Print cooling fan not working properly

SnowCrash

@red-sand-robot said in Print cooling fan not working properly:

@snowcrash Not sure if I am understanding exactly how PWM affects the fans, but could setting a high frequency (as in the mHz range as you mentioned) damage the fans or the board at all?

I'm running a 24v blower fan and any time I set it lower than 100% it emits a high pitched whine (at 500hz). Curious as to if changing the frequency to much higher will remove the whine without damaging anything

Raising the PWM frequency to the levels I mentioned won't cause any damage to the fan or board (btw, I previously verified with @dc42 in another thread that the board can generate these frequencies and indeed it can). I'm not surprised your blower whines at 500Hz. Give 25MHz a go and see if it solves the problem (don't forget the value is entered in Herz, so 25000).

samlogan87

@SnowCrash so 25kHz is what you recommend for even a 2 wire fan that is pwm’d? The wife got sick of the whine so I ordered some 4 wire fans but in the mean time should I bump it up that high as it might be a few weeks before the others show up.

Regards,
Sam

SnowCrash

@samlogan87 said in Print cooling fan not working properly:

@SnowCrash so 25kHz is what you recommend for even a 2 wire fan that is pwm’d? The wife got sick of the whine so I ordered some 4 wire fans but in the mean time should I bump it up that high as it might be a few weeks before the others show up.

Regards,
Sam

Hi Sam,

No reason not to try and see. Though certainly not all 2-wire fans like to be PWM'd as they weren't designed for this kind of operation, still in many instances there will be a significant improvement with higher frequency in terms of noise (and functioning).

Btw, if you're hooking up a PWM fan to the Duet by running the blue (=pwm signal) wire to the GND pin of the fan header and independent power lines to the fan, don't forget to add the protection diode on the signal line.

For anyone interested, I dug up 2 previous discussions on this topic in which @dc42 took part (and also include instructions on the said protection diode):

https://forum.duet3d.com/topic/6054/help-configuring-pwm-fan-in-thermostatic-mode-for-duet-drivers/4

https://forum.duet3d.com/topic/5248/maximum-frequency-of-pwm-fans

Red Sand Robot

@snowcrash Hey, thanks for the tip to turn the frequency up to 25kHz! completely fixed the electronic whine I was hearing from my blower fan, didn't even think it turned on when I set it to 30% speed!

SnowCrash

Very cool, @red-sand-robot, I'm happy this helped

samlogan87

@SnowCrash Thanks for that. I am running them at 25000hz and I think there is a slight improvement but still noticeable. I am guessing the fans that I got are not the best.

Kind Regards,
Sam

dc42

@snowcrash said in Print cooling fan not working properly:

Hi @jckray,

Glad you got it sorted out.

Just thought I'd chip in about the issue of PWM frequencies for the benefit of anyone else struggling with pulsating fans and potentially reading this thread. 100Hz (or even 1000Hz for that matter) is way too low for PWM fan signals. Although there's no one frequency that's suitable for all PWM fans, conventional wisdom puts the number at 25,000Hz (or 25MHz), while some even opting for a little higher (around 30MHz). Often times, changing the frequency to that level makes a significant difference in terms of the fan's acoustics.

The 25kHz figure is quoted for 4-wire fans with a separate PWM input. You shouldn't use high PWM frequencies with ordinary 2-wire fans unless you use a series inductor followed by a parallel capacitor to smooth the PWM into DC.

Most 2-wire fans were never designed to be used with PWM, and it's amazing that they work as well as they do.

Red Sand Robot

@dc42 Is there any danger to the electronics from using a high PWM frequency (in the 25kHz range)? Or would the other danger be to the fan itself?

SnowCrash

The 25kHz figure is quoted for 4-wire fans with a separate PWM input. You shouldn't use high PWM frequencies with ordinary 2-wire fans unless you use a series inductor followed by a parallel capacitor to smooth the PWM into DC.

Most 2-wire fans were never designed to be used with PWM, and it's amazing that they work as well as they do.

@dc42, to pick up on @Red-Sand-Robot most recent question, while I completely agree with your second statement, I'm a bit baffled by the first.

What's the specific electrical/mechanical problem with using high PWM on 2 wire fans?

dc42

High frequency PWM is ineffective if the fan has a capacitor across its power terminals. Switching losses in the MOSFET increase with PWM frequency, so if you do use 25kHz PWM to switch the power input to a fan, watch out for the MOSFET overheating.

SnowCrash

@dc42 said in Print cooling fan not working properly:

High frequency PWM is ineffective if the fan has a capacitor across its power terminals. Switching losses in the MOSFET increase with PWM frequency, so if you do use 25kHz PWM to switch the power input to a fan, watch out for the MOSFET overheating.

Thanks for the clarification. @dc42.

For me, this is one of those cases where theory & practice seem to diverge somewhat - at least in my experience. I've been working with every possible kind of fan for years now and run them off every conceivable (dc) supply type and although it's certainly true that some 2-wire fans really don't like to be PWMed, I never had a fan fail in the manner you describe. On the other hand, I've had quite a few 2-wire fans with which the noise level was significantly reduced when run in the 25-30KHz frequencies and they worked fine throughout their lifespan.

Perhaps I've just been extraordinary lucky in this context or possibly the danger here is very marginal. Either way, everyone's millage is different, as they say, and I guess each of us can make his/her own choices as to the level of risk they'd like to take and how to go about it.