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

[gotoluc]

Oh no.  This encoder generates 360pulses per revolution, so you can use it to trigger at any angle and as many times per revolution as you want.

I didn't know that! so that same model can do 3 phases?  how do you program it to do what you want?

Do you realize, that it is also possible to have a fixed optical reflective switch with only 1 pulse per revolution anywhere, and electronically divide the time between these single pulses into 360 equal periods using the 4046 chip and some counters, in order to trigger at any shaft angle degree and as many times per revolution as you want ...without a shaft encoder ?

No, I did not know that!... I'm relying on people like you who know about electronics to inform me of these things when you see me struggling with making timing reflector wheels on the devices I built.
It's a good thing you're telling me now because I'm presently modifying an off the shelf 12 pole 3 phase Switch reluctance motor and spent the day building a 3 optical switch (see pick) to control each phase.

How many of these 4046 chips and any other components do I need to make a 3 phase work?  so I can look into it to see if it's worth going that direction at this point.

Thanks for your time.

Luc

[verpies]

I didn't know that! so that same model can do 3 phases? 

Yes.

how do you program it to do what you want?

The rotary shaft encoder is commonly used by counting the pulses, that it outputs.  BTW: There are also rotary encoders that output 3600p/rev.  ...and more.
The counting is done digitally by several cheap CMOS chips and BCD switches (without programming) ...or alternatively by one microcontroller (with programming).

For example if you set it to output the 1^st, 121^st, 241^st pulse then it produces a 3 phase output 120º apart and if you want to advance this timing by 5º then you can set it to output the 6^th, 126^th, 246^th pulse.

How many of these 4046 chips and any other components do I need to make a 3 phase work?  so I can look into it to see if it's worth going that direction at this point.

Just one 4046 chip and four counter chips like 4059 (or 40102), which are cheap, but the BCD switches to program them can be expensive.
The alternative, one chip solution with a microcontroller costs ~$50 but it requires a connection with a computer to program and work.

Be aware that dividing 1 revolution into e.g. 360 smaller time periods with the 4046 chip, gives you an accurate shaft angle down to 1º only when the rotor's speed is constant.  If the rotor is accelerating or decelerating, then this method of computing the shaft angle makes an error that increases with the acceleration.
The rotary encoder method does not make this error, regardless of acceleration.

The digital timing diagram below illustrates how that appears on a scope with a x18 pulse multiplication.
If a rotary shaft encoder is used instead of the 4046 pulse multiplication, then the purple waveform comes directly from the encoder.


P.S.
Your 3 optoreflectors and the pie-wheel, constitute a home-made optical rotary encoder of low pulse count:  12pulses / rev.
Just another way to look at it

[citfta]

Hi luc,

I am not sure if this will be any help or not.  How much control do you want to have over the pulses going to the motor?  If you are willing to let a circuit make those adjustments automatically there are some 3 phase motor controllers already built that are pretty cheap.  They are used in Radio Control vehicles.  I fly RC airplanes and the little motors on them are very powerful and controlled by what are called ESCs or electronic speed controls.  They come in all sizes from ones for 15 amp maximum output up to 150 amp output and probably even higher.  The only problem with them is they are usually for 12 volts except for the larger and more expensive ones that are for higher voltages like 30 volts or so.  You also need some way to control the esc if you don't have a RC transmitter and receiver.  But I think a simple circuit could be built to control the esc.  If you think you could use something like this let me know and I'll do some more research to see what would be needed to control the esc.  The better escs let you make adjustments to the program for fine tuning to different motors.  The controllers automatically adjust the pulses going to the 3 phases to control the motor from 0 rpm up to 20,000 rpm or more depending on the design of the motor.

Carroll

[tishatang]

Hi Luc
citfta is right about using an RC controller for model airplanes.  If you go that route, use a controller for a model car, it has a fan for cooling.  Airplanes have the air stream for cooling.

I went this route to drive my alternator setup.  I converted a car alternator to run as a 3 phase motor.  It works fine.  However you are limited to around 12 volts or so and the electronics are frail. 

If I were to do this rig again I would go for a DC motor controller for an Ebike or scooter.  They are cheap and rugged.  And motors sometimes for free as many scooters  are in the recycle yards when they break, also with controllers in them.
Then I would use this DC motor to drive a car alternator without the regulator and diode bridge in it.  The alternator now becomes a 3 phase generator to drive your test motor with pure 3 phase power.  All this very rugged and cheap. 

If you are interested I could post photos and videos how to convert the car altenator and hookup the RC ESC using a simple $5 device.  No need to have a radio transmitter to run the ESC. 

Hope this helps,
Chris

[tishatang]

I just realized that the in-hub motors in Ebikes are brushless 3 phase motors, the same as RC models.  One of these controllers might just drive your motor straight away without the use of the alternator?

On the other hand, it might limit your experiments.  For example, maybe you could put a diode on each leg to the test motor.  Maybe it will run in attraction mode?  Then, when you collect the back spike, you can easily compare output to input.  You can't do this experiment with the ESC controller because they depend on feedback to sense the position of the rotor.  Here the DC motor/alternator makes more sense.
Just tossing out some ideas.