Self running coil?

[Magluvin]

Hi Mags,

no I don't have one. I can add it. How high of value can I go?

Luc

I ran a 10k for my orbon. If your setup runs well with it, then maybe you can discount the 555 powering your cap.
I dont think it is, the fact that you remove the magnet and good effect stops, tells us that power is not coming from the 555. But to satisfy the needy, the 10k res, if the effect is the same, then your on your way to stardom.
Mags

[e2matrix]

Luc,  the point at which the caps started going negative in voltage after the coil was disconnected and the caps were quickly drained with the 10 Ohm resistor and before the 1.2 volt battery was put back on -- I take that as a good sign - as the coil was hooked back up at that point and since the only power in the circuit could be coming from the timer circuit and it was actually reversing the charge - it seem that may be important to note since if they were charging in a positive way then you might have to consider the voltage was leaking through from the timer circuit.  I think (may be way off here) that would indicate the timer circuit is NOT putting voltage into the caps.  And that's great news!

[gyulasun]

...
So even if I had an OPTO isolator this would happen also?
...

Well, what does your circuit do you assembled?  I think the point is you created an LC tank circuit, the L is a magnet-tuneable toroidal cored coil and the C is a voltage-dependent capacitor made from the drain-source output capacitance of a MOSFET.

The energy input to tank circuit can only come from the 'kicks' as the gate-source input pulse 'leaks' through the gate-drain capacitance (also voltage dependent) to the drain-source side and pumps energy into the tank just like you push a child on a swing periodically to maintain the swinging speed.

So if you try to use a device for kicking the tank and the device has very small unwanted coupling capacitance for instance, then you reduce the energy input that otherwise would be there like in case of a MOSFET. However, using a different device for the switching, it would need the same 'cut and try' i.e. much fine-tuning to find the best match between the tank and the input device. 
Or maybe originally you did not wish to create a tank circuit you finally got as the result, but you wished to make something useful with a toroidal core +coil on it + permanent magnet combination. Then the use of the opto isolator is also a question if it is needed for you.  Sorry, do not misunderstand me, I try to answer your question. Magluvin's suggestion of a series 10 kOhm (or maybe higher) can be an answer where the energy comes from, this test is equivalent with using a MOSFET that has only ,say, a much less (some pF, not 500-1000pF) gate-drain capacitance than your FETs have now. 

Gyula

[mscoffman]

@gotoluc,

Excellent demo. What I would do, though is somewhat rhetorical
is record the frequency and duty cycle that seems to be optimal
and build a fixed frequency source that uses very little power
and operates from the main battery/capacitor with just the
required frequency and duty cycle. Some PUT Programable
Unijunction Transistor circuits seem to be inherently high
impedance circuit designs. Note that the mosfet doesn't require
very much gate current because of it's extraordinary gain but
the voltage needs to be at least .7Vdc above mosfet source to
drain voltage for it to turn fully on.

Below the purpose built frequency source I would build a voltage
doubler - voltage multiplier power supply ...producing just enough
current to keep the frequency source running. With design skill you
should be able to use small enough power to keep the toroid coil
running and gaining from the capacitor. You may be able to
have the PUT oscillator circuit bootstrap it's own power.

I think this is doable...but then you will need to figure out where
the extra energy is coming from; internally or externally of the
toroid. You will have achieved an overunity design in any case.
Which is a step in the right direction.

:S:MarkCoffman

[Magluvin]

Good suggestions MS

Luc   have you tried to run it at 12v instead of 1.2?  If it takes no current it should run safe. Put a 12v stop light bulb and use it inline with the bat as a fuse that never blows and shows problems. The stoplight bulb will deliver good amperage for what you are doing while the filament is cold.  We use 100w bulbs inline to test amplifiers that have problems.  Lil trick

But if you can get the coil to run even 5v or greater to keep the 555 going all together, it would be something to see.

Mags