H-bridges - what am I missing?

IIRC there are chips that do the logic such that you have a direction input and drive input. Most microcontrollers designed specifically for motor drive essentially put out a "gate on" command, which you are then expected to hook up to a gate driver and external FET.

I have yet to find an all-in-one H-bridge chip that has drivers, logic that matches today's microcontrollers, and the capability to run in the tens of kHz. There may be one out there, but to date I've ended up either doing slow PWM or dropping FETs and drivers on my boards.

(Note that if you've got a 5V supply to the motor that a 74ACTxx part makes a fairly good driver for an itty bitty FET. Otherwise you're looking at more traditional driver chips that require 12V or more supply, but will drive FETs capable of steering 100W or more around).

Sorry if I'm being dense Lewin, but this part is for driving a DC motor in one or the other direction, for example car electric window.

Where does PWM come into it ? What kind of motor are you looking to control ?

Thanks in advance for the clarification, Best Regards, Dave

PS: careful on lead time of H-bridges - I'm having difficulty finding what we need that's actually obtainable...

That may be caused by looking at not entirely the optimal IC for what you have in mind. I'll second Dave's assessment here: that IC appears to be intended for simple DC use, not for PWM'ed operation.

The IC you're looking for would effectively need five output states (two different ones that end up as high-z) and thus _three_ inputs: e.g. direction, brake (for "00" state) and enable (turn off to reach "zz").

You could PWM both inputs, and thus have

PP:P0 --- PWM forward P0:0P --- PWM backward

I need to drive a DC gearhead motor forward and reverse, with some speed control (2 or 3 steps in each direction). Fine control of motor speed is not necessary, and this is open loop. It would be a low frequency PWM signal, but the switching times mentioned in that specific datasheet seem to imply up to maybe 30kHz is theoretically doable. Control is from a general-purpose micro with two 10-bit PWM outputs and various available GPIOs.

I could do what I want to do with discrete FETs but I don't want to have to design all the shoot-through protection, etc. etc. and worry about what I might have forgotten - much rather get a canned part.

I thought it was just me - almost all the parts I want are no stock everywhere...

If employers are afraid to hire even as the work picks up, I see no reason that manufacturers wouldn't be afraid to stock parts, even as demand picks up.

Perhaps the "jobless recovery" is a "partless recovery" as well.

We have a serious problem with this at my day job - most of the big guys are giving us random leadtimes. I think the recovery is uncertain enough that it isn't a sure thing that it's going to be worthwhile to fire up a new shift or bring a spare fab online or whatever it is - the mfrs just don't want to sign up for increased fixed costs.

On a related note, I found a device that has exactly the inputs I was expecting - National LMD18200. Seems like a very old part though.

I don't know what you call old----but I have that part in systems designed in 2003 and it wasn't new when I first came across it. OTOH, it is still available through DigiKey---although it is currently out of stock and there won't be more until ~10/15/10.

If you need one for testing and can't find an easy source, I have a leftover in the miscellaneous parts bin and can mail it to you.

Mark Borgerson

The even older L298N might also be viable. Stil available from multiple sources. In stock at Digikey.

Regards Anton Erasmus

Lovely part. Huge voltage drop compared to a MOSFET H-bridge.

Cheap, too. But dropping 10W(ish) in the chip for a 2A load?! :)

I need to drive a DC gearhead motor forward and reverse, with some speed control (2 or 3 steps in each direction). Fine control of motor speed is not necessary, and this is open loop. It would be a low frequency PWM signal, but the switching times mentioned in that specific datasheet seem to imply up to maybe 30kHz is theoretically doable. Control is from a general-purpose micro with two 10-bit PWM outputs and various available GPIOs.

I could do what I want to do with discrete FETs but I don't want to have to design all the shoot-through protection, etc. etc. and worry about what I might have forgotten - much rather get a canned part.

If shoot-through is what I think it is - both ends of a half bridge being on together during a transition and shorting the supply - then there are plenty of half bridge drivers that look after that. Just connect to a pair of N-channel fets and drive with logic signals (typically H/L and enable). 2 should produce an H-bridge unless I'm missing something.

I've driven FETs up to > 350 KHz using these with no problems.

I thought it was just me - almost all the parts I want are no stock everywhere...

There are some other like the LMD18200 (approx. 40V/3Apk) chips aviable

Freescale MC33186 ST: L6203 Infineon TLE7209, TLE5205

all are more or less aviable @ digikey (Austria)

hth

- Michael Wieser