24V Gate Driver, cheap

How would you connect that as a highside driver?

I tend to use a small SSR when I can, rather than messing with mosfets and highside drivers and all that. But they are fairly slow.

Why not use two jellybean complimentary emitter-follower BJTs driven by a CMOS gate?

I just did a test instrument board that used some P210x optoisolators to drive the high-side MOSFETs, but they are not cheap.

--sp

The logic signal from the controller is level shifted to the 24V rail plus bootstrap vcc with a current source and high side resistor.

The CD4050 boost that signal, riding on the switch node to drive the powet N FET gate.

I could replace the 4050 with complementary BJTs as emitter followers, but then the voltage drop of the npn makes it difficult to drive the gate properly off

Cheers

Klaus

Like this one:

There are lots of hi/lo driver chips around, that do all that stuff except the bootstrap cap. That would reduce your parts count, and maybe risk.

Do you need speed?

I'd consider a PMOS high side element, myself... But, why is the BJT an 'npn'? Oughtn't you use a PNP, pulling down the collector to the upper NMOS source with a suitable resistor?

I like the Micrel drivers MIC50xx, but their $1 or more. How about a MAX232 charge pump driver?

Cheers

You can connect a resistor from the CMOS buffer output (and bases) to the two emitters which will draw the gate down to almost exactly the source voltage over a bit of time, and uses no steady-state current.

There are lots of gate driver chips around.. Microchip has some that are advertised at 6A or more drive current, though the actual specs don't look quite that rosy.

If you are switching continously (or can arrange to do so) there are also the bootstrap driver chips- they provide the high side supply and level shifting, for prices starting at under 20 cents (1K). Eg. NCP81161 Parts are available to handle >>1kV for motor controls, and some switchmode chips might be suitable for your application.

--sp

The IRF2xxx drivers were known for being especially tough where other guys' popped. Cheap too, IIRC.

Cheers, James Arthur

I really do not need very high speed, but I need very low cost

Cheers

Klaus

THe MAX232 only goes to +/- 7V as far as I can see. I am running at 24V rail, so I need something to go that high

That is a really nice part, thanks (and cheap too). It might do the trick

Cheers

Klaus

Thanks

But wont the gate be damaged when driving low, then there's -24V on the power MOSFET?

Cheers

Klaus

OK, this is sort of minimalist:

That's nice, 15 cents in quantity.

The other possibility is to eliminate the mosfets entirely and use a gate driver chip to do the whole job. UCC27517A or something like that. Or a dual.

$0.1533@2500 and the diode is build in

but you have live with something closer to 5V for gate drive, because it is limited to 35V Vboost

-Lasse

This might actually do the trick:

It's more expensive, but may remove the need for the half bridge MOSFETs

Cheers

Klaus

The old capacitive level shift trick is nice, but cannot offer 100% duty cycle: