Hello All

I'm working on a high voltage power supply of my design. It applies high voltage AC to an ozone generator tubes
and the power supply works just fine. The PSU works offline (from 22v mains) and is built around a IR2151 chip, The trnsformer is an U core ferrite
It has the primary and low voltage secondary in one leg, and the high voltage secondary secondary in the other leg

As I said, the power supply works fine, the transistors run cold without heatsink, and HV is generated ok to the tubes. The problem is that
the tubes run very hot and air cooling is not enough, so I decided to lower the AC voltage by lowering the HV secondary turns, until
the tubes begin to run in a lower sustainable temperature. Even running partially on (with corona patches).

But, here's my problem. When I change the secondary turn to a lower turns ratio (I have many windings for testing, it's very easy to change in a U core), High voltage is generated, but the drive transistors begin to get very hot, at the point that a heatsink it's not enough
I don't know what's going on there. Maybe I'm messing somenthing with the impedance configuration. Maybe, by lowering the secondary turns I made the HV secondary to resonate with the ozone tubes (they are a 'pure' capacitive load of 355pF)

The operating frequency is 30KHz by now. I do not know what are the recommended methods to lower the output voltage, using a primary series capacitor, changing the duty cycle to <50%
using a primary series inductor, or changing the turns AND the operating frequency.

If someone has some experience to share please, or some suggestion about what can be the cause of this problem, please tell me

I'll be very grateful

best regards