4
\$\begingroup\$

I have a power supply that uses a PWM fan with 4 wires.

Wires: 12V (red), GND (black), PWM Control (brown), TACH Sig (white)

The fan speed is controlled by the PWM control wire (brown). Fan speed is duty cycle, so 50% means 50% fan speed. Max speed is 22,000 RPM.

Tach signal (white wire) is an open collector with two pulses per revolution.

At full speed we should have 44,000 pulses on the Tach signal, or 733 Hz.

The power supply checks if the fan works, so using another fan or removing the fan altogether is not working. It also seems to check if the fan speed is anywhere near the commanded fan speed.

Here is the datasheet of the fan: http://www.nmbtc.com/pdf/dcfans/1611ft.pdf

How would one build a very simple circuit that simulates the fan?

I imagine a small circuit that gets the PWM signal, and depending an duty cycle outputs the open collector pulses. ie. 0% duty cycle 0Hz, 100% duty cycle 733 Hz.

I have seen this question asked again and again, so coming up with a small circuitry would be very beneficial not only to my own small project.

Edit: I'd like to add that the fan needs to go because a far superior way of cooling will be employed for this power supply. I am not trying to find a cheap solution not to replace a fan or save some dollars. Power supply and its components are sufficiently cooled. I just need the fan on the PSU itself to be gone.

\$\endgroup\$
4
  • \$\begingroup\$ Simple: Get a different power supply, or use a fan! But snarkyness aside, there is probably a very good reason why the power supply wants a fan and you haven't given us any indication that you have addressed that. \$\endgroup\$
    – user3624
    May 5, 2013 at 0:13
  • \$\begingroup\$ It could be an interesting question if the "small project" was explained. Otherwise it sounds like a cheapskate fix to make a power supply work until it overheats. @DavidKessner snarkyness is a good word - I haven't looked it up but, it sounds like what it should mean. \$\endgroup\$
    – Andy aka
    May 5, 2013 at 0:18
  • 2
    \$\begingroup\$ Thanks guys and apologies for not explaining. I am building a system that is cooled in another way (immersion in dielectric fluid). The power supplies I have used so far have just issued a warning (beep, blinking LED) but kept working after the fan was removed. But this particular PSU simply switches itself off if the fan is not present. This has nothing to do with overheating, as the system and PSU is sufficiently cooled (= better than with the built in fan). \$\endgroup\$
    – Alex
    May 5, 2013 at 2:01
  • \$\begingroup\$ here is a circuit you may find resolves what your looking for: techidiots.net/notes/fake-fan-sensor \$\endgroup\$
    – Raiden
    May 14, 2018 at 11:29

2 Answers 2

Reset to default
4
\$\begingroup\$

By far the lowest component count solution to the problem of simulating the TACH signal from a fan is to use a small microcontroller (MCU) that can measure the PWM input signal duty cycle. The PWM signal can be measured two ways with the MCU either by using a timer to directly measure the high / low pulse widths OR by filtering the PWM via an R/C circuit to produce an analogue level that you read into the MCU via an A/D converter channel. The output TACH signal can be generated easily using an MCU timer that is capable of driving a pin. In fact I have even done TACH signal simulation using a timer that generated a periodic interrupt and then let the MCU software toggle an I/O pin in the interrupt service routine.

\$\endgroup\$
7
  • \$\begingroup\$ Are you sure there is no easier way (ie. passive circuitry without MCU), for instance something like using a 555-timer IC? en.wikipedia.org/wiki/555_timer_IC My home is on the thermal side, I have general electronics knowledge, but building a system (especially on the weekend) with a MCU is definitely beyond my capabilities. \$\endgroup\$
    – Alex
    May 5, 2013 at 2:04
  • \$\begingroup\$ @Alex - I did not say the MCU approach was the easiest way. It would be the lowest component count way to do it. In addition to that it offers a more exact way to translate the PWM value to a corresponding TACH frequency. \$\endgroup\$
    – Michael Karas
    May 5, 2013 at 2:13
  • 1
    \$\begingroup\$ @Alex - The analog way using a batch of descrete components is certainly a fully possible way to build the TACH generator. You would have to convert the PWM to an analog voltage and then use that to set the frequency of a voltage controlled oscillator (VCO) that in turn becomes the TACH signal. Be aware that PWM filtering can be fairly linear. On the other hand VCOs are often not linear in their V-F transfer function. Note that fans are also not normally linear in their PWM to RPM characteristic, particularly when there is a build up of pressure in front of the fan versus in free air. \$\endgroup\$
    – Michael Karas
    May 5, 2013 at 2:20
  • \$\begingroup\$ It seems I assumed that this is easier to solve than it really is. I thought with a 555-timer, a few resistors, and maybe an octocoupler I could take the PWM signal as input, and create 0-733Hz open collector as output. \$\endgroup\$
    – Alex
    May 5, 2013 at 2:32
  • \$\begingroup\$ @Alex - A 555 chip can be configured as a VCO. The circuit of discrete parts would require "tuning" to get the PWM to frequency range correct. \$\endgroup\$
    – Michael Karas
    May 5, 2013 at 2:37
0
\$\begingroup\$

You did not specify the reason to remove the fan TACH signal from the equation with your power supply. I think that needs to be considered even though it is not a direct answer to your question.

First of all consider that the power supply designer saw an important consideration add a fan to the power supply. To help cool hot components!!

Secondly they went to extra concern to check for and validate proper operation of the fan through monitoring and range checking the TACH signal from the fan. Certainly there was great concern to ensure that the designed in cooling in the power supply continued to operate properly so that hot components did not get too hot!!

If your issue is replacement of a failed fan it would be much better to replace the fan with a proper equivalent unit that includes the TACH signal. Four wire fans that include a TACH are not that so hard to find from fan suppliers. The extra cost for such fan is likely smaller than the cost of a cobbled up TACH substitution circuit that would allow the use of a 3-wire fan. And with such substitution you loose the safety checking of the fan operation that the power supply designers felt was imperative.

If your issue is one of trying to remove the fan because you were trying to reduce the sound level near the power supply then please stop. This particular power supply was not designed to operate that way. Instead search out an alternate power supply all together that is designed to operate without a fan or has some other type of lower sound cooling solution.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Thanks Michael. As explained above, power supply is sufficiently cooled and I am not looking for a cheap way to work around a proper fan. I simply can't use a fan (and don't need it) in the way I am using the power supply. \$\endgroup\$
    – Alex
    May 5, 2013 at 2:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.