Silencing a PWM Speed Regulator

How about raising the oscillator frequency above 20kHz!?

"How about raising the oscillator frequency above 20kHz!?"

Tried that. Raising the frquency just raises the pitch.

You just gotta take the frquency high enough so that over the range you use it doesn't excite the coils at audio frequencies.
I realise that this is easy to say, but may be nigh on impossible to achieve.
I had the same prob' with my golf trolley speed controller...mk 1 was driven by a microcontroller, but it couldn't go quick enough to get out of the audio band...I ended up doing it with CMOS oscillator stuff.
Wire a sig gen and power amp up to it and sweep it....or tickle it with a feather duster.

I expect you know all this already...but hey sue me.
Del

Your ideal speed is between 4 & 20 RPM. If the pot will allow a much greater speed change the belt to pulley ratio. May then be able to push the frequency of the noise above audible level.

"If the pot will allow a much greater speed change the belt to pulley ratio. May then be able to push the frequency of the noise above audible level."

Two of you suggested that and I think it's a good idea. The motor certainly has the torque needed to continue running under load, even without the belt and pulley. However, I loose some of the precision in regulating turntable speed but that isn't critical.

It's not the fix I was looking for but what the Heck. . . .

Thanks

L.J.

If you want to eliminate the source the best would be to make a few tests:

- change the PWM frequency (if possible measure it) and see if the "quite speed" for the motor does change or not and measure it as well.

The resonance could be also due to a subharmonic generated by the combination of PWM frequency and motor speed. It is possible that the amplitude is amplified by a component of supporting structure. I would not, from the start, take the belts off since due to their compliance they have a beneficial effect on the speed regulation or better to say on the uniformity of table rotation (combined with table inertia) working as low-pass mechanical filter. It could be possible by taking the belt off that you allow some higher frequencies to enter the table system and deteriorate the surface grinding quality.

If the noise is related to the motor speed this means that the system has a mechanical resonance some where.

Could you change the transmission so that the motor will work in the "non noise" range?

Is the "quite speed limit" related to the PWM frequency?

I assume this is a control you purchased but there is a possibility that if it is a current regulating device, it may be a parasitic audio frequency oscillation in the regulation loop that doesn't affect it's operation but induces a squeal in the motor. You could contact the manufacturer and discuss this or it it your own design you should take a look at the output of your error amp to see if you have high frequency AC.

Otherwise, you might try to find a different motor that is wound tighter or tighten any frame bolts on the existing one since in order to radiate audio something inside the motor must be vibrating at that audio frequency. Potting the windings would be about the only way to keep the wires from vibrating. Good luck.

Your idea about the motor could be in the money to my mind, furthermore, the OP should realize that the noise generated will eventually cause a wire to break (fatigue) in the motor as it is flexing at this time to make the noise.

The output to the motor should be checked on a scope for illegal low frequency components....compare the noise generated to a signal generator (connected to a small speaker), to know what frequency you are looking for......if it is the PWM frequency, you need to change a component or two to increase the frequency dramatically.....

A significantly higher frequency should fix the problem, or a new better quality motor where all the coils on both rotor and stator (if not a PM motor), need to be so firmly embedded in the "lacquering" (as it used to be called, probably Epoxy today) so that they cannot move....taking the motor apart and epoxying any loose strands may help, but may cause an imbalance in the rotor.....

Best of luck to the OP.....

I will throw my hat into ring with the higher motor speed too. It will be the cheapest and most effective way to accomplish the desired results.

All the fixes suggest raising the frequency... however, then what may happen- is instead of the people around you howling, the dogs around you will start howling, and you'll still have the problem of irritating the people around you!

Why would dogs hear it? The frequency at which motor wiring can act as a loudspeaker are to my mind going to be mostly relatively low. Even medium loudspeakers have problems reproducing such frequencies....

But if that does turn out to be a problem, then the frequency needs to be upped even more well beyond a dogs hearing levels.....

Anyway, whats a dog doing at the business?

I have played around with PWM and I found it quite easy to increase the frequency and stop the motor "squalling"......it did not affect the function seemingly, but it was only upto 20 KHz or so, a dog might have heard it I feel, even if I did not!!

Jaguar, You mentioned two things that make me wonder about the output of your PWM speed controllers. First that varying the speed makes the motors get softer or louder, and second, that when you tried to increase the frequency of the PWM that your workers could hear the pitch or wine of the motors go up. Some responders mentioned moving the PWM frequency up to around 20 kHz but I suspect yours are operating at a very much lower frequency, perhaps around 1 kHz or so. In this audio range the only way to quiet the motor windings is to improve the quality of the DC fed to them. Since your motors are just variable speed and not truely servo motors. There is'nt any feedback to the controller except for their ability to sence the current draw. All this means that you can probably just filter the output with a capacitor (sonething like 20 k ufd electrolytic at 40 wvdc) also to help isolate the PWM, you might try feeding the output through a power diode before the filter cap. I have a small metal lathe that I run with a treadmill motor (1 1/2 Hp) that had this problem. I solved it by trial and error gradually increasing the filter capacity to where the motor ran quietly. In that case 50 k ufd, and the motor and power supply have been living happily together for the past 5 years.

Hi Zamaron!

Here's what is posted at his e-Bay web site:

Technical information.

- Power Supply 8 -30 VDC
- Frequency Adjustable 400HZ to 3kHz
- Jumper select Fixed at 100Hz or Adjustable 400Hz to 3kHz
- Duty cycle 0-100%
- DC motor current up to 30A ( Max )
- LED power on indicator
- Including Plastic Knobs as photo
- 2 Power MOSFET with Big heat sink

I am using the unit "as is"; just as it came in the box.

I used a laboratory quality 24VDC supply to feed this unit initially. The plan is to either buy two 12 VDC car batteries for it or simply build a transformer isolated full wave rectifier with a capacitor across the output to clip the signal some.

Your suspicion that the PWM is operating at a low frequency was correct.

Here is a picture of the unit

Thanks

L.J.

Sometimes a 555 is used to make such a unit, made a few myself for various friends/reasons. If there is a 555 on the PCB, we may be able to add an extra resistor in parallel with one already there and significantly increase the base frequency.

Can you look to see if any of the chip(s) have the number 555 somewhere on them?

Or I can send you a really simple circuit/schematic I use to make such a device, that I found on the internet....I will need your normal email address, send it via CR4 email, do not post it here....

Hi Andy!

Here is a circuit diagram provided with the PWM unit.

I figured this would be more useful.

Thanks for the support. All of you!

l.J.

If I have understood the circuit correctly, the first OP Amp is the Oscillator and the second part is the variable mark/space ratio (speed).

I have made a slight change to the circuit with an extra resistor and an extra variable resistor in parallel with the original circuit. Please send me your email address via CR4 email.

Set the original pot to the highest frequency and then adjust the new pot for the best frequency, which will be much higher.....

The reason I have made it in this manner is so that you can make up a small PCB with the two components on it and attach it via 2 wires to the original circuit, so you do not need to modify the original circuit at all.

But you could also remove the 10 k resistor involved (I cannot read the resistor name CR4 does not allow clear copies) and replace it with a 2K and the effect would be roughly the same....its the same one that you will parallel with the circuit I have modified/suggested.

Any questions, just ask.

I cannot be 110% certain that it will be stable and work as hoped without building a similar unit, but at least you can easily return the circuit back to its original state if thats the case.....

Thanks Andy.

Leaving for a long weekend in Florida. I'll be in touch w you next week.

L.J.

Andy I've been out of action for over a month. A combination of projects at work, a very aggressive bronchial infection (no not Swine Flu) and an illness in the family, all combined to keep me out of here.

There is no question that the design and materials used to mount the motor have created a massive sounding board that amplifies the noise. I will address this first.

I have managed to obtain a set of low durometer isolation mounts which I expect to tone things down quite a bit.

These simply address the symptoms, not the source. It's been suggested that constant exposure to these high amplitude cyclonic forces might eventually cause damage to the motor.

I know enough about these things to suspect that the caveate merits further study. If it's possible to change the frequency without causing ill side effects, that alternative should be explored. Fortunately, I have a very generous EE at the Lab who I can take to lunch.

I'm going to disappear again for a little while. Thanks

"I shall return" with an update of results.

Thanks for your interest and support.

L.J.