Generating a square wave of 40Vdc 125A 20kHz output?

Hello Steven,
Thank you for posting.
I will need to do some research on this.
I will reply by end of day Friday.

Thank you for your patience.

Steven, what is the input voltage range of your design?
Power supply specifications need a Vin as well as Vout Iout.

Do you have slew rate specifications on the square wave edges?

What kind of load will you be driving?
A reactive load (capacitive or inductive) adds levels of complexity.

Any other details of what you need will help the search.
Thank you.

Hi Ed

We have a Sorensen power supply. It output 40Vdc and 125A. Unfortunately ~20kHz square wave is the critical component for this test we are doing.

This power supply cannot achieve that, but we can use it as Vin? if not 3 Phase Line-Line: 200VAC-240VAC is preferred.

No specification on the slew rate at this stage. I assume the faster the better but not critical. 

load is  a resistive load. 

The test need to run multiple loads. Saving some cost we are hoping to use one power supply for two set of loads by switching the square wave out of phase. Not sure if this is achievable.

Thank you

Steven

Hello Steven,

I was thinking along the same lines as your drawing, at first.
Your mosfets are shown in a bridge configuration which would allow the output to be +-40V switched.

But do you need -40V?
Creating a +-40V would require driving 4 mosfets as you show.

A simple switch between the power source and the load would modulate the power to the load.
This would produce a 0 to +40V square wave output.

This is a picture from a TI app note; 
Cut-Off Switch in High-Current Motor-Drive Applications
http://www.ti.com/lit/an/slva991a/slva991a.pdf

Your power source may have enough capacitance at its output to keep the turn on spikes low enough.
But you could add some more capacitance if needed, before the switch, so the caps source the current spikes into the load.

TI has power N Fets up to 100V VDS
http://www.ti.com/power-management/mosfets/n-channel-transistors/products.html#p267=59;100

This mosfet is one to consider.
www.ti.com/.../csd19532ktt
100 V, 4.6 mΩ, D 2PAK (TO-263) NexFET™

I would suggest 3 or 4 in parallel to keep losses, and heat, low.


The NPN PNP gate arrangement would be lowest cost.
You will need to experiment with how much current you drive the gate with in order to get the waveform dynamics you want.

The wires connecting the load and switch, both positive and ground, have inductance which will induce voltage spikes as the switch.
V=Ldi/dt.
You should consider some clamps and/or snubbers in the design to control these spikes.
Otherwise they could damage the mosfets by exceeding their VDS rating.

I welcome your thoughts.