DPRG
DPRG List  



[DPRG] Cheap high power H-bridge

Subject: [DPRG] Cheap high power H-bridge
From: Chuck McManis cmcmanis at mcmanis.com
Date: Tue Nov 27 12:22:06 CST 2001

At 05:36 AM 11/27/01, D. Daniel McGlothin wrote:

> > Interesting idea. Having been in the throes of designing a FET based
> > h-bridge I thought I'd check it out.
>
>Are you aware of the OSMC http://groups.yahoo.com/group/osmc, an Open Source
>Motor Controller.

Yup. And it is a damn fine project. The mailing list and the board layouts 
are excellent. Much better than the 'cheap' one mentioned at the beginning 
of this thread.

>Quoting from their FAQ (in
>     While the design has not been battle tested yet, it is expected to have
>an
>     upper limit of less than 100A. To date, extensive testing/usage has been
>     in the 60 Amp range.

This is an issue of course. Two things have to be true, one you have to do 
a thermal model of your FETs and then you have to validate your model. On 
my controller <http://www.mcmanis.com/chuck/robotics/projects/speed2.html> 
I spent a lot of time doing that and found out many misconceptions about 
MOSFETs (like the 'max Id' is in fact the current the FET can carry, it 
isn't) For TO-220AB lead frames the metal "legs" melt at 75 amps. 
Interestingly you can do better than this (get more current) if you fold 
the legs flat and solder as close to the "body" of the FET as possible.

But the trick here is that to _test_ it, you really need a 100 - 150Amp 
power supply that can deliver 100 - 150Amps all the time. I use two HP KW 
supplies (0-60V, 0-60A) with one "slaved" to the other to provide a maximum 
of 120Amps. I know of at least one person who salvaged a 5v 100Amp supply 
out of an old mini-computer and rewired it to deliver 12V @ 50amps. Another 
solution of course is a pile-o-batteries. That limits your test time, but 
its a whole lot cheaper to set up.

I've been tempted to bring my test setup to BattleBots and issue the "speed 
controller challenge" which is to say, tell me what you think your speed 
controller is rated at, and we'll run it at 90% of that capacity and see 
how long it lasts. :-) But that would be kind of silly. If the speed 
controller works for your application you don't want to mess with it.

>2.5) What is the maximum rated voltage?
>
>     You can go up to a maximum battery voltage of 50V. This is a function of
>the
>     (typical) maximum voltage the FETs and FET driver chip used.
>
>     Note: Batteries have a significantly higher voltage "fresh off the
>charger"
>     so you won't be able to use 48V worth of batteries with the OSMC.  It is
>     recommended that you limit your "rated" battery voltage to 36V or so.

There is also the margin for turn-around. So you've been sending your motor 
"forward" and have built up a huge back-emf field, now you switch it to 
reverse. Your low side FETs will "see" the voltage of the input + the 
voltage over the motor for a brief period of time. That can be annoying if 
you're running near the FETs rated voltage.

--Chuck


More information about the DPRG mailing list