FeedBack Path loop compensation in DC to AC inverters

hi,

What type of feedback amplifier (type-2,type-3 or PI) is used in a pure sine wave inverter?

Suppose an inverter consists of a h-bridge and a low frequency iron core toroidal transformer. H-bridge is derived by sinusoidally modulated PWM (20KHz). At the secondary of the transformer there is a filter to get filtered 220v sine wave. The output voltage are scaled by a small step-down transformer and are feed to ADC of the micro-controller. The micro-controller adjusts the pwm duty cycle according to the output voltage.

What is the best technique of sensing output voltage and current in this type of inverters?

What you need to do is adjust the output voltage up or down in discrete steps.

Set yourself some realistic limits.
If the voltage exceeds the high limit, step down a notch to something just above the low limit. If it's low, step the voltage up.
It should work rather like auto ranging in a multimeter.

If you try to run full dynamic negative feedback using a PI loop, you will never have both good regulation and stability because of the wide range of horrible highly reactive loads the inverter will sometimes see. That does awful things to the feedback phase.

All you should really be interested in is maybe the peak, average, or rms value (does not really matter which) and try to keep the output within some realistic band. Ignore phase, just measure the output voltage in a basic way, and correct that.
Its simple and can work fast without any problems of instability.

That should be quite sufficient, because the grid supplied mains voltage jumps around and wanders all over the place anyway, and everything works just fine. Precise output voltage is not as critical as some people think.

Thanks for reply.

What is the main cause of H-bridge mosfet failure in such type of inverters?

The peak voltages are going to be pretty much restrained by the supply rails and inverse diodes.

The current can sometimes spike to destructive levels from a variety of causes, and that seems to be the biggest mosfet killer.