COP 20.00 (2000%) Times, Reactive Power Energy Source Generator,

[popolibero]

Hey Barry,


that shouldn't be problem. Even if both switches are ON during the discharge phase the energy can only go from cap to transformer in that moment, as far as I can see.
The sine wave you see on my shot is on transformer primary. The current wave is across the light bulb.
Strange, your wave looks like if you had an inductor in series with your resistive load. Meaning instead of having the current rise immediately it rises slowly, like initially impeded by an inductor. But it still would have to decrease towards the end of the half cycle since cap is being emptied.


Where did you put your probes?


regards,
Mario

[popolibero]

Barry, also, I wonder if your mosfets are turning off completely? The switching sequence you had before your last update should actually work, from the transformer secondary point of view, even if the caps alternate, it still sees charge/discharge. That sequence is actually even better because in the last one you did you are not able to control when the cap stops to charge and starts to discharge (sine peak), current can go either way since both your switches are on.
Well the best option is to have 1234 all separate signals, but I don't know if you can achieve that with your circuit?


regards,
Mario

[listener191]

Hi Mario,

After delaying discharge pulses to around 135 deg, I obtained attached waveforms. 

Just a quick non precision check with an energy meter on the ouput of the transformer, (the line transformer in the amplifier), shows 31W, 85VA PF 0.35

I note the timing is very critical and the current waveform can easily collapse back into 0 phase shift with respect to voltage.

I did manage to fine tune with larger capacitors and got even closer to your waveform however, I had MOSFET failures (all 4 short) not been able to determing cause yet.

This happened when I increased the load, which is a rheostat. I cannot see any voltage transients to speak of, so suspect current transients are the problem,

I was using IRF840s which are good for 8A , 32A pulsed .

I can change these out for some 20A 1200V IGBT's which would be more robust.


Barry

[popolibero]

Hi Barry,


where are you putting your probes? If the red wave is across the load I'm having a hard time understanding how your current can rise that way. Is it possible that your rheostat is acting like an inductance?


regards,
Mario

[listener191]

Hi Mario,

The current probe (clamp on)  is between the load and the switches.

The load is 208 ohm and has an inductance of 249mH

I found the problem that was destroying the MOSFET's. I am using 60uF non polarized caps and I have two more sets of these but I decided to place 660uF electrolytics  in place to make a large increase in capacitance. The problem occured as the range of my timing adjustment allows the charge period to be moved into the opposite phase and the electrolytics were shorting when this happened. I have reverted to non polarized.

Barry