How to choose the frequency for a PWM speed control ?

[Externet]

The pulse widht will determine the speed a DC motor runs, but what dictates which frequency ?

It is for a 130VDC / 2.75 HP 21 Amperes maximum treadmill DC motor that will operate from a bank of batteries at 72 Volts.
How to calculate the most convenient frequency ; and what undesirable behavior happens when frequency is over/under the optimal ?

 

[Reloadron]

First you may want to give this a read. The PWM frequency used is going to be a function of the motor. Next, you mention a 130 VDC motor and a 72 VDC battery bank. PWM runs a motor at its rated working voltage by turning that voltage on and off at a fast rate. How will your 130 VDC motor respond to a 72 VDC PWM signal?

Ron

 

[ronv]

It can get complicated. There are probably a lot of people with more experience in this than I have, but oh well.
A lot depends on your motor and your requirements. A lower frequency will give better low speed torque and start and run at a lower speed. But to low and power is wasted in the form of heat.
To high and the motor will tend to have more trouble starting at low speeds.
In between is good but audible noise may annoy people.
How much help was that?
Try 1 to 10 KHz and see what you think.

 

[misterT]

The PWM frequency should be selected so that it does not create too much current ripple in the motor windings. Current ripple results in heating of the motor.
In theory there is no drawback in using very high frequencies, but in practice high frequencies mean more trouble with the driving circuitry, not the motor iself. (more complicated circuitry and switching losses).
So, choose the pwm frequency so that the ripple stays "small enough".

Also, Low quality dc-motors usually have high friction and therefore need a high current pulse to get started. That is why low frequency PWM seems to work better with some motors. But, lowering the PWM frequency just to get that "starting torque" is wrong.. you will end up with a current ripple which overheats your motor. (DC) Motors always run better, smoother, cooler and quieter with higher pwm freq.

Here is some theory, but you really don't need that. Just measure the ripple current at 50% duty and choose your pwm frequency so that it is "small".. I mean.. maybe 5% or less from the full swing.
**broken link removed**

And, if your electronics can't do "high enough" pwm frequency, then a perfectly good solution is to add inductance in series with the motor. That smooths out the ripple also.

 

[alec_t]

So some motors might benefit by having the PWM frequency (as well as the duty cycle) increase as the speed increases? Surprisingly, a quick google doesn't suggest this is common practice.

 

[misterT]

So some motors might benefit by having the PWM frequency (as well as the duty cycle) increase as the speed increases? Surprisingly, a quick google doesn't suggest this is common practice.

If you are referring to post #3..
It might look like that the motor runs better at low PWM frequencies when high torque is needed at low speed. But when using too low frequency the only thing that is "smoothing the action" is the mass of the motor and load.. that is.. the mechanical time constant. The current in the motor winding is switching from "maximum" to "zero" at the PWM frequency. You are not actually controlling the torque, you are just switching it to "full blast" or "zero".. even when you only need small torque.
If you need more dynamic action from your motor, then better solution is to use some kind of feedback loop. Good thing is that motors are inherently integrators.

 

[MrAl]

So some motors might benefit by having the PWM frequency (as well as the duty cycle) increase as the speed increases? Surprisingly, a quick google doesn't suggest this is common practice.

Hi,

What made you ask this question?

Also (not related to your question), if the PWM circuit is made right it seems to me that the PWM frequency should not hurt the start up. That's because if the PWM is made right then it should be able to get up near 100 percent duty cycle, regardless of frequency, and that's full power applied to the motor. If there is too much dead time then that would make lower frequency more attractive, but then so would a better PWM circuit.

 

[alec_t]

What made you ask this question?

The post #4 statement that
"Also, Low quality dc-motors usually have high friction and therefore need a high current pulse to get started. That is why low frequency PWM seems to work better with some motors. But, lowering the PWM frequency just to get that "starting torque" is wrong.. you will end up with a current ripple which overheats your motor. (DC) Motors always run better, smoother, cooler and quieter with higher pwm freq."
I took it that "the PWM frequency" was referring to a fixed frequency. I agree that a fixed low frequency might not be good for a motor.

 

[ronv]

Ok, I think we are all in agreement - now back to your question.
Do you need it to start and run at low speed? Maybe you could test it on 1 battery at a time to see if the speed and torque is ok.

 

[MrAl]

Hi,

Wasnt his original question about what undesirable effects would be present with too high or too low a frequency?

I would think too low would result in higher current pulses to the motor, as if it had to start and stop every cycle. That would heat up the motor and could even cause problems in the load. Too high would result in more switching losses.

So what is the best? I would think it requires a measurement. Measure the speed with a given load at higher and lower frequencies, and watch the current waveshape so that it does not rise and fall too much, and watch the switching speed of the controller circuit. That should turn up a range of operating frequencies that work ok.

 

[ronv]

I assume he has an application in mind since he has a motor and a voltage. Other than that we don't know much about what he wants it to do or if it is open or closed loop. I would think open, but maybe not.

 

[4pyros]

And here again a 130 volt motor is going to be way under power running off 72 volts.
Thats almost half the voltage and only 1/4 of the rated power, less then 1hp.

 

[Externet]

Thanks, gentlemen.
Yes, will need to start under load; it is an electric quadricycle I built with 2 joined bicycles and now am tackling the PWM. Open cycle, less than 1 HP. If not enough power, will add more batteries (all riding mower Pb type) later.
Will take a picture one of these days.
Edited, added: A couple of opinions found:

https://electronics.stackexchange.com/questions/80153/calculate-dc-motor-frequency
https://www.electro-tech-online.com/threads/suitable-pwm-frequency-for-motor-control.19246/

 

[4pyros]

less than 1 HP. If not enough power, will add more batteries (all riding mower Pb type) later.

Then your controller will need to be rated at 120 or 130 volts.

 

[ronv]

Here is a cheap and dirty circuit you can use. I put some cap values in if you want to try various pwm frequencies.
The FET and big diode will need a heat sink