Active Regulation of DC Resonant Chargers

Original Poster: "Marco Denicolai" <Marco.Denicolai-at-tellabs.fi> 
>>Original Poster: Greg Leyh <lod-at-pacbell-dot-net>
>>Avoid the big DC filter cap if at all possible.
>>If you really need 1% stability, active regulation
>>is far simpler, and more accurate.
>Without the filter cap the voltage will change from V to 0.866 V (ripple),
>that is a stability of
>+/- 6.7%. I would like at least 1% of stability, that's why the capacitor.
>What do you mean with active regulation? SCRs ? But you cannot remove
>ripple with SCRs without using a capacitor, can you?

Indeed you can!  A 'De-Qing' regulator works in the following way:

The resonant charger is set to overcharge the primary cap by about 20%,
so that the regulator has some headroom.

A compensated, fast HV divider is added, that monitors the voltage on 
the primary capacitor bank.

A shunt winding is added to the charging reactor core.

On each charging pulse a regulation circuit monitors the HV divider,
and when the desired primary voltage is reached, the regulator fires 
an SCR which is connected to the shunt winding of the reactor.  The
SCR diverts the remaining stored reactor energy into a resistor bank,
and the voltage on the primary cap stops promptly at the set point.

This circuit works quite well, and the pulse-to-pulse accuracy is
limited only by the accuracy of the HV divider and regulator circuit.
Compensated dividers with better than 1% accuracy are commercially
available from Ross Engineering.

In addition, the De-Qing regulator would provide stability against 
line and load fluctuations, unlike a simple passive regulator.  
The costs and hazards of the big filter cap, the 3 primary reactors,
and the charging diode stack would all be avoided, as well.