Ha Ok I am going to get other resistors and continue experimenting with this value 4.7 Kohm, on the other hand I did another experiment without using the capacitor gave me the following result:
COIL NO. 1.- 11.38 VOLTS OUT
COIL NO. 2.- 11.75 VOLTS OUT
INPUT VOLTAGE 12 VOLTS X 0.20 = 2.4 W

My e-mail is: skycollection@hotmail.com
In case you have some time and you want to send me some advices

Thanks a lot

Dear Jorge, 

Please give some more details to better understand your latest experiment. 

So you removed the puffer capacitors from the output of the R full wave rectifiers if I understand you correctly. Is this correct?

And what about the LED bulbs? Did they remain at a decent brightness like they were with the capacitors in place as you showed in the video earlier? 

Thanks for your email address but I would prefer communicating openly as we have been doing now.  If you have special question, then please send a personal message via this forum, ok? 

Thanks
Gyula

For my last experiment I removed the diode rectifier and the capacitor as well, and i connected a new diode rectifier without capacitor, both output voltages give almost 12 volts each coil, related to the BEMF and the led bulbs kept their brightness the same as before , it only decreased the engine consumption a little.
My second experiment was to connect a 7 amp acid battery and it began to charge very well, the BEMF bulb went out and the motor consumption remained the same, without changes, 2.7 Watts.
I think that the capacitor is not necessary to charge a battery, my motor continued to rotate even with the charging battery connected.

Hi Jorge, 

Okay, thanks for the further explanations on your latest tests. 

The diode bridge you labeled as R in your video schematic surely rectified both the voltage induced by the rotating magnets and the BEMF (coming via magnetic coupling from the switched-out winding) and the puffer capacitor stored both of these as a filtered and averaged DC voltage. 

When you remove the capacitor, then an unfiltered DC appears across the output and the voltmeter cannot show the voltage values correctly, this is what I think. 
Okay that the LED bulbs kept their brightness the same as before during this test. 

I assume that both the earlier and the new diode rectifiers have had at least 250 V or higher reverse breakdown voltage ratings. For such bridge rectifiers the best is to assemble the bridge from individual fast switching diodes like the 1000 V, 1 A rated UF4007 or similar types. Off-the-shelf diode bridges with the 4 built-in diodes could be used but these are not fast switching types and good for the mains frequencies but not so good for pulse rectifiers. 

Regarding your second experiment charging a 7 Ah battery and you found the BEMF bulb went out while the battery was being charged: This is normal and expected because the 12 V battery virtually appears across the output winding (and via the magnetic coupling it also appears across the switched winding) and will keep all the induced and the BEMF voltage levels at its 12-13 V amplitude while being charged. Batteries are very good voltage limiters. 
So this max 13 volts level is not enough any more to light up the LED bulbs which may be able to emit light when the peak voltage across them is at least as high as 75-80 V, maybe little higher (manufacturer's specification clarifies this lower limit).
You nicely experienced the trade-off in voltage hence power levels between the diode outputs and the BEMF outputs by observing the effect of the 1 kOhm and the 2.2 kOhm resistors on the brightness of the LED bulbs: 

- with the outputs unloaded the output voltage across the capacitors was 66 V, LED bulbs were bright and rotor speed was at maximum, 
- with  1 kOhm load the output voltage was 4.5 V only and the rotor slowed down, LED bulbs went off, 
- with 2.2 kOhm load the output voltage was around 40 V, the rotor speed was not affected and LED bulbs went also off as you wrote- with the charging battery the output voltage was surely around the battery voltage, the rotor speed was unaffected and the LED bulbs were off.

This shows that when the averaged voltage level was the unloaded 66 V DC across the outputs, the peak voltage spikes across the switched coil was surely around the 100-110 V voltage for the LED bulbs, these spikes gave the brightness for them in a pulsed operational mode. And the moment these spikes amplitude was reduced i.e. attenuated by the resistor loads (or by the charging battery) below the lower threshold voltage for the 110 V LED bulbs, the bulbs went off and the induced and the BEMF power went into the resistors (or into the charging battery).

Hopefully these additional comments help understand better the operation of your interesting circuit setup.

Gyula

Gyula, very interesting, i really appreciate your time in this project, i am experimenting with the circuit, now i have some fast diodes, i am going to build a new rectifier and i get some capacitors 2200 Uf 16 volts, i will make new experiments with this components.
In my new circuits I am incorporating a green terminal where the 1K resistor is connected, to make the conversion to 24 volts I am changing that resistor for a 2.2k , so easy with a screwdriver and I can now apply 24 volts.
Regarding the BEMF I am connecting a 120 volt 6 watt led bulb in order to absorb the voltage peaks, which I have been very successful, the neon bulbs (a pair) are kept off and in case the bulb of led fails, the two neon bulbs will come into operation while I replace the damaged led bulb, so far I have not lost any transistors.