Bidirectional High-Voltage Converter

[Leitz83]

Hello!

I am in search of a bidirectional high-voltage dc/dc converter for capacitive loads:

- Output Voltage: min. 2kV
- Output Current: 0...100mA.

Maybe someone can give me some hints on how to realize such a converter.

Regards Leitz

 

[Orson Cart]

Bi-directional - what is the voltage on the LV side? presumably you want bi-directionality to discharge the cap? Regards, Orson Cart.

 

[Leitz83]

The voltage on the low voltage side will be 325V. Yes, I want to charge and discharge the cap.
I thought about a bidirectional flyback-converter (one switch primary, one switch secondary), but the limited drain-source voltage of MOSFETs is causing me headaches...

Regards Leitz

 

[mtwieg]

If you want bidirectional operation, then you'll need synchronous rectification (two switches), and I don't think there's any avoiding the need for high voltage devices. Look at high voltage IGBTs instead of mosfets.

 

[Leitz83]

Thanks for the hint. I thought there is a possibility to connect MOSFETs in series, but that probably entails other problems.

I thought about a bidirectional fly-back converter.

 

[Orson Cart]

A bi-directional flyback could be made to work, but you would need 3 x 1kV mosfet or IGBT in series on the HV side (due to the flyback effect when you are discharging the cap), also when you are discharging most of the energy will arrive back at the input, so you will need large-ish caps on the LV side if the voltage there is not to increase too much during discharge of the HV side. It would be quicker and simpler to build two converters, one for charge and one for discharge, or more simply get some 2.5kV igbts to do the discharge in a controlled linear fashion, regards, Orson Cart.

 

[Leitz83]

Thanks for your suggestions.

Do you mean that building 2 converters is easier, because the dimensioning of the components for each converter is simpler?

regards Leitz

 

[Orson Cart]

Two converters is slightly easier, yes, better still is to have a controlled dissipative ckt to discharge the caps, using a 2.5kV IGBT say - much easier to implement and control.

 

[RCinFLA]

Don't think about trying to put together series MOSFET's to achieve higher voltage breakdown. You will never be able to match and control the turn-off, turn-on times, so poof will be the result.

I agree the best chance is parallel reverse IGBT's push-pull converters. Parallel flyback design will likely work also but push-pull will give you smaller transformer size for 200 watts.

 

[Orson Cart]

Don't think about trying to put together series MOSFET's to achieve higher voltage breakdown. You will never be able to match and control the turn-off, turn-on times, so poof will be the result.

Usually the above is true, Power Integrations have a pretty neat app note which shows how to series a mosfet with the mosfet in their chip to raise the max volts from 700V to 1400V - can be used with any pair of similar mosfets, Regards, Orson Cart.

**broken link removed**

 

[Leitz83]

Okay, i will think about the suggestions you made. Thanks.

@RCinFLA:

What do you mean with "parallel reverse IGBT's push-pull converters"?

regards Leitz

 

[Orson Cart]

Suggest IXGK75N250 from IXYS as suitable for a 2kV discharge device - used in linear controlled mode - Regards, Orson Cart.

**broken link removed**

 

[RCinFLA]

Okay, i will think about the suggestions you made. Thanks.

@RCinFLA:

What do you mean with "parallel reverse IGBT's push-pull converters"?

regards Leitz

One full converter forward, one full converter in reverse. Avoids having to deal with syncronous rectifiers on bi-directional converter.

 

[Leitz83]

@RCinFLA:

Thanks for your explanation. But I still would have to handle the high output voltage when discharging, wouldn't I?

regards Leitz

---------- Post added at 12:10 ---------- Previous post was at 11:16 ----------

I mean, we will definitely need more than 2kV as output voltage in the future. But I will clarify this with my superior on monday.

regards Leitz

 

[Leitz83]

The galvanic isolation is a must. Furthermore it should be possible to increase the output voltage up to 10kV in the future and a bidirectional power flow has to possible.

Won't be easy...

regards Leitz

 

[Orson Cart]

Going from 325VDC to 10kV is relatively easy (we make a module that goes from 20V to 6kV) but going from 10kV back to 325 is a bit of a waste of engineering time (although it can be done, we have a 14kW, 4kv to 350VDC down-converter module, which could be easily extended to 10kV - using bi-mosfets, especially at the lower power levels you require), however series mosfets or igbts in a controlled linear dissipator still looks to be the easiest and quickest solution to discharge the caps.
Regards, Orson Cart.

 

[Leitz83]

I share your opinion concerning the controlled linear dissipator. It would be much easier that way.

regards Leitz