GZPA Class D 4000W with IRFP264

[mgm2000ro]

along with the signal restoration into rectangular waveform i will lose the dead time => more cross condution?

[Eva]

Originally Posted by B.I.G

i just changed all the small transistors to MJE340/350 and ktc4793 to 2SC4793 and KTA1837 to 2SA1837 and the oscilating freq has doubled ...

Be careful with that, self oscillating circuits rely heavily on propagation delays, so replacing parts by different ones without shifting the operating frequency too much is tricky.

Do waveforms improve with the new transistor set?

[Andrew Eckhardt]

Originally Posted by mgm2000ro

along with the signal restoration into rectangular waveform i will lose the dead time => more cross condution?

Accounting for faster gate rise time, yes, the total propagation delay is reduced and so is time between turn-off and turn-on. Buy increasing peak gate voltage you also increase the turn-off charge to be yanked out of the gate, which can also add to undesirable overlap. This might not be the entire reason switching frequency doubled, but it is probably a major factor. I'm not certain but I don't think the differential/level shift transistor changes you made would have that much of an effect. They'd just be easier to cool. You could always put the MPS transistors back and see.

[stewin]

hi B.I.G did your last schema work at +/-90vlts ? is this approach of discreet class d clear as the ucd one?

[stewin]

Originally Posted by stewin

hi B.I.G did your last schema work at +/-90vlts ? is this approach of discreet class d clear as the ucd one?

[stewin]

any feedback?