I am designing circuit with a high-side N-channel MOSFET switch. The switch will be operated at 100% duty cycle. I decided to use a high-side gate driver since they have internal circuitry to boost the gate drive voltage to the required levels. My first prototype used a IRS10752LPBF but I quickly discovered that the "bootstrap" configuration could not operate at 100% duty cycle. I am redesigning and looking for a gate driver appropriate for this application. What sort of gate driver must I select to operate at 100% duty cycle?
There are several kinds of high-side gate drivers.
A simple driver can achieve 100% duty cycle on a high-side PMOS. You get the disadvantages of a PMOS: higher RdsON, lower speed, higher FET cost. But it's simple, and driver cost is low, so the whole solution could end up a better compromise than the other options.
- NMOS Bootstrap
To drive a high side NMOS, a voltage higher than the power supply is necessary. If that isn't available, then one way is to use a bootstrap capacitor that is recharged then the low-side FET switches on and the high side FET switches off. This cannot achieve 100% duty cycle as the bootstrap cap must be recharged.
- Additional power supply
To generate the gate drive voltage above the power supply voltage, several options exist. A charge pump is a good option at low voltage, here is an example gate driver chip that does this. Unfortunately neither digikey nor mouser allow to search using this feature.
There is no bootstrap diode outside the chip or inside it, so if you want to quickly look through lots of datasheets this is what you need to look for, along with the words "charge pump".
I've also done it with a "hand-made" charge pump, basically a microcontroller GPIO output that feeds a diode voltage doubler. With this you can generate a voltage a bit above the power supply voltage, and that can be enough to drive the gate of a high side FET. If the GPIO stops switching, a pulldown resistor discharges the gate. Crummy and slow, but simple and cheap. With a 3V3 micro though, the two diode thresholds will probably lose too much voltage.
You could also use a higher voltage supply if that's available somewhere nearby.
If you need isolation from the driving circuitry, then another option is to use an isolated gate driver, and add a small isolated DC-DC power supply to generate the driver supply.
Another option is the photocell optocoupler, which is basically a LED and a tiny photocell in a chip. Lighting the LED will make the photocell generate a voltage, which can drive the gate of a FET. Output current is very low, so it is quite slow, but it is simple, does not generate any switching noise from a charge pump or an isolated power supply, and it is isolated. It's a good option if you use a pair of anti-series FETs to switch AC voltage.
I asked a question yesterday that may help you if you want to check out the thread I made last night.
Your solution may be either a boost converter to get your Vgs or voltage doubler and then controlling that signal to your gates via an optoisolator/transistor. I will likely go the opto route just because they're so simple to control.