170 watts in - 1600 watts out - looped - Very impressive build and video

[listener191]

The Magneto Motive Flux (MMF) distribution on a single pole (concentrated) winding is rectangular.

After further analysis...

Pierre's winding and coil connection scheme is essentially a 6 coil distributed winding that creates a stepped MMF i.e. due to the coil overlap, the center of the pole has a higher MMF than the edges.

Referring to the attached image.

This illustrates how distributed windings normally work in an AC machine. This example showing 5 coils.

Note: Normally due to stator space, this would be reduced to 3 coils in a practical implementation.

Ok, so Pierre uses 6 coils however, the problem is that his 6 coils only have a 6 slot pitch. This results in a waveform that is triangular in distribution rather than sine.

A gap is needed  to widen the top step, although at the moment I cannot see a solution to a 6 pole configuration using 6 coils.

Two coils on and two coils off per pole may work, although the distribution would not be ideal.

If the poles were reduced to 4, then 3 coils on, 3 coils off could be used, which would yield a much better sine approximation. Then just clock the switching sequence at a higher rate to achieve  60Hz (50Hz).

L192
 

[listener191]

For those that want to implement the half bridge ladder scheme that Pierre uses, It would seem the the L298N is still the lowest cost solution.

Tie together OUT 1 and OUT 2 and tie together In1 and In2.

Repeat for other side of L298N.

This should be Ok as the manufacturer already permits paralleling the bridges.

This now gives you two independent half bridges, with each Enable input turning off both transistors.

This will then require 36 boards instead of 18 but it is the cheapest solution I can find, unless you want to build your own half bridge boards, which will likely be more expensive.

The combined output stages will provide about 3.5A of continuous current capability.

L192

[x_name41]

as shown by the demonstration, capturing energy through diodes to supercapacitors is not enough to obtain self-powered so it is necessary moreover return of a certain quantity of energy from the exit to the entrance, there are 2 energy returns in the system (if it can so be called)

here's something like that only that with the capture of energy from windings that are located on the original ones (something like a transformer is getting)

[e2matrix]

Attached BTN7960 half bridge.

Would work well for Pierres's switching scheme but a little pricey.

L192

i'm seeing some on Aliexpress.com for around $3.50 to $5.00 a piece if you buy 10 which includes shipping.  Smaller quantities are only a little more.   Some parts from there are good but some are not the best quality.
I also see this ebay seller has 6 of them for $19.99 with free shipping so only $3.33 a piece for those who may want to try that method.


I wrote the above before seeing Listener191 changed to say the L298N can be used and it is still quite a bit less money as he stated in post #373.   

[listener191]

i'm seeing some on Aliexpress.com for around $3.50 to $5.00 a piece if you buy 10 which includes shipping.  Smaller quantities are only a little more.   Some parts from there are good but some are not the best quality.
I also see this ebay seller has 6 of them for $19.99 with free shipping so only $3.33 a piece for those who may want to try that method.


I wrote the above before seeing Listener191 changed to say the L298N can be used and it is still quite a bit less money as he stated in post #373.

Hi e2matrix,

For anyone that has already purchased the L298N boards it would be better to utilize them and just add more boards however, the BTN7960 offers higher current capability.

Regards
L192