# High Frequency operation of Three level DC/DC Converters

1. ## High Frequency operation of Three level DC/DC Converters

Hello all,

I am designing an isolated high voltage power supply with the aims of reducing the overall system size and weight, increasing power density. This is leading me to find converter topologies which are capable of operating at high frequencies efficiently. I have been studying the series resonant, parallel resonant, LCC and LLC converters for this task as they can absorb transformer non-idealities into the resonant tank design, meaning higher frequency operation can be achieved without parasitic effects. I have recently came across, however, a technique called the three-level DC/DC converter (Neutral point clamped, flying capacitor, diode clamped etc). Every paper I read on this states that it allows the MOSFET/IGBT switches to be operated at much higher frequencies with much more efficiency than standard 2 level designs. However, as of now, none of them offer why this is the case.

My ideas are as follows:

1) The voltage stresses on the devices are reduced, therefore the stress at a switching instant is reduced by half. However, this seems pointless if wide band-gap devices or soft-switching techniques are used.
2) the transformer is subject to a quasi-resonant square wave, removing the effects of operating a transformer at high frequency driven with a square wave voltage (due to harmonic content, etc).

And that's about as far as I can tell the benefits. However, I wish to use WBG devices and use some kind of soft-switching techniques, and my input voltage is only maximum 300V, so I'm unsure if the reduced switch stress is a massive plus either.

If anyone has any input, or any recommendation for other possible high frequency isolated, high output voltage DC/DC converter topologies that I could investigate (or general tips) then I would be very grateful for that.

I have attached a standard DC/DC 3L converter image.

•

2. ## Re: High Frequency operation of Three level DC/DC Converters

For low voltage like 300V I don't see any advantage of multi-level topology. Half- or full-bridge LLC seems like an obvious solution. Do you mind to mention the intended power level?

3 members found this post helpful.

3. ## Re: High Frequency operation of Three level DC/DC Converters

Hello FvM,

The power level is approximately 650W. The issue with the LLC converter from my point of view is that my converter has two outputs both in the kV range. I've never really seen an LLC converter at this voltage level, and since it doesn't take into account the high parasitic capacitance of the transformer the performance at high frequency will be affected, unless it is nullified in some way. But I think still the LLC will have problems at this high output voltage, and since the magnetizing inductance of the transformer is typically high in HV applications, it doesn't really effect the resonance operation (to my understanding).

Thank you

•

4. ## Re: High Frequency operation of Three level DC/DC Converters

LLC and other resonant topologies are still the best way to deal with winding capacitance, but it's only reasonable up to certain level. A voltage multipiplying rectifier is much easier.

5. ## Re: High Frequency operation of Three level DC/DC Converters

for LLC with HV o/p's put the resonant comp's on the sec side ...

- - - Updated - - -

Is what you have posted even a 3 level converter? is there a connection missing ... ?

•

6. ## Re: High Frequency operation of Three level DC/DC Converters

Yes google says that's a 3 level. It puts out 0, +/- 0.5, +/- 1.

There are a million demo boards and reference designs of LLC and Phase Shifted Full bridge's running off 400V. 300V is no problem.

Such a multi-level inverter is comparatively exotic. Best to avoid it unless you really need too.

For high output voltage first turn too a full bridge rectifier instead of center tapped. Then consider multiple isolated secondaries in series.

7. ## Re: High Frequency operation of Three level DC/DC Converters

google may say it - but the drawing don't show it - that dwg shows a 2 level only ...

8. ## Re: High Frequency operation of Three level DC/DC Converters

Post #1 schematic needs for the join between VD5 & VD6, to be connected to the join of the capacitors at the left. Then it becomes 3-level diode clamped. Those diodes conduct inductive kickback, when the transistors are gated correctly.
Simulation of the concept:

The screenshot has the upper diode (led) conducting briefly. Both led's point upward.

- - - Updated - - -

The system is 5V to make it easy to control using logic gates.

--[[ ]]--