Guru
Joined: 09/08/2007
Location: AustraliaPosts: 4406
| Posted: 10:40pm 15 Apr 2018 |
Its really an unavoidable problem, and wobbles or freedom from wobbles seems to be more a combination of lucky coincidences than anything else.
Stray resonances are always going to be there somewhere, they just may not see enough energy at the specific frequency to be provoked into action.
Unipolar drive may have something to do with this as well.
One side of the transformer is being driven with a 50Hz square wave, and that will be pretty rich in odd harmonics.
So if there is an unlucky resonance at a few hundred Hz that coincide with one of those harmonics, that might be enough to set it off. That is all just a theory....
With bipolar drive, both sides of the bridge see high frequency PWM, and the only low frequency energy will be at 50Hz. And because its a 50Hz sine wave, there should be no significant harmonics of 50Hz present at the transformer. In fact it should be pretty clean right up to the PWM switching frequency.
I have never tried comparing unipolar PWM drive to bipolar PWM drive as far as the effects on any kinks or wobbles. It might be something to try one day. Some of the driver boards have a link to select either mode of operation.h
My own multistep inverter has a very good sine wave shape, but with a very small amplitude ripple all over it everywhere. Total harmonic distortion unfiltered is less than 1.5% so its not too bad.
I have tried many times to filter that out, but every filter I have tried so far brings up wobbles and self resonances that actually make things worse than without any filtering at all.
With high frequency PWM there absolutely must be some output filtering, and that is the problem.
Cheers, Tony.