Gate driver

The first problem was how to make a gate driver circuit for the brick(s). The datasheet makes scary reading. What you need to look at for driver design is the graph of "Gate voltage vs. gate charge". You find the value of gate voltage you want to get to, and the chart tells you the amount of charge you need to dump into the gate to make it happen. Unfortunately the chart only went up to Vcc=800v and Vge=20V. We're heading for Vcc=1kV and Vge=30V! So it's time for a bit of guesstimation, that suggests the required charge is 6000 nano-coulombs (nC).

Now we know how fast we want to turn on, that's about 0.5us. We also know that Q=I*t. So we can say I=Q/t= 12 amps!!! OK All we need now is a 12 amp 30 volt gate driver. I designed this:


Brick driver circuit mk.1

Basically it's a simplified half bridge made with MOSFETs. M1 is driven straight off the 555 timer. It turns the brick off, and also drives M2. How does this work? When the 555 turns M1 off, its drain is pulled towards 12V by current through R3. As M1 drain voltage rises, first D1 turns off, disconnecting M1 drain from the brick. Next M2 begins to turn on, bootstrapping itself through C1 as it does so. Within 500ns M1 is fully on and the output is high.

When M1 is turned on by the timer, the reverse happens. The drain voltage begins to fall, first turning M2 off, then pulling the output down through D1. Note both MOSFETs can never be on at once in this circuit.


How the circuit was laid out


Gate drive waveform measured at terminals of brick

The ringing on the waveform is caused by the gate capacitance of the brick resonating with the inductance of the wires that go from the driver board to the brick. If you measure it at the driver board terminals, it's clean. Anyway, now we have a driver, time to test the brick out.

<<< Previous: A Modest Proposal Next: Testing the brick >>>
Return to TC Page