Kapanadze Cousin - DALLY FREE ENERGY

[verpies]

If you still need to use transistor, attach diode in series (just make them 10+ in parallel to lower resistance) between coil and transistor. The mechanical break will still produce the best spikes but sometimes the electronic counterpart is OK too.

For 1A of current, the effective resistance across 1 diode is  0.6V / 1A = 0.6Ω, because the voltage drop across the diode is 0.6V (almost independent of current).
10 diodes in parallel will still have a 0.6V voltage drop collectively, so their effective resistance will not change.

[gyulasun]

Hi verpies,

Really not wishing to nit-picking on paralleling some diodes but it brings up memories for me.
At my workplace I used to have access to a transistor curve tracer (the good old 575 type from Tektronix) and once I tested some Si diodes in parallel. The display clearly showed a certain (small) drop in forward voltage, the more I paralleled (I tested max 5 or 6 pieces) the more reducement in forward voltage happened. The relationship was non-linear in the drop with increasing the number of diodes and the amount of the drop was somewhere between 70 to 100mV in case of 5 diodes (at a given current) if I recall correctly. In case of 10 diodes the drop may amount to say 120-130mV hence in your example at still 1A forward current the resistance may change to 0.47-0.48 Ohm or so because the resultant forward voltage may become 0,47-0.48V. The drop in forward voltage due to paralleling the diodes also depends on the amount of the forward current of course, the higher the current, the more drop happens i.e. the more the resulting forward voltage decreases.  At 1A current this reducement in power loss is not so significant but for higher currents it may be so,

For those wishing to see or recall some memories on the 575 curve tracer, here is a picture on it: http://en.wikipedia.org/wiki/File:Transistor_curve_tracer.jpg

Greetings, Gyula

[verpies]

Really not wishing to nit-picking on paralleling some diodes but it brings up memories for me.

Nit-picking is fine with me
All diodes have some small leakage current in reverse and forward, below their major forward voltage threshold (e.g. 0.6V).
This leakage current can be simulated by a large resistor in parallel with the diode. At small currents (e.g. used by curve tracer) this leakage current can be a significant fraction of the intentional current.  At larger currents (e.g. 1A) it becomes insignificant.

[gyulasun]

Yes I agree, the resultant reverse current increases when paralleling diodes, that is one 'price' to pay for the lower forward voltage.
For Schottky type power diodes the reverse current can be much higher than for Si diodes, so one has to consider that in their case too.

[itsu]

Still playing with this coil shortening thing.

I build the "syncer" as mentioned by sergey in this post:

http://www.overunity.com/12736/kapanadze-cousin-dally-free-energy/1680/#.UcAsi5ya8kI

I had to modify it somewhat as i did not like the circuit being connected to 220V mains.
Also my 2 N-channel MOSFETs need both positive pulses on their gates compared to their sources, so needed to convert the negative pulse to a positive one.

Result is a rock steady "sync" and a short on top of the both sine wave peaks.
However, the small (but high) spikes during turnon/off time do not at all compensate for the big chunk missing in the sine wave, and is causing the small bulb to dim.

Problem still is the lack of "width" control and the fact that the shortening is not complete (to ground), so i guess i need to implement MOSFET drivers to more quickly and fully open/close the MOSFETs.

Video here: http://www.youtube.com/watch?v=-T5wolRnamA&feature=youtu.be

Regards Itsu