the chip. The problem is, I don’t have proteus anymore so I don’t want to be programming and simulating in real life. It’s much more difficult to debug a chip on the fly in real life than in a virtual simulation module like proteus. What I’m describing is basically what a PLL does. You set the center frequency of the oscillator using the required caps and resistors, etc. Then the square wave you feed it is passed through and RC filter to make it a voltage. The voltage controlled oscillator in the chip produces a square wave to match the center frequency, offset by how far it is from center voltage. A divider is used between the square wave input and the comparator input for the VCO to essentially double the frequency generated.

I just wish I could find a pll that was simple, cheap, and designed to work at low frequencies. Maybe someone who knows more than me will chime in. I’m a hobbyist.

- - - Auto-Merged Double Post - - -

Have you though about just plugging that fan into a spare header on the motherboard and using Speedfan to read gpu temp and control said fan?

my board only has 2 pwm headers... one for the cpu and one for the fan i put on the vrm
"sinpler circuits"? i assume you are referring to a forum, but assuming this is a typo, google is not helping me

Yeah that was a typo. Was supposed to be “simpler” but i was tired and forgot to paste the link I had found. I’m thinking about a simple circuit. Maybe an RC filter that drives a voltage based square wave frequency generator. Crank your fan to full blast on the bench, use a multimeter to measure what voltage comes out of the RC filter. Then set that as the max frequency voltage for the signal generator.

So i can give the fan sense a PWM square wave in, what is the switching frequency i need to be in spec?