Hi Kone thanks for clarification

This is much better calculated BEMF output

Maybe try SHORTING the coil right after the swtich-off of the primary in near future...try and "short" the backemf into caps rather than jsut swtihci it out....by shorting I mean to connect the two leads of the coil together very briefly during sinewave peakl-period, and in your case right when motor coil switch OPENS - which is an event that occurs during the sinewave peak period if looking at the sinewave the rotor magnets-only create...
the put FWBR AC legs across the coil ;eads being shorted, with DC side of FWBR into capacitor...use your arduino timing to dump this cap to load but have no load on this cap when it fills..

Kone at bottom is circuit schematic I draw to control drive coil shorting

Please check if this is what you suggested for drive coil shorting?

JoeFR

Hi Joe

It might be redundant to have both the backemf/recoil-collecting diode circuit, and the coil shorting circuti across the same motor coils, (might not be too) as the coil shorting should pull out as well, and switch-short too, the backemf/recoil into cap - only difference is swtich short will collapse the coil (big difference)...

so idea is to actually be shorting the backemf/recoil and then you wouldnt need that first circuit with the single diode made especially fro the backef/recoil...thats working pretty good already - its a shame to ditch it but it will be interesting to see what happens if you combine the two extraciton methods - swtich short and steering diode -

however it might be better to send backemf/recoil into one cap, and then after that "settles down" then swithc-short the same motor/drive coil into a 2nd cap, before the rotor-magnet induced sinewave of it starts to drop.
I dont know which will be better - two caps or one combine the backemf with swithc-short or not...a few experiments should tell story...

Other thing I did in thing I built for someone last year, somethig similar to Romeros rig but much larger with huge magnets,  is to dump the backemf from a motor/drive coil ,first into capacitor, then after that event, then swtich short a generator coil into same capacitor, so the backemf/recoil-collectiing event sort of primes the cpaacitor to some good voltage, sthen when the swtich short occurs, the cap is already filled up fairly well and it has lower resistance, so the swtich-short works better into same cap.
this is with using two coils- a drive/notor coil for the bakcemr/recoil. and a seperate gernator coil for the switche-short....so anyways with the same drive coil like you have, it seems like you could first fill the cap with the backemf/recoil then after that, swtich short the same coil into cap too....probalby no overlap should happen in the timing but that is a big probably, since maybe you can combine the two maybe not....You might want  to SWTICH out the backemf by a swtihc between the coil and that diode....so you can control when it happens - right now it is sort of doing it all the time but when you throw in the swtich-shorting you might want them to not overlap but then again mabye this will be good so you will see...

Anyways you got the bidireectional mosfets right, and coil leads on the drains of the two mosfets right, so its going to work now you just have to get teh timing of it all right and pulse width and cap sizes too...

Hi Kone thanks again

Could you please draw simplified schematic what will you correct in my schematic or make different to get the best result out of it ?

JoeFR

hI JoeFR

I think your schematic looks great as is really...I was jsut rambling on about maybe this maybe that...that FWBR should pull out the backrmf spikes fine - but I dont know if the coil-shorting will snuff out the spikes, or make them fill cap even greater...seems like good idea is to fist take out spikes jsut like you did before, then after that do this coil-short cirucit to it...one after the other keep them seperate....but I might be totally wrong and it will work great as is wtih it all combined..

Mabye you shouldnt step into the coil-shorting right now, mabyeit wil be dsitraciton since you did super great with the backemf/recoil being pulsed out into light bulb...maybe you should refine that some more and make it over 100% via the formula for watts-out....  jsut do some preliminary fiddling with those numbers in the cap-discharge formula and it really has some high hopes  - like for instance if you go with 2000uf cap instead of 680, and everythign else reamins same now you have 3 watts out, or do 4 pulses a second insted of two now you have 2 watts out....pulse width of the SSR willd ecide the votlage-drop alot too...

Ran my 2nd RomeroUK "variant" today - 10V and 190ma input and it went 950rpm...added the magnets behind the cores, now it went 1510rpm at 10V and 150ma so quite a difference with magnets back there.

It is fring 20 times revolution as it has 5 mosfets, one from each coil-position ( five pairsof "facing" paralell coils) and 4 trigger-magnets in the timing disc, since it has 4 rotor magnets (allN) sp this is pure-motor, but when motor coil switch turns off, its going to be pure "switch-shorted at peaks" generator is idea...I made two identical motors so plan is one is going to run other and vice versa for looper that way.

i plan on shorting the coils right after the motor coil swtich closes, like what I suggested to JoeFR but havent got that far yet - here are some pictures: