PhysicsProf Steven E. Jones circuit shows 8x overunity ?

[MarkE]

A 2N3055 takes a good deal of energy to switch between highly conducting and non conducting states.  They were the work horses of open frame linear power supplies where they were great as they are rugged and cheap.  The power gain from a 2N3055 in a circuit operating at low voltages like this is very limited.  A low threshold voltage, low gate capacitance FET solution would yield better power gain, faster switching, and better efficiency.  I would use something like a DMG1012.

http://www.diodes.com/datasheets/ds31783.pdf

[JouleSeeker]

MarkE -- I agree that a 2N3055 is not the best choice for this circuit.

Slider reports on different transistors he has tried with this particular Joule-thief-like circuit:

Transistors tried include C1815, MPSA14 (Darlington), C2236, but the '1P' SMD version of the 2N2222A that arrived in the mail today works best by a long way. 40 seconds was the record with a MPSA14, now it's up to 3 minutes.

He shows videos and discusses anomalies he is seeing; here:
http://www.overunityresearch.com/index.php?topic=2436.msg38412#msg38412

[MarkE]

MarkE -- I agree that a 2N3055 is not the best choice for this circuit.

Slider reports on different transistors he has tried with this particular Joule-thief-like circuit:
He shows videos and discusses anomalies he is seeing; here:
http://www.overunityresearch.com/index.php?topic=2436.msg38412#msg38412

A superficial review suggests that the circuit relies on interwinding capacitance to close the return path from the red winding through the LED and the FWB to the base of the transistor during transistor on times.  Interwinding capacitance would also be necessary to close a path between the red winding through the diode to C1 in order to clamp the windings.  Absent such capacitance, there would be nothing to drive current through the transistor base and nothing to clamp the transformer flyback.

[JouleSeeker]

 LaserSaber is excited in this latest video (this morning, Mother's Day in the US) and for good reason:
https://www.youtube.com/watch?v=6B79UJGoNJE&feature=em-uploademail

Note that it runs long, and operates just fine inside an unplugged microwave oven (as a Faraday cage).

[MarkE]

His circuit's endurance is impressive.  Now comes the task of measuring / calculating the energy delivered to the LEDs.  If he gets very clean voltage measurements across the LEDs and separately characterizes their I-V curves then he can determine energy delivered each cycle.  He'll need to keep track of that as the capacitor discharges.  Given his long run times things shouldn't be moving so fast as to make that very difficult.  The trick will be getting very accurate with the I-V curves over the voltage range that he measures actually applied to the LEDs.  The flatter the curve the larger the calculated error bars will be.