Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!elroy.jpl.nasa.gov!sdd.hp.com!hplabs!hpcc05!hpldsla!tonya
From: tonya@hpldsla.sid.hp.com (Tony Arnerich)
Newsgroups: sci.electronics
Subject: Re: motor speed controller, transistor question
Message-ID: <1990012@hpldsla.sid.hp.com>
Date: 3 Apr 91 18:59:41 GMT
References: <68137@eerie.acsu.Buffalo.EDU>
Organization: HP Scientific Instruments Division - Palo Alto, CA
Lines: 51

The motors won't go faster with higher 555 output frequency, if the duty
cycle is the same. It is even possible that *less* power gets to the motor
at high frequency, due to greater inductive losses. Set the chopping fre-
quency and vary the duty cycle. A good frequency to use is at least 500 Hz,
and not so much that all you do is fight motor winding inductance (<30 KHz
for a small swiss motor I have).

For variable speed on the car, it might be best to think about varying the
motor *torque* as opposed to the free-running speed. In that mode, you get
constant speed operation when the motor power equals the power requirements
due to drag losses. Want to accelerate? Put in more power. Braking of a sort
is accomplished by cutting power. Real braking comes from shorting the motor
windings (not the power supply, though!).

You can use a 555 timer as a pulse-width modulator. I haven't tried that
personally, but I have driven DC motors using an H-bridge of power MosFets
controlled by a 3524 PWM chip. Works like a champ. I noticed that motor
speed was proportional to pulse width when there was a significant drag on
the motor. If left to run free, motor speed was essentially constant for a
wide range of duty cycle.

You can get power MosFets from International Rectifier (such as the IRF530
and IRF9530) that have only 0.6 to 0.3 ohms on-resistance. That's bypassed
for just about all intents and purposes. Not bad for something that can be
driven directly by CMOS. Just be sure you drive the gates with at least 9V;
12V works extremely well. 4V is just barely on (high resistance).

Which diagonal pair of transistors in the H-configuration is on determines
motor drive direction. You can gate the on-signals (transistor gate drive)
of either the P-channels or the N-channels from the PWM control signal to
set the duty cycle.

I'm putting the finishing touches on a synthesized sinewave power inverter
that uses exactly this power output stage. The only difference is that in
my application, the "motor" (transformer winding) changes direction 60 times
a second, and the power output varies sinusoidally all the time. The circuit
worked out very well.

Yes, you can parallel some transistors. It's a real pain with bipolars, because
when they get hot they go all the way and fry. MosFets, at least those I am
familiar with (IR), are self-regulating and are inherently safe to parallel.
No external circuitry is required, just bus them. That's exactly what is done
in commercial motor drivers for RC - look for the gang of TO-220 packages. I
understand from a RC-enthusiast acquaintance that they can handle 400 amps.
You need that when you have inductive loads at high power. Don't forget the
protection diodes!

For more info, get IR's Power MosFet (HexFet is their copyrighted name)
data book, and Motorola has an application note on the 3524 PWM controller.

tonya@hpldsla.sid.hp.com