Air pump too fast at minimum pwm

alankilian · 23 Jan 2021, 16:59

@alankilian said in Air pump too fast at minimum pwm:

I think things will look better if you connect the scope ground to ground and scope-signal to E1-.
You'll see a zero-volt signal when the pump is "on" and a woozy-doozy signal probably getting up to 24 Volts when the pump is "off".

PERFECT!

I think you've demonstrated that everything is working exactly as it's supposed to even at 22 KHz.

That minimum of 10% was certainly buried in the details. Thanks @fcwilt for finding that.

Now, how do we get your darn motor to slow down more?

Or, can you just make a "leak" in the output hose to let off 50% of the air and then you'll be able to control it enough?

Or can you restrict the output through an orifice to limit the air?

We'll get you there somehow.

Great investigative work.

fcwilt · 23 Jan 2021, 17:13

@Nxt-1

Hi,

Are you sure the L0 parameter took hold?

The image labeled 5% at 2kHz looks like10% - which is the default minimum - what you get with no L parameter.

Frederick

Nxt-1 · 23 Jan 2021, 17:39

@fcwilt said in Air pump too fast at minimum pwm:

@Nxt-1

Hi,

Are you sure the L0 parameter took hold?

The image labeled 5% at 2kHz looks like10% - which is the default minimum - what you get with no L parameter.

Frederick

You are right in between setting the different frequencies the L parameter got overridden. If I set it again and look at the trace it has a half as long off period than before, as it should be. The rise time the same or similar.

fcwilt · 23 Jan 2021, 17:40

@Nxt-1

So can you get the pump running slow enough now?

Frederick

Nxt-1 · 23 Jan 2021, 17:45

@fcwilt said in Air pump too fast at minimum pwm:

@Nxt-1

So can you get the pump running slow enough now?

Frederick

Yes, no, maybe? I think it's slow enough now, it does make some progressively unhappy noises as low percentages as I drop the frequency. For now, I have some things I really need printed so I am going to roll with it @2kHz I think

For sure many thanks for those who contributed.

fcwilt · 23 Jan 2021, 17:48

@Nxt-1 said in Air pump too fast at minimum pwm:

Yes, no, maybe? I think it's slow enough now, it does make some progressively unhappy noises as low percentages as I drop the frequency.

If a motor is designed to run on pure DC then PWM DC can make it sound unwell.

There are filter circuits you can add between the board connections and the motor that can smooth out the pulses giving something more like DC.

It might make a suitable improvement.

Frederick

alankilian · 23 Jan 2021, 19:43

@fcwilt said in Air pump too fast at minimum pwm:

PWM DC can make it sound unwell.

Also, 2,000 Hz is a pretty annoying frequency.

dc42 · 23 Jan 2021, 20:08

Those traces look much better than the previous ones. Can you confirm that the motor was connected when you took those traces? If so, then I suspect that the slow rise time is caused by an interference suppression capacitor connected across the motor terminals, possibly inside the motor itself.

Nxt-1 · 23 Jan 2021, 23:29

@dc42 said in Air pump too fast at minimum pwm:

Those traces look much better than the previous ones. Can you confirm that the motor was connected when you took those traces? If so, then I suspect that the slow rise time is caused by an interference suppression capacitor connected across the motor terminals, possibly inside the motor itself.

Yes, motor and diode were connected for those traces.

dc42 · 24 Jan 2021, 09:41

In that case, I suspect an interference suppression capacitor is causing the slow rise times, or possibly just the capacitance of the motor windings. When the mosfet turns off, the motor inductance, resistance and capacitance from a LCR resonant circuit that start to ring; but the diode clamps the oscillation it when it reaches 24V.

The basic issue you are having is that lightly-loaded DC motors respond more to the peak voltage they are supplied with than to the average voltage.

If you never want to drive the pump as maximum RPM then the simplest fix may be a power resistor in series with the pump; or you could power the pump positive terminal from 12V from a buck regulator instead of 24V. If you do need the full range then I suggest an LC filter between the Duet and the pump, with the flyback diode connected to the input (Duet end) of the filter. However, you will need to use quite a large inductor. Even at 65kHz PWM frequency you will need at least 1mH with a suitably-high rated current.