Shorting coil gives back more power

[gyulasun]

Hi Laurent,

From the RSonline link above here is two 200V FET types I would choose for tinkering:
http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=6887153   

and
http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=6886973

Of course there are many other types, these are maybe the cheaper ones, with still good parameters.

Yes I agree you try to establish a working plan / order. If you wish to replicate Romero setup, you have to use at least 24 white LEDs in series (I assume his 72 LEDs are connected as 24 in series, he has 3 such series line and the 3 lines are paralleled.  This means that the lowest forward voltage needed for lighting the 24 LEDs in series is 24 x 3.2V=76.8V or so. (White LEDs has a forward voltage range between 2.9 - 3.4V to conduct current.)
By the way,  this means that under this voltage level, there is NO current consumption from the generator coil. So the best loading test would be what Stefan suggested first with a low power incandescent light bulb  (it has no forward voltage threshold like any LED has). Of course such a light bulb could be substituted by a power resistor of appropiate value.

Unfortunately, the fan motor coil you also use has a big DC resistance coil and when you try to load it, the current drops voltage from the output and it heats the wire, you cannot use it for LED or bulb driving. You measure 3V after the full wave bridge, while your unloaded AC output is about 80V peak to peak which is about 28V effective value...  The only solution is to make another gen coil with thicker wire, to bring down the several hundred inner DC resistance to under 10 Ohm or less. 

Gyula

Hi Gyula and all

Thanks for advice and i am looking for the best fet.

But more important at this point is to make some order in the research.

Because the general setup is really interesting and after 3 days of full scrumble upon i begin to see a way to go on.

So my aim is to replicate exactly what ROMERO UK did in his video, where he can light 70 LED very brightrly with only one REED and one 47 uf 250 volt cap  and this WITHOUT LOOSING RPM.

So  far from today, is that i replicated this and finally i could get my (all that i have at disposal for the moment)  13 led IN SERIE to very good lighting.
But the price for this is a slight increase in current draw on the motor.

So the aim of the AC cap (as per Konead) is to suppreseed the lugging of Mister Lenz. So i think that, the MATCHING  of the coil + core + cap + magnet rotorspeed  is at atmost importance.

And for today  when the rotor is spinning (without the shorting the coil ) it gives 3 volts after the FWBR  (and the 13 leds in serie does of course even get a sign of light) and  as soon as i connect the shorting circuit with the 2 mosfet , the same led ramp light very happily and the the voltage is about 36 volts.

Very interesting

i will make some thinking on this subject to prepare a video

good luck at all

laurent

[gyulasun]

Hi Laurent and All,

I think you have to see this new schematic drawn by Doug Konzen today because it shows the series non-polar capacitor in series with the AC input of the diode bridge and NOT in series with the switch that does the shorting.

This should make a difference in the Lenz drag as per bolt's explanation in the previous page, namely the series capacitor forms a high pass filter. This should mean that the induced sine wave is mostly supposed to be blocked by this capacitor (i.e. its capacitive reactance is chosen as a relatively high impedance at the sine wave frequency), however the spikes created by the shortings mainly contain harmonics for which this capacitor should show low impedance to be able to pass them towards the full wave bridge.
If I recall correctly, Romero used his 0.47uF capacitor in series with the reed switch.  And schematics so far from Doug Konzen showed the capacitor in series with the switch too. 
I think the capacitor in series with the input of the AC bridge makes more sense if the role of the capacitor is indeed working as a high pass filter.

This needs to tested of course.

rgds,  Gyula

PS The schematic also shows a MOSFET driver IC, capable for very fast switching. You find them under different brand names like TC4422 or MIC4422 etc.  According to Doug Konzen, this driver circuit works nicely, controlled by a normal Hall sensor. The RF filter shown in series with the gates of the FETs are small ferrite beads with a few turns (kind of choke coils), you can replace them with series 47 or 56 Ohm 1/8 or 1/4 Watt resistors. The goal is to prevent spurious oscillations in the gate circuit should it happen.

[woopy]

Hi Gyula

Thanks a lot for this info about Doug shematic

will study and try it ASAP

for today, i could replicate Romero uk result  and i tried LED in serie and in parallel i will report the result because it is not so simple and i want to be sure before saying anything

For info i tried a bigger coil with much less impedance and also inductance with good results also ,i will go on the testing ( see the pix with very strong and probably too large pulse at the top and bottom of the sine wave)

good luck at all

Laurent

[e2matrix]

Woopy, you have good nose

With reasonable confidence I can say that to access this "effect" one needs very sharp gradients and/or exiting coil at its resonance frequency,

Easiest way to hit resonant frequency is to use spark gap, reed in this application works as spark gap. Thats why people have seen good results with reed and virtually nothing with transistors etc. Ordinary solid state components when driven in ordinary way are slow and transparent for HF and just do not work in this application.

For those who are trying to make solid state version, good choice would be russian "kacher" type circuit. There solid state components are "shockwave" driven and give desired effects. (NB! Kacher can and will affect your health, strong RF emission! NB!)

I dont see how one can easily combine very sharp gradients at exactly right frequency with rotary version and transistors etc. So I would bet on heavy duty reed or some other component exibiting wide spectrum sharp gradient "white noise" emission (and physical breaking of circuit).

Also series AC cap is a must. When there is AC cap in series, Ohm and Kirchhoff do not apply. And thats already half-done deal.

Also forget about directly measuring amps out of this contraption! This power is only good for driving not resistive loads (or charging acid batteries)! To use it for resistive load you must first transform it.

And as always, trust no one, follow your instincts, progress only after replicating basic effects

Totally agree with everything you said as part of why this works with a reed switch but will be difficult using solid state.  I'm thinking now about a possible cam setup with multiple automotive type points and larger coils. 

[Arthurs]

    I always think that the core of the coil is very important, it seems very suitable for the coil core Fe3o4

YouTube Video Links:
http://www.youtube.com/watch?v=xA8HalLJ92c&feature=feedwll&list=WL
http://www.youtube.com/watch?v=vZE-pIXipm4&feature=mfu_in_order&list=UL