TheBackShed.com

Forum Index : Windmills : RF interference at 20Khz PWM?

Author

Message

Haxby

Guru

Joined: 07/07/2008
Location: Australia
Posts: 423

Posted: 04:08am 16 Apr 2009

Copy link to clipboard

Print this post

My idea is to have a microprocessor controlled charger for a wind turbine.

The processor will have two inputs: The turbine RPM and the wind speed taken from an anenometer.

Based on mill RPM and wind speed, it will have a lookup table and will appropriately PWM the power going into the battery bank through Mosfets.

I expect that I will be switching maybe 30 amps and I wonder what sort of RF interference this may cause. The switching frequency will be just above audible 20Khz or so.

Has anyone done this before, and what were your results? My guess is that the frequency is too low to be picked up by radios and TVs, but I know little about RFI theory.

Robb
Senior Member

Joined: 01/08/2007
Location: Australia
Posts: 221

Posted: 04:44am 16 Apr 2009

Copy link to clipboard

Print this post

Make sure its in a shielded case and all leads are well filtered.

Its not the 20Khz content you have to worry about but rather the HF content the square edges of your PWM will produce.

Hope you don't have a cat, dog or kids

Haxby

Guru

Joined: 07/07/2008
Location: Australia
Posts: 423

Posted: 06:47am 16 Apr 2009

Copy link to clipboard

Print this post

I'll give that a go.

I think its impossible to shield everything because the cable going to the batteries will have the high current PWM going through it.

I could put a large capacitor on the mill side, and an inductor on the battery side and I could check with my scope to see the edges of the PWM, but I don't know what I am aiming for in terms of smoothing out the square wave.

Is there a general rule of thumb as to what slew rate the square wave should have that will minimise RFI?

GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127

Posted: 08:04am 16 Apr 2009

Copy link to clipboard

Print this post

Hi Haxby,

RFI caused by the sharp switching edges of the mosfets is a major problem with Heath Robinson designs. Pay special attention to LC filters on the DC lines to the input and output of the converter. The actual switching square wave is not smoothed at the mosfet.

There should be no pulse current outside the converter. I would not just us a DC chopper type cct with a series inductor on the output, without some steering diode as well.

I have not found it possible to measure accurately the windspeed that the windmill is directly subjected to. beware of timing delays used in frequency measurements and the effects caused in the control loop.

Gordon.

become more energy aware

Print this page

To reply to this topic, you need to log in.