Is there an optimal frequency for PWM?

[SolarTrap]

Hello all,

I am using PWM in my solar charge controller and by default the frequency is 900 Hz of the micro controller?
with 900Hz you and hear the MOSFETs making noise when not full on. I tried setting the frequency to 9kHz which makes the sound not audible any more.

Is there an optimal frequency?

Markus

 

[MikeHalloran]

Some people will still hear a 9kHz noise.
Most people will not hear 18kHz.

... but you need to look at your waveforms to see how fast you can go and still get clean switching; you might need more gate voltage, or different reactance in the load.

... and there's a limit to how fast a microcontroller can toggle a switch, given all the other stuff it may be doing simultaneously.


Mike Halloran
Pembroke Pines, FL, USA

 

[itsmoked]

Frequently there is/are specific components mounted in the wrong way that cause the board to become a sounding board. There are various techniques for damping that or eliminating it completely depending on exactly what part the noise is emanating from.

You should put in some effort to figure out exactly what is making the noise, likely it is not really the transistors.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[MacGyverS2000]

I have made quite a few PWM circuits over the years... the power versions always sing. That said, it's typically what I'm driving, not the driving circuit itself. For example, when making a heater controller, the heating elements themselves put off a nice tune... I could tell when the unit was approaching the desired temp as the PID algorithm would kick in and the frequency would ramp down. Music to my ears.

I'd suggest it's the inductors on the board that are singing, not the FETs, and Keith's suggestion is sound (pun intended).

Dan - Owner

http://www.Hi-TecDesigns.com

 

[itsmoked]

On SMD boards the most likely candidates are the capacitors. They actually get larger and smaller every cycle. Routing a gap in the board dividing it beneath the singer often stops the noise.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[benta]

In my experience, 99% of the acoustic noise coming from converters is caused by the inductors/transformers, due to magnetostriction.
There is not a lot that you can do about it except placing dampening material around the cores or filling them with epoxy or equivalent.

A better solution is to move the switching frequency above 50 kHz, then you won't annoy your dog or other pets in the vicinity.

Cheers,

Benta.

 

[SolarTrap]

Thanks for the suggestions but I am having a hard time locating the source. The sound seems to be radiating everywhere. I might have to get a tiny microphone and move that around the parts.

Is there a low end of the PWM frequency where it becomes more or less useless for regulation purpose?