PulseWidthModulation comments

[Raymond C Drouillard]

A common theme here is that people keep forgetting about the freewheeling
diode.  With a freewheeling diode, the current continues to flow (due to
the inductance) when the <transister, SCR, FET, whatever> is switched
off.  The inductive reactance doesn't inhibit the performance of the
motor.  In fact, it helps to smooth out the current.

If the frequency and inductance are high enough, the current will not
drop very quickly when the transister is turned off.  Also, it won't rise
very quickly when the transister is turned back on.  The motor will see a
DC current with a very small AC component.

A switching regulater (like the one used in this computer) works the same
way.  It consists of a switching transister, a freewheeling diode, an
inducter, and a feedback circuit.  The goal, in this case, is to get the
AC component (ripple) as small as possible.


+ supply ---<switch>--------<motor(inducter)>-----|
                        |                         |
                        |                         |
                        |---------|<|-------------|
                           freewheeling diode     |
                                                  |
                                                -----
                                                 ---
                                                  -

When the switch is off, the current in the incuctor will want to keep
flowing.  This forward-biases the freewheeling diode (rather than
creating a really nasty inductive "kick").  Current continues to flow
through the diode until it decays, or the switch is switched back on.

That way, the current through the motor, rahter than being chopped from 0
to maximum, actually only changes a fraction of a percent to a few
percent - depending on the inductance and frequency.

Ray Drouillard




On Thu, 4 Feb 1999 14:36:29 -0600  Terry_Sare at dell.com writes:
>>From another list for robots: PWM comments. Funny that it shows up on 
>two
>different list at the same time.
>
>>>>>>>>>>>>>>>>>>>Inserted from srs list <<<<<<<<<<<<<<<
>Hi,
>
>        The speed of the PWM for speed control of a DC motor is 
>*highly*
>dependent on the motor you are controlling.  For example, the little 
>3/8
>inch
>diameter motor that I use require a PWM frequency of at least 20 KHz 
>(which
>isn't doable with a standard HC11 setup).   Luckily, most of us aren't
>pushing the performance envelop, so most of the typical motors used in
>mobile
>robots can be run with a wide range of frequencies.
>        As far as the inductance, if you run the frequency up, the 
>inductive
>field doesn't collapse as much, so the impedance isn't such a problem.
>Depending on the motor, there are bad ranges, however.
>
>                Cheers,        Bill Harrison
>                                    http://www.sinerobotics.com
>
>Peter McCollum wrote:
>
>> 1000 ticks at 8 MHz would be 8 KHz, right? I'd say that that's much 
>too
>> high. The inductance of a DC motor is probably high enough that the 
>8 Khz
>> signal is seeing a very high impedance. Thus, very poor power 
>transfer.
>>
>> I would try 20,000 or 30,000 ticks instead - giving you 300-400 Hz.
>> That's the range I'm using on two projects so far, and it seems to 
>work
>> OK.
>>
>> Pete M.
>>
>> On Sun, 31 Jan 1999 22:27:08 -0500 (EST) Derek Konigsberg
>> <octo at evcom.net> writes:
>>
>> >       Anyways, does anyone know what period works best for maximum
>> >torque in PWM motor control?  Right now, low speeds yield almost no
>> >torque
>> >and I can effortlessly stop the wheels with my hands.  I'm using a
>> >period
>> >of 1000 clock ticks (8MHz clock) on a 68HC11E1.  My high/low (duty
>> >cycle)
>>
>> ___________________________________________________________________
>>>>>>>>>>>>>>>.end insert
>-----Original Message-----
>From: rauscher at icst.com [mailto:rauscher at icst.com]
>Sent: Thursday, February 04, 1999 1:21 PM
>To: DIY_EFI at efi332.eng.ohio-state.edu
>Subject: Re: PulseWidthModulation comments
>
>
>
>Ludis wrote:
>>Years ago as a summer intern I worked on a motor controller which 
>used
>>PWM.  A brushless DC motor ran on up to 120 to 170 volts at up to 10 
>to
>>15 amps.  A switching (PWM) power supply controlled the voltage 
>applied
>>to the motor.
><snip>
>>An improved version used a microcontroller controlled digital PWM.  I
>>think the PWM frequency was about 20 KHz and the PWM had 5 bits of
>>range.  The same microcontroller monitored the motor position/speed 
>to
>>form a closed loop system.
>
>Was this motor designed specifically for this freq? When I first 
>developed
>this PWM motor controller, I was trying to run it at 20KHz also. But 
>the
>motor was down on power. Motor was an AstroFlight cobalt, 9.6V at 
>35Amps,
>such as those used in R/C model aircraft. Going to the lower freq 
>brought
>the power back up. I looked at possibly skin effect, but the wire used 
>in
>the 
>winds wasn't that large.
>
>>Some comments applicable to fuel pump / coolant fan PWM control:
>>
>>The pulse width and frequency needs to be very stable, otherwise all 
>the
>>inductors in the circuit will squeal like stuck pigs.  The original
>>analog system screamed constantly.  The digital system (running the 
>same
>>switcher!) was totally silent as long as the pulse width wasn't
>>changing.  Software in the microcontroller updated the pulse width 
>every
>>millisecond.  When the pulse width needed to be changed often, this
>>produced a 1 KHz tone from the electronics.  The amount of noise was
>>directly proportional to the load on the motor.  You could actually 
>hear
>>the computer "straining" against a high load.
>
>Sounds as thou the noise is from the P/S, and not the motor. I've only
>heard a light whine from the motors, and only at lower speeds. So 
>hopefully
>the noise won't be enough to drive folks out of the car <g>.
>
>>The main switching transistor was a TO-220 package mounted on a dinky
>>little heat sink.  You don't need a massive TO-3 transistor.  In the
>>above system, cooling the rectifier diodes was actually more of a 
>problem.
>
>Oh so true, in the writeup I showed that even at 15 Amps, the MOSFET 
>only
>needed to dissipate a little more than 5 watts. Only a small heat sink
>required.
>The key to the proper MOSFET, is low voltage drop through it. This 
>keeps
>losses down and heat also.
>
>>BTW, in one mode, this motor was reversed every several tenths of a
>>second.  This applies to a thread from several weeks ago.
>
>My brain is fading here, possibly relative to a motor driven FPR?
>
>BobR.
>
>--
>
>

___________________________________________________________________
You don't need to buy Internet access to use free Internet e-mail.
Get completely free e-mail from Juno at http://www.juno.com/getjuno.html
or call Juno at (800) 654-JUNO [654-5866]