Sharing ideas on how to make a more efficent motor using Flyback (MODERATED)

[verpies]

decreasing R2 from 10 to 3.3 Ohm made the C2 cap pratically go to zero, see below screenshot where i have ch1 (yellow) set at max visible amplitude.

That poor Q3 must be carrying a Helluva current pulse now and I hope that it does not break, but the waveforms look really nice.
Next year,  I shall calculate the "flyback" energy recovery efficiencies and losses, in the circuit's resistances and diode voltage drops.

P.S.
I'd like to notice that it is the designer's fault (my) that Q3 closes only for such a short time (proportionally to L2). 
If I had added more diodes to Q3's gate circuit, then Q3 would stay closed until the Pulse Sequencer had explicitly turned it off (not the decaying current in L2).  I did not want to do that in order not to complicate the circuit too much, but maybe I'll do it next year.


Happy New Year!

[itsu]

Thanks,

I measured this Helluva current pulse in Q3 (IRF840) with my new current probe, see screenshot below.

The yellow trace is the voltage across the cap as before, purple the Q1 Drain signal as before and green
the current through Q3 with the current controller set to 5A/div., so we see about 15A pulse current.
The IRF840 should be able to withstand 32A pulse current, so should be fine.


Looking forward to next year


Happy New Year!

[sm0ky2]

What you're describing sounds like a shaded pole motor, is this correct?

Thanks

Luc

something that looks similar to this picture at the bottom
they are in almost every household appliance that has a fan in it.

not this one in particular, they come in different shapes and sizes, some are double-motors attached tandem.
I think this one is a Braun, which kind of sucks because of the windings.
Westinghouse makes some good ones that can be scoped out to find a resonant node.
speed increase with simultaneous current drop is the prime indicator.
the other is a "cleaner" waveform at the frequency node.

if you have a good scope you can expand the scope image you can see variances in the signal, tiny spikes, that reduce in a resonance mode

when you are dealing with a pre-manufactured device, you don't have the options of engineering everything to a specific set of resonant frequencies.
the best you can do, with a complex system of multiple inductors and coils, is to scope the signals across a broad frequency spectrum.
and try to find a resonant node.
pay attention to voltage, current, and signal coherency.
resonant nodes will be noticeable when you are looking for them
the inductor or housing of the device may physically vibrate at that freq.

if your system has multiple frequencies, you can do a little math, and/or scope out both frequencies while you change freqs and record results.

[gotoluc]

Hi sm0ky2,

please make a video of your shaded pole motor demonstrating the advantages.

Thanks

Luc

[itsu]

See this sim of two 555 working without inverting drivers.

I did not complete the full drain circuits of Q1 and Q3 - I just completed their gate circuits.
The adjustment of the astable pulse width is done by varying the 22K resistor and the monostable pulse width is done by adjusting the 21K resistor.
( Yes, I know that changing that 22K resistor will change not only the pulse width/duty cycle, but it will also change the frequency of the entire astable waveform - that's why I hate working with these 555s and why people come up with such monstrosities as in Fig. 4.4.8 )
BTW: The "out" labels what is shown on the sim scope.

I managed to adjust my duty cycle pot of the first 555 so i got a similar out signal as your sim (40Hz @ 80% positive duty cycle).
This signal is fet to U2 (non-inverting)/Q3.

It is also fet to the input of the 2e 555 timer (pulse extender) and its output adjusted by its 100K pot to give a 18% positive duty cycle signal
which is used to drive (no driver in between) Q1.

So the feedback hall sensor is not in use anymore.

Below screenshot 1 are the signals again using this setup (C2 is 2.2uF cap).

Blue voltage across the csr point A to B
yellow the voltage across the C2 cap point D to B
purple the voltage across the drain of Q1 point C to B
green the signal at point F to B

A DDM across the C2 cap reads 78V DC

Using a 10uF / 400V cap makes the yellow signal fail to go to zero again, so probably need more L2 inductance and/or less R2 resistance.

Screenshot 2 is the current spike (green trace) through Q1 measured with my current probe at R2 (3.3 Ohm), the current controller still set to 5A/Div.
Yellow is the C2 voltage when shorted by Q1


Itsu