[SOLVED] Continuous charging and uncharging a large capacitor with 20 mF,10 Hz and 400Vpp

Hello!

I need to charge and uncharge continuously a capacitor with 20 mF (application industry aggregate). The amplitude is 400Vpp (400V <-> 0 or +200V <->-200V). The signal should be a rectangular. Let’s say with a max. rise and fall time of about 10 ms. The frequency should by 10 Hz -> period of 100 ms.

For me it is not clear how to solve this problem. How would you realize this problem?

With a half-bridge?
Are there transistors, which are robust enough to carry the high charge currents of 1000A for milliseconds?
Are there commercial signal supplies which can do that?
Are there commercial inductors to limit the current (10-100 mH) with very low parasitic resistances?

How would you realize this problem?

Thank you for help and ideas!!!

to charge 20mF to 400v in 0.1sec takes 80A if charging current (not power) is constant.


...But maybe you start by charging with higher current, and then the current can tail off as the voltage increases.

Hi,

400V in 10ms means 800A.

The energy in the capacitor is 1600Ws.

10 times per second means 16kW.
And in addition to that the loss in the transistor.

Not impossible with a linear current source/sink.
You need a lot of transitors in parallel
You need water cooling.

What's the use of your application?

I'd use a switching power supply, maybe flyback topology.
But this is no job for a hobbyist.

Klaus

The requested charging time is 10 ms, so current is 800 rather than 80 A. It's surely doable using IGBT converters. Inductive components for this current level can't be ordered off the shelf but need to be custom made.

The instantaneous power required to charge the capacitor is really large (several 100 kW), you'll preferably have a larger storage capacitor or super capacitor to provide it and get only the power losses out of a power supply.

All in all, this is big power electronics project, needing considerable resources and knowledge.

Yes - it would be better to transfer the power from one set of caps to another and back again - with a power supply providing the losses ( and initial charge )

800A to get from -200V to +200 V on 20mF ( assuming OP meant milli farad - not uF ) the peak energy is 400J at +200V, to get there from zero in 0.05 sec is 8kW average - so the converters must handle 8kW average and ( theoretically ) 160kW peak when putting 800A into the cap near the 200v finishing point.

The best way to do this would be a resonant circuit with a quarter period of 10mS, switched by SCR / diodes with a power top up to the storage caps from a psu, let me see, C = 20milli-farad, T = 0.04 sec ( 25Hz) thus L = 2.026mH, Zo = 0.318 , peak applied V for transition ( -200v to +200V) = 400V therefore peak current = 1256.6A = peak current in choke - doable.

An half bridge of SCR/diodes running off a suitable +/- 200V supply with output caps 10X greater, say 200mF with a series choke of 2.026mH ( 1300Apk rated ) would do the job - there - done )

Thank you very much for the ideas!!! It is very helpful.
Transfer the power from one set of caps to another and back again sounds perfect.
But I have not jet understood how do you arrange the resonance circuits.
Do you mean something like the attached circuit?

Hi,

With a resonance system I expect about sinusoidal voltages....in both polarities.

But from the descripion I assume you want to fully charge a capacitor within 10ms, then wait 40ms, then discharge within 10ms, then again wait for 40ms....repeating.

Thus I'd use a controlled polyphase switch mode system.

Klaus

O.k. I think now I understand. The right timing is everything!!!. Thank you! I think this can be the solution!

if i understood your first post, you want to charge a 20 milliF capacitor to 400 V in 10 ms,
and discharge it in 10 ms and do this continually to get 10 Hz.
you've accounted for 20 ms out of 100 ms per cycle.
what is supposed to happen the other 80 ms?

once you have charged the capacitor, what did you intend to do with the stored energy?

Easy Peasy suggested moving the energy back and forth between 2 capacitor banks (if I understood what E P wrote)
that doesn't let you do anything useful with the energy.
you'll have a LC circuit at resonance, shuttling the energy back an forth between the L and the C except you won't have an L and you'll have to force the transfer both ways??

please accept my apologies if i misunderstood anyone

Dear Ritsch77, you are a fast learner, well done.

MCC500-12Io1 try aliexpress search