I have a circuit where microcontroller drives a H-bridge which is powered from 12V. I am having a problem with driving the high-side MOSFETs. I will be turning the H-bridge on/off only couple of times a day for up to 5 minutes at once.
I've designed a boost converter that is capable of converting 12V to 24V in hope of driving the high-side MOSFETs. However I am currently stuck with connecting microcontroller to boost converter for driving the gates.
This is how I am thinking of wiring it up:
simulate this circuit – Schematic created using CircuitLab
But…
Boost converter is working in continous current mode. And if I were to leave boost converter working on forever, if I wont drive H-brdige, the voltage output from the boost converter would rise untill something would fail.
And if I were to enable the gate driver IC only when I would want to turn the H-bdrige on, same thing would happen, because I would need to power it for up to 5 minutes.
Does anyone have any idea on how drive high-side MOSFETs without blowing something up?
Best Answer
I agree with Austin that a voltage doubler should be more than sufficient for your mosfet - the gate uses virtually no current at all.
From your diagram, it is unclear whether the gate driver and the level shifter are controlled from different MCU outputs or a common one.
Is the "Gate Driver IC" the core of your booster? From what is shown, I don't see why you'd need the "level shifter" - drive the driver from ONE pin, and when it is high, the H-bridge FET will switch.
Unless I'm misunderstanding something in your diagram, I'd think you could simply use an optocouper driven from an MCU pin to control current into a voltage doubler circuit to drive your H-bridge. Any number of OC's would work, since you're not driving much current. You're working with 12V, which your uC doesn't run off of, so the OC is necessary.
In fact, a Dickson Charge Pump clocked from a uC pin (i.e. program a pin for PWM) would provide an isolated voltage doubler: opto + 2 caps and 2 diodes. If you're not clocking, it's not doubling and your gate won't have enough voltage to conduct.
Basically, you'd be keeping the uC (MCU), H-bridge, and V1, and the Dickson would replace the rest of the components shown there, outputting to M2, running to ground, and sourcing from the 12V Vcc, with the clock as a solitary input from the uC. Opto is < US$0.50, caps and diodes are cents apiece.