Pierre's 170W in 1600W out Looped Very impressive Build continued & moderated

[listener192]

Hi All
Here is email I just got from Gary Porter, and a couple drawings too:

"I figured out what he did and it matches with his pulsed sawtooth and BEMF spikes. The attached pic shows one circuit of the 36. I need 72 FET 36 for GND and 36 for power. Your right the odd coils are versed polarity to get South poles. So their are 36 coils I found that with my stator slots I can have a max of 34 17ga wires so two coils at 17turns each will just fit the slots. I'm about half way thru inserting the 36 coils man it is a pain I have to put one wire in the slot at a time and near the end I have to push them down into the slot with wood.

I'll have to redsign the PCB into two boards one as a controller and 9 others each having 8 FETs. The Arduino 2560 has 54 digital outputs so I'll have to use 36 of them twice to get 72. so each pin connects to two buffers each have an enable line.. Pierre did this on his control board look at his top left and you'll see two FET by themselves. These provide a ground path each of his 36 relay groups. He uses 72 diodes to gather BEMF positive spikes only that's why his waveform has so much negative noise after each relay is opened.

I want the sawtooth with noise to pass into a 1:1 transformer cap on primary for 60 hz resonance. this should clean up the signal."

If you alternate polarities between coil groups, you loose the continuity of flux, as then you cant overlap fields?
I also tried this with relays, without the overlap, and the output was not very much.
L192

[r2fpl]

The passage of the magnetic flux is only when the coil is within the rotor range from one or both sides. When the coils are active on the sides there is no magnetic flux.
When the coil is shorted, all the coils respond and add to the primary coil.  This is my explanation. I do not know if correct.

[jerdee]

Hello all,

Been thinking more about how to change the inductance and current in rotation without loosing field rotation. We have many possibilities of changing inductance with 6 fields.   As long as you keep the 6 field orientation  (NSNSNS), you can combine any of these 6 coils in series or parallel to increase and decrease inductance.   For example, 2 could be in series, while the remainder 4 could be parallel.   You never loose the 6 field NSNSNS rotation, you only add the change in inductance and current as you drive each grouping that has variations in them.  Otherwise, everything else stays the same.

For example, to keep this easy, let's say the inductance for a single coils is only 10mH per coil.  6 in series would = 60mH while 6 in parallel would = 1.67mH

For a 36 slot stator, you have 6 groups of 6 fields.  So you can cycle the inductance with 6 increments/decrements in inductance in rotation, I'm thinking forwards and backwards here in the cycle as you want NS...then SN on the 6 groupings.

1.  Group 1 starts with 6 fields in series. =  60mH
2.  Next Group 2 (one increment in NS or SN slots)  with 4 in series, while 2 are left in parallel = 45mH
3.  Group 3 with 2 in series, while 4 are left in parallel = 22.5mH
4.  Group 4 with 0 in series, while 6 in parallel = 1.67mH
5.  Group 5 back to 2 in series, while 4 are left in parallel = 22.5mH
6.  Group 6 back to 4 in series, while 2 in parallel = 45mH
Then starts over opposite polarity with same cycle.

For this example, you are generating a variation in inductance and current as it cycles back and forth between:

60mH
45mH
22.5mH
1.67mH
22.5mH
45mH

this happens over and over in inductance shifting as it rotates WITHOUT CHANGING NSNSNS ORIENTATIONS in rotation.  Wouldn't this create d(LI)/dt?  Also wouldn't this simulate the magnet getting closer and further away?   

You still keep the code the same, you can do overlap mode as well, and you have full recovery from off states.   For example, you can recover from any group when off using a single mosfet on the d.c. side of fwbr.

This would only require 6 H-Bridges to work in both polarities, 12 pins from MCU, and 6 more pins for recovery FETs, for a total of 18 pins of MCU.

I may be far off on this thought process, but at least I"m trying to think of ways to increase and decrease variation and intensity as it rotates NS then SN in 6 coils groupings.  We already know that just rotating fields is not enough.

Otherwise, I really like the concept of more variation and intensity from Konehead's work/experience on the coil shorting.  I can see how Pierre gravitated to his scope images.

Just thoughts right now, but this is how I'm thinking at the moment. 

Jerdee

[ARTMOSART]

Tous,
Je vous prie d'examiner ce petit détail sur cette capture d'écran à 2'15 sur la vidéo 1 ?
personnellement je voix comme un connecteur ,qu'en pensez-vous?
cordialement Mosha.En:

All, Please review this little detail on this screenshot at 2'15 on video 1? personally I see a connector, what do you think?

[r2fpl]

Hello all,

Been thinking more about how to change the inductance and current in rotation without loosing field rotation. We have many possibilities of changing inductance with 6 fields.   As long as you keep the 6 field orientation  (NSNSNS), you can combine any of these 6 coils in series or parallel to increase and decrease inductance.   For example, 2 could be in series, while the remainder 4 could be parallel.   You never loose the 6 field NSNSNS rotation, you only add the change in inductance and current as you drive each grouping that has variations in them.  Otherwise, everything else stays the same.

For example, to keep this easy, let's say the inductance for a single coils is only 10mH per coil.  6 in series would = 60mH while 6 in parallel would = 1.67mH

For a 36 slot stator, you have 6 groups of 6 fields.  So you can cycle the inductance with 6 increments/decrements in inductance in rotation, I'm thinking forwards and backwards here in the cycle as you want NS...then SN on the 6 groupings.

1.  Group 1 starts with 6 fields in series. =  60mH
2.  Next Group 2 (one increment in NS or SN slots)  with 4 in series, while 2 are left in parallel = 45mH
3.  Group 3 with 2 in series, while 4 are left in parallel = 22.5mH
4.  Group 4 with 0 in series, while 6 in parallel = 1.67mH
5.  Group 5 back to 2 in series, while 4 are left in parallel = 22.5mH
6.  Group 6 back to 4 in series, while 2 in parallel = 45mH
Then starts over opposite polarity with same cycle.

For this example, you are generating a variation in inductance and current as it cycles back and forth between:

60mH
45mH
22.5mH
1.67mH
22.5mH
45mH

this happens over and over in inductance shifting as it rotates WITHOUT CHANGING NSNSNS ORIENTATIONS in rotation.  Wouldn't this create d(LI)/dt?  Also wouldn't this simulate the magnet getting closer and further away?   

You still keep the code the same, you can do overlap mode as well, and you have full recovery from off states.   For example, you can recover from any group when off using a single mosfet on the d.c. side of fwbr.

This would only require 6 H-Bridges to work in both polarities, 12 pins from MCU, and 6 more pins for recovery FETs, for a total of 18 pins of MCU.

I may be far off on this thought process, but at least I"m trying to think of ways to increase and decrease variation and intensity as it rotates NS then SN in 6 coils groupings.  We already know that just rotating fields is not enough.

Otherwise, I really like the concept of more variation and intensity from Konehead's work/experience on the coil shorting.  I can see how Pierre gravitated to his scope images.

Just thoughts right now, but this is how I'm thinking at the moment. 

Jerdee

Jedree,

You've come up with it. It would be like changing the intensity of the field and the movement of the magnets.
If you do it with ordinary magnets and subtract friction, it will still be too little. Maybe it would be 99% efficiency. I think Pierre does it all but he also coils up on the coil as the konehead says.