On 20 Apr 2006 at 2:15, ccool_j wrote:

> Also, my inputs are 2 50% square waves generated by a D-latch. These
> 2 signals then go through 2 sets of push-pull fet combination (n and p
> fets). And like JL Naudin's design, I used nFet to "drain" the
> current. My fets are IRF520

Have you tried lower pulse width with a deadtime in betwwen?
Say something like 2 x 40% with 10% OFF.

Mark Jordan