Hello,
I have a 15pF capacitive load that I need to drive with a 15-cycle pulse train. I'd like to use a half-bridge where the high rail is 15V and the negative rail is -15V. (One end of my capacitive load is required to stay grounded). My system will run from a single Li-poly battery so I will have switching converters to generate the +-15V rail supplies. However, my microcontroller that will generate the control signals for driving the half-bridge will run from 0V to 3V. I was wondering if anyone might have any suggestions for the most efficient way to do the level translations for driving the half-bridge? That is, I'm looking for overall system efficiency. For instance, I'd like to get the +-15V switching converters to start up as quickly as possible and I'd like to turn them off as soon as I'm done driving the pulse train, and keep them off until I need to repeat this sequence. It would be good not to waste a lot of energy in the gate drivers. Any suggestions? Thank you.
Hello, I **may** be able to help (no promises!), but first I need to clarify the situation, hope these questions are not too impolite...:
1. For the worst case waveform, (ie:corresponding to maximum power usage), please describe the ON time and OFF time of this "pulse train". When the pulse train is ON, what is the pulse repetition rate? ie: what frequency or time between pulses within this "15-cycle pulse train"?
2. What is the **slowest** dv/dt that you can tolerate as the voltage across the 15pF load changes state?
3. Does the "hot" end of the load really have to go both positive AND negative wrt the "cold" end? ie: is it OK if the "hot" end is connected to a pulse train switching between 0V (gnd) and, say, +30Vdc ?
4. When the pulse train is running, does the voltage waveform across the 15pF load spend any time at 0V? Is it OK if it spends any time at 0V, ie: does it matter if the waveform is "square" or "quasi-square"? Can it spend any time at some voltage other than the +15Vdc, -15Vdc, and 0V ?
5. Continuing with the voltage waveform: can it be a sinewave? If yes, then this will allow the capacitor load be part of a resonant tank, which **could** result in lower power loss, cleaner waveforms, less EMI, etc. But it will require a resonant inductor, of course... ;-)
cheers..!