Lasersaber strikes again. A joule thief king ?

[TinselKoala]

In the first part of the second video I posted I am _not using_ the third coil winding at all. The device is wired just exactly as lasersaber's schematic, but with slightly different turns ratio. The LEDs are definitely on. When I connect a single end of the third winding I get brightening of the LEDs. Later I discovered the real effect of the third winding and its proper connections, and that is what is shown in the first video I posted, showing brilliant LEDs running with less than 0.2 V on the capacitor.

The video with the scopeshots (third coil winding connected) shows that there are several superimposed oscillation frequencies happening all at once. I have never seen that slow and stable variation in primary frequency in any JT before.

[scifi123]

Hi. Try this  LT simulation. It works...

  Skycube,

  I see that you added a parasitic capacitance between the primary and the secondary (C1 in your circuit).
  I thought about doing that, but you were faster. 
  Good work on making it oscillate.

  Anyway, I changed some values in your schematics and managed to get the circuit to oscillate in bursts, which I believe is what the real circuit does (see the attached picture).
  I'm also attaching the LTSpice circuit.

[scifi123]

  Lasersaber, TinselKoala and others who replicated the experiment: did you measure the input/output energy and did you find overunity ?

  If you didn't measure it but only guesstimated it, do you have any reason to suspect overunity ?

[TinselKoala]

No, I have no reason to believe that it is OU. My own interest is merely in very low voltage operation with _useful_ levels of light output. In the sense that these circuits allow the extraction of residual energy in batteries that would normally be discarded as unable to run anything... well, that's "free energy" isn't it? Or rather, it's energy that you paid for when you bought the battery but can't normally use.

By the way, your latest series of scopeshots looks pretty good, very similar to what I showed in the last video above but with higher Q. It's too bad that a DSO or simulated DSO can't also show that much slower variation in burst frequency that shows up on the analog scope as "shrinking and expanding" along the horizontal axis. It would be interesting to see if your sim behaves that way, but I don't know how you'd be able to catch it on the screen.

The idea of coupling the coils with a small capacitor is interesting. I may try that in my systems a bit later on.

[scifi123]

By the way, your latest series of scopeshots looks pretty good, very similar to what I showed in the last video above but with higher Q. It's too bad that a DSO or simulated DSO can't also show that much slower variation in burst frequency that shows up on the analog scope as "shrinking and expanding" along the horizontal axis. It would be interesting to see if your sim behaves that way, but I don't know how you'd be able to catch it on the screen.

The idea of coupling the coils with a small capacitor is interesting. I may try that in my systems a bit later on.

  You can catch it with the simulator if the simulation duration is long enough. Let's say that the "shrinking and expanding" phenomenon has a period of 1s. Then you simulate your circuit for 1s and you should see the duration between two big consecutive spikes varying.
  But I don't believe that this LTSpice circuit reproduces that phenomenon since there is nothing in it to generate an oscillation with such a long period (around 1s).

  If you're interested in simulating that phenomenon, here are some ideas that might explain it (in decreasing order of probability):
1. The transistor's temperature (or maybe the LEDs') varies periodically (with a period of about 1s) which changes the duration between the spikes. So we have here a "thermal oscillator" with a low enough frequency (about 1Hz).
2. The circuit generates two oscillations and their frequencies are close: for example 100Hz and 101Hz. The oscillations create interference between them (or "beatings") with a frequency of 1Hz that you see on the scope.
3. Maybe there is no oscillation with a period of ~1s but there is something wrong with the measurement set-up.

  In my opinion the first explanation is quite plausible, the second one is unlikely and the third one even is more unlikely if the oscilloscope operates correctly.

  Anyway, let me know if you want to do more experiments to find out the cause of the "shrinking and expanding" you see on the scope, maybe I can help.