What passes as Pulse Width Modulation in DC Motor Control?

Can anyone please tell me what passes as PWM in Motor Control?

Is it limited to supplying the motor with a square wave, where a set and constant level of voltage is switched on or off to the motor?

Or can other waveforms consitutute PWM control of a dc motor? Such as where voltage does not drop to zero, but has periodic peaks having square, triangular, sawtooth or some other waveform?

Can PWM control consist of dc pulses that rise and fall in voltage in a sinusoidal fashion?

At the moment I am believing that PWM is just on/off of a contant level of dc voltage. But wonder if that belief is correct or not. Thanks.

Reply to
Rich
Loading thread data ...

--
Yes, but not necessarily a square wave, which would represent a 50% duty
cycle PWM signal.
 Click to see the full signature
Reply to
John Fields

Hi.

So PWM is not just simply switching on and off of a steady dc voltage.

If you employed a trangle wave, or sawtooth wave, or some other waveform, these waveforms require something other than a simple switch action. And therefore they cannot theoretically be as efficient as square-wave PWM.

But as long as there is no constant "pedastal" dc voltage in the dc supply to the motor, such waveforms would pass a PWM.

This is the logic behind what you said.

Reply to
Rich

--
Yes, it is.  

Pulse Width Modulation = constant amplitude, variable pulse width:
 Click to see the full signature
Reply to
John Fields

This is confusing actually when I think about it.

The question is asking is, is PWM limited to the switching on and off of a contant dc voltage.

You say "Yes,..."

Okay.

But then you say, "...but not necessarily a square wave, which would represent a 50% duty cycle"

I know about duty cycle, so I'm thinking JF is saying the dc voltage can be some other waveform other than square.

Which is a bit confusing because you said that PWM is limited to simple swithing on and off of a constant dc voltage.

"No." just tells me the dc voltage must go to zero at some point.

You confirmed that PWM *is" the switching on and off of a constant dc voltage.

What it was is that you saying above, "Yes, but not necessarily a square wave,.." Probably illogical, but it sent me off in a direction to think any dc waveform that was turned on and off constituted PWM.

Anyway gotcha. PWM is the on//off switching of a dc voltage that remains

*constant* in time.

Thanks.

Reply to
Rich

Square defined as 50% duty cycle. Other duty cycles are perceived as "rectangular" in this definition. Short on, long off. Long on, short off. Still going to constant voltage when switched on/off. Clearer now?

--
Cats, coffee, chocolate...vices to live by
Reply to
Ecnerwal

--
Yes, it is.

The problem seems to lie in the terminology, where a "square wave" is a
 Click to see the full signature
Reply to
John Fields

Well, I think technically PWM is "square" but I suppose in general anything could work.

The point here is that one can vary the function so that an average can be computed.

The time average is

1/(t1 - t0)*int(f(t),t=t0..t1) which is over one period but generally is taken to be infinite(they end up being the same though unless f is not periodic).

So the idea is that we have some parameter that we can adjust to modify the average.

The "standard" PWM scheme is to use rectangular pulses. It is defined as

f(t) = 1 if 0 < t - k*P < d for some integer k, else 0

P = period, d = "duty cycle" If you take the time average you get d.

This means that if f is a voltage then we can easily control the average voltage.

But we can actually choose f to be many things. The problem is that the relationship between the "control parameter" and the output becomes more complex.

You could use a triangular function such as

f(t) = 2t/d if 0 < t < d/2, -2t/d + 2 if d/2 < t < 2, else 0

then d controls the "base width" of the triangle.

Integrating over a period P gives

1/P*int(f(t),t=0..P) = d/2

(it's pretty easy to see because it's 1/2 the area of the rectangular pulse)

Now this is relatively simple but it's much harder to generate a triangular pulse than a rectangular pulse.

Essentially though any function will end up simply a*d where a is some constant

Reply to
Jon Slaughter

Hi. I thought that's where the problem lies. "Square-wave" to you was meaning the on/off waveform where the duty cycle is 50%. It really is square when that pertains.

I was simply meaning a waveform that had a very steep rise and fall.

All sorted I think. :c)

Reply to
Rich

Switching between a supply voltage and open circuit is also called "PWM", which is a bit of a misnomer but is in common use.

Otherwise, what John said.

--
http://www.wescottdesign.com
Reply to
Tim Wescott

I think you might say a sawtooth waveform or any waveform where there is a period of voltage cutoff, fails to qualify if there is no *sharp* cut-off of voltage.

It's the sudden "chopping action" that establishes a PWM voltage supply.

What I was trying to establish is what would pass as PWM motor control in

*DC motor control*.

Seems to me, that you make up your own mind, but many might say switching between a constant voltage supply and open circuit is normally what is meant by PWM in DC motor control.

Where the supply has a "pedestal" with peaks, no matter if the peaks are square-wave, that's clearly not PWM.

Some might think, as long as there is a sudden chopping of the supply voltage, irrespective of the waveform of the supply, be it DC , triangular, sawtooth, whatever, passes as PWM in DC motor control.

I think you will think it does. Rich

Reply to
Rich

No, it is the averaging. It's a way to reduce an input's value. PWM is one way, there are others such as PCM.

PWM is efficient because it achieves the reduction without wasting power. It only wastes it during the transition from high to low or on to off.

Because PWM is generallly done by using mosfets, which are very efficient switches... probably closest thing we have to mechanical switches, the best method is simply your rectangular pulses. This is the most efficient with method and generally the easiest.

Now, you can use triangular pulses too but you need to realize that it is much more complex to do and probably is not as efficient.

I think your trying to understand it back-asswards. PWM is a techinique for reducing the maximum output of something(usually voltage from a power source). There as many other ways such as a resistor, linear regulator, SMPS which uses PWM but makes it more useful for devices that need continuous power).

Check out

formatting link

There might be a technique definition for PWM but generally it shouldn't matter to much in practice. Rectangular pulses are almost always used because it is the easiest and most efficient method.

It seems you are failing to understand why PWM is actually used for motor control? (besides the obvious stuff you read) PWM is basically the idea of averaging. A solar sail has many photons that hit the sail. The average effect is force acting on the sail even though at any given moment no significant force is acting on it.

An engine has the same effect. The pistions are fired by "pulses" from the fuel igniting. At any given instance there is no force being created but on average there is and it is felt as if it were a continuous force. This is a form of "PWM"... at least abstractly.

V-----Motor----GND

If we just connect power to the motor it will run at full force. We cannot control it's speed. It is pre-determined.

But suppose we put a switch in there

V---S---Motor---GND

If we turn the switch on and off then we reduce the average current to the motor that simply depends on the off time to on time.

e.g., if we turned on the switch for 1 sec then off for 1 sec then we would get only 50 of the power to the motor(ideally). If it was 1:2 then it would be 33% of the power.

Since the power delivered is proportional to the speed(ideally) and hence proprotional to the current, we would expect the speed to be reduced!!!

this is quite amazing!!!!! We can reduce the speed without costing any power!!! If we used a resistor we would waste power in the resistor!!!!!

The method described is exactly what is done except a voltage controlled switch, e.g. a mosfet, is used instead of our hand.

Note that if we do it too slow the motor will be jerky but if we do it really fast it will seem continuous. This is because the motor has inertia and for the switch was off it will still move as fast assuming it is not off for too long.

Reply to
Jon Slaughter

--
Interesting.

You know nothing about PWM and all of a sudden you\'re telling us what
 Click to see the full signature
Reply to
John Fields

--
Nice post. :-)

The other thing that\'s nice about PWM\'ing a motor is that there\'s no
 Click to see the full signature
Reply to
John Fields

Hi.

What I meant by "It's the sudden "chopping action" that establishes a PWM voltage supply." was that turning off and on, in a mechanical switch-like manner, of any waveform, is a form of PWM. Because whatever the waveform is, it's being averaged by the switching on and off in a switch-like manner.

Therefore I agree with what you said: "Now, you can use triangular pulses too but you need to realize that it is much more complex to do and probably is not as efficient."

My understanding is that you probably would not want to go for anything other than turning on and off a steady dc voltage to achieve the efficiency possible with PWM motor control.

I did in fact wonder if the following circuit was a curious case where a waveform other than steady dc was being switched on and off. I even wondered whether there was any full turn off of the voltage to the motor.

formatting link

I was having trouble getting a hang of what kind of reference voltage the circuit was meant to be fed with, and what the sawtooth wave did to proceedings.

No-one yet seems to have explained this circuit and whether it is a PWM motor controller.

Reply to
Rich

I would say that PWM has to be bi-level, i.e. with negligible rise/fall times. One of those levels is usually zero, but I wouldn't exclude switching between two fixed, non-zero levels from the definition. After all, they're still pulses, and their width is being modulated.

E.g. for a analogue signalling circuit, if you have "off" = 1V (so you can distinguish "off" from a broken cable) and "on" = 10V, and the level is represented by the width of the pulse, it's still PWM.

Use of the term "PWM" in power conversion is a bit misleading to start with, as it isn't actually the width of the pulses which matters, but the mark-space ratio. Keeping the pulse-width (mark) fixed while changing the spacing will change the power.

Reply to
Nobody

"John Fields"

** Completely wrong.

PWM of the DC supply to a motor has the same effect on available torque and free running rpm as varying the DC voltage to that motor.

...... Phil

Reply to
Phil Allison

The chopping action is just a convienient method because it's easy to do. Mosfets are almost ideal switches and generally it's not a good idea to operate them in the linear region. There reason is because they dissipate more power(they act almost as a resistor).

Essentially it goes from 0 resistance to infinite resistance just like a normal switch but it has a resistance inbetween. R = 0 ==> no power dissipation. R = inf ==> no power dissipation. Else There is power dissipation. Generally this is not a good idea for motors because of the large currents. (I'm speaking in ideal terms here because in reality a mosfet takes time to switch on)

Yes, this is generally the case.

Note they are not using mosfets!!! This should tell you that they probably are not using PWM. (rarely would you use anything but mosfets for PWM)

They are simply using a current driver for the motor. They are using feedback to determine how much current to dump into the motor using the bjt's. The bjt's are current controlled current sources. They dump so much current it, if the motor is going to fast they reduce the current.

It's not PWM but a simple current driven system with feedback. Think of it as a current power source supplying the motor.

This is not the best method in general but maybe for some specific application it is good.

So here you are getting confused in thinking that they are using PWM when there is no pulsing.

In some sense this is just an active "resistor" where the resistance is proportional the motor speed. 0 rpm's means 0 resistance => causes motor to spin faster. large rpm's means large resistance => cause the motor to spin slower.

PWM works differently but abstractly it is the same. I wouldn't say it is PWM because it is not pulsing anything. There is no "duty cycle".

Reply to
Jon Slaughter

Thanks. I think what you wrote is beginng to help me understand that particular circuit. I need to ponder what you said until I fully grasp things.

As a novice, I think I just happened to pick on a tricky circuit to explain. I was looking for a PWM circuit that had speed regulation and I thought that perhaps that circuit was such a circuit. It now seem I have the answer I have been seeking - this is not a PWM controller that will drive well at slow speeds, unlike PWM that tends to do that quite well.

Just one more thing: What do you think the reference input is meant to be where it is marked "Input voltage from computer"?

Reply to
Rich

Yes, its an on/off pulse. The duty cycle dictates the resulting voltage through a filter.

Many circuits can filter a PWM signal to produce useable DC control even though the circuit wasn't designed with that in mind.

formatting link
"

Reply to
Jamie

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.