Confused - electrolytic vented? (and LC filters)

[Things]

OK, will give the other diode a try

I'm going to order some beefier inductors from e14, is that 47uH the best choice for 32Khz? I decided on that vs 62Khz as my gate drive is only 5V.

[DanielS]

I'm going to order some beefier inductors from e14, is that 47uH the best choice for 32Khz? I decided on that vs 62Khz as my gate drive is only 5V.

There is no "best choice" since that depends on a hundred other parameters like how much you can afford to pay, how much space you can afford for components, transient response requirements, regulation, availability, desired efficiency, current, etc.

You cannot pick any random inductor based on inductance value alone either since you may run into unexpected issues with core losses, core saturation and a handful of other parameters.

[Things]

This is not something that'll be going into production, so in terms of availability and price I don't really care. Just a hobby project.

Basically all I want, from 12V 32Khz PWM, is as close to DC as I can get (though realistically even 500mV ripple is probably acceptable), at a max current of around 6A - but the output voltage will be variable. I have found numerous LC filter calculators online but they all seem to calculate for a specific voltage.

[kingofkya]

Reading the first post, your not trying to control a pc/server fan right? Because you can just feed a square wave into the 4th wire on most of them.

section 2.4 has some info.
http://www.formfactors.org/developer%5Cspecs%5Crev1_2_public.pdf

[Things]

Reading the first post, your not trying to control a pc/server fan right? Because you can just feed a square wave into the 4th wire on most of them.

section 2.4 has some info.
http://www.formfactors.org/developer%5Cspecs%5Crev1_2_public.pdf

Unfortunately no, I'm essentially dealing with a car radiator fan and a bunch of DC brushless centrifugal blowers (with no PWM input).

[wraper]

OK, will give the other diode a try

I'm going to order some beefier inductors from e14, is that 47uH the best choice for 32Khz? I decided on that vs 62Khz as my gate drive is only 5V.

for 32 kHz better use 100 or 68 uH. As of gate voltage, frequency is not dependent on it at all. However pwm source must have enough current output capacity to provide fast rise/fall time.

[Things]

Have decided to move all the motor drive stuff to it's own board and enclose it an an aluminium box .. hopefully that'll keep the microcontroller from resetting all the time due to EMI.

I also went out and bought a couple of toroids, the specs say they're good for up to 50Khz so should be OK. Bought the heaviest magnet wire Jaycar had (1.25mm) which should handle the current fine, and wound about 15 turns on. Not exactly sure what inductance this gives me but I'll soon find out of it works or not I guess

[wraper]

That looks like a iron powder core. IMO you would need like a 2x more turns for that. off course you should measure actual inductance rather than guessing. If you don't have anything to measure inductance and no signal generator, you can use your MCU + 2 mosfets to make a square wave.

[Things]

I've just had a chance to test it this morning, and it seems to work well Nothing gets even warm and the fans are dead silent in terms of PWM noise

Also just moving all the power stuff off to it's own board seems to have stopped the resetting issue.

[Things]

I've completely finished rebuilding the controller now, and everything works perfect The inductors and diode get a little warm after a few minutes, but I just put a small 40mm fan in the box and that fixed it right up.

No more squealing fans or power supplies, and no microcontroller EMI resets, woohoo! Thanks everyone for the help