24V Gate Driver, cheap

It was playing off the half H-bridge circuit you linked up-thread, and assumes there's a load that pulls down Q4's source terminal so that Q4 acts as a source-follower.

If Q4 drives a capacitive load, you're right, -24Vgs is possible, so make appropriate provision. Otherwise, no, Q4 never sees more than 12V Vgs.

I posted it mostly for fun--the high-side gate driver ICs cost a little more but make up for that in parts-count and real estate.

A second reason for posting was to point out that 100% duty-cycle is possible if you drive the bootstrap with a nice squarewave.

John's is the simpler slower version, & might be fine depending on all those various considerations for your use which you know but we don't.

Cheers, James Arthur

Ooops, IR21xx, not IRF. But those are for hundreds of volts, and too pricey for you.

whit3rd suggests a P-FET, beat me to it. P-FETs are getting pretty good.

Cheers, James Arthur

That was my thought too. Klaus needs a steady bootstrap voltage.

Cheers

Does he? I missed that. Might as well go PMOS.

Meanwhile, back on the funny farm, a minimal 100%-duty, *slow* gate driver.

D2 .------|--||---+-->|--+----||-------+ : | C2(opt) | --- | | D4(opt)^ | +24V | | | -+- | === | | | | | | | ||--' | | || | | D3 | [R1] '----|

We are using the IR drivers, but for my application they are too expensive

Cheers

Klaus

und:

duty cycle:

rge the boost capacitor

I may be able to live with not being able to do 100% duty cycle, so the cap acitive level shift is very nice for that. Maybe it needs some protection d iodes, since burst voltage coupled into the 24V rail may couple via the MOS FET and capacitor and destroy the microcontroller driving the signals

Cheers

Klaus

Hundreds of volts is not a problem, you can even connect the upper driver to GND and obtain a very decent full-bridge oscillator using just 3 components (IR2153 etc.). But the price is, indeed.

Best regards, Piotr

Right, the problem with "hundreds of volts" is the price ($$) you pay for that capability.

Cheers, James Arthur

Anachronistic thread discussing 1980s hairball FET driver solutions when there wasn't a whole lot to work with. Hello, this is 2015, billions $ of monolithic solutions, use them

Hey, Fred, designing circuits is fun.

Here, let me help you with that concept: