170 watts in - 1600 watts out - looped - Very impressive build and video

[cheors]

Detail how L298 module could be used :

https://i62.servimg.com/u/f62/18/62/17/81/dzgene13.jpg

[listener191]

The coil switching conditions are similar to that in a switched reluctance motor.
In the situation where coil inductance is reduced i.e. not in registration with the rotor, usually current has to be limited particularly at low RPM.
As this generator runs at a fixed Phase RPM current limiting is not necessary, as currently operated.

No doubt there is energy wasted due to large angle over which the stator coils are not in registration with the fixed rotor, that can only really be mitigated by more rotor poles.

I think for an initial experimental build, the recovery that an H bridge provides is OK.

I believe Pierre is running the coils at a conservative voltage/current and I would imagine the current waveform is a typical sawtooth, with some variation for the coils in registration with the fixed rotor.

A faster rise time current waveform may be obtained by increasing the applied voltage. Current limit still to be observed though for the switch devices.

One interesting way for this to be achieved is to feed the cap bank to the switching devices rail and a smaller value cap via a suitable fast power diode.

The smaller value cap isolated by the diode, will develop a voltage larger than the supply rail due to the large recovery voltage spikes and the fast charge rate of the smaller cap.
This will help reduce the coil current  rise time. Now when you do this, you have to ensure your switching devices voltage rating will be adequate.

I have attached a scope shot (taken on an old digital scope), illustrating the effect of the higher voltage on a particular type of switched reluctance motor. Note the very fast current rise time in red and the resulting flux waveform in yellow (in this case saturating).

L192

[listener191]

Apologies I have done it again and I still cant find a way to delete the attachment.
L192

OK fixed it!

[seaad]

No doubt there is energy wasted due to large angle over which the stator coils are not in registration with the fixed rotor, that can only really be mitigated by more rotor poles.
L192

When primary field not in registration with the fixed rotor. Why keep on making a rotating field there in that stator sector (no-man's-land
) at all ?? With one fixed rotor. (Six poles !!)

[listener191]

Attached is an old circuit that I used on special type of switched reluctance motor.
It illustrates my previous post.

Note: I would only use MOSFET's these days not IGBT's as shown.

MOSFET's are far more robust for experimentation.

L192