0
\$\begingroup\$

This is a hypothetical question but it's been bugging me for a while.

Lets say I hook up a DC motor with a propeller on it to a 10v battery and let it run.

Then lets say I hoop up a second identical DC motor and prop to a 100v battery, but with a PWM controller. And lets say I tuned the PWM duty cycle so that the RPMS of the motors are identical.

The power output of the props should be identical now as well, since it's related only to the RPMs.

Lets also say that hypothetically the mosfets in the PWM controller have 0 on resistance, and switch instantly.

Will both systems be identical in efficiency? Or are there some inherent losses due to the PWM throttling?

\$\endgroup\$
1
\$\begingroup\$

Lets also say that hypothetically the mosfets in the PWM controller have 0 on resistance, and switch instantly.

Will both systems be identical in efficiency? Or are there some inherent losses due to the PWM throttling?

You haven't stated the PWM frequency. But assuming it's around 100 kHz, then it's very close to being the same as if you didn't use PWM.

If it's in the MHz - GHz region then you can expect some of the energy being radiated outwards because some of your wires will act like an antenna. I'd call this a "loss".

If it's in the sub 20 kHz range then you can expect to hear the PWM sound, this might drive you mad. I'd call this a "loss".

If it's in the 30 kHz range, you won't go mad, but your dogs might. (It's like giving them tinnitus. I'd call this a "loss".


When your transistors are off, the motor will act like a generator (if it is rotating), giving current back to your system. This means you will need some clamping diodes to guarantee that the terminals don't reach unsafe voltages. So that's some losses, depending on how you handle that excessive current.

\$\endgroup\$
  • \$\begingroup\$ Hi HS, for the record, the idea that PWM should use frequencies over 20 kHz so you can't hear a tone is a myth. It depends on what the loading (averaging) device is and how it behaves. I've used it on solenoids and transformers and you hear tones when well above 20 kHz. And I have a phone charger running at 100's kHz or a MHz or so and that whistles away, with the pitch clearly changing with load. \$\endgroup\$ – TonyM Dec 8 '17 at 22:24
  • \$\begingroup\$ Oh the memories of being called HS (Head Shot). Reminds me of when I went to school. - Well I designed a buck converter with 4 kHz PWM, and I could clearly hear it switching. Then I designed one in the 500 kHz range and I could not hear that one switching. - So I just went with my experience. I am known to have bad hearing so I can definitely be wrong. - But I did manage to identify where the noise came from. It was cheap ceramic capacitors acting like piezo speakers. - Maybe the 500 kHz had some better capacitors, maybe I'm going deaf, maybe you got super hearing. I can definitely be wrong. \$\endgroup\$ – Harry Svensson Dec 8 '17 at 22:31
  • 1
    \$\begingroup\$ I'll stick with 'Mr Svensson, sir' then :-) I've done a reasonable amount of DC motor control and we kept PWM to 2 kHz, as recommended by our large (350M units/pa) motor manufacturer and also noticing this was the common frequency in remote control. Definitely hear it, though. With solenoids we thought it was tiny movement of the coil, that's where it was coming from. Cheers. \$\endgroup\$ – TonyM Dec 8 '17 at 23:08
  • \$\begingroup\$ The other issue with electromechanicals is you can excite much lower frequency resonances in the motor, and elsewhere, even when you are switching in the MHz region. It also happens if you are speed controlling a motor that the carrier modulation frequency used to control the speed can create audible noises. That happened to me once on a stepper motor project, PWM was silent, speed control created a very distinct hiss. \$\endgroup\$ – Trevor_G Dec 9 '17 at 0:22
1
\$\begingroup\$

Yes, there are eddy current losses in the motor iron, if you consider just the motor.

But it's not really a fair comparison to just draw the box around the motor. You should consider the efficiency from mains or battery source input to shaft horsepower output, and consider power factor too in some cases. Generating a smooth-ish DC voltage from the mains or a battery efficiently generally involves PWM and filtering, and the filter inductor will have losses and the switching losses will occur in either case.

\$\endgroup\$
  • \$\begingroup\$ Doubled up, tks. \$\endgroup\$ – Spehro Pefhany Dec 9 '17 at 2:28
  • \$\begingroup\$ So you're saying that there will be eddy current losses in the motor being PWM driven that will not exist with DC drive? I'm not concerned with the realistic efficiency of the system, what I really want to know is how PWM compares to no PWM assuming ideal components. For example, an ideal led powered through an ideal resistor is not 100% efficient, there are inherent losses even if the components are "perfect". I want to know if PWM motor drive has any similar losses. \$\endgroup\$ – Drew Dec 13 '17 at 21:58
  • \$\begingroup\$ Usually you keep the PWM at a high enough frequency that the current ripple is relatively small. Because of the motor inductance. This is a rather important detail because I^2R losses increase with RMS current, not average. Depends how simplistic your models are. \$\endgroup\$ – Spehro Pefhany Dec 13 '17 at 22:56
0
\$\begingroup\$

Lets also say that hypothetically the mosfets in the PWM controller have 0 on resistance, and switch instantly.

What's the point of making that assumption? That's EXACTLY what would be different: it's called switching losses.

\$\endgroup\$
  • 1
    \$\begingroup\$ But at low frequencies, the FET switching losses would be so tiny as to be irrelevant. The OP's question is on more than that. Downvoting I'm afraid, this isn't answering the question about PWM vs continuous voltage. \$\endgroup\$ – TonyM Dec 8 '17 at 22:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.