Is joule thief circuit gets overunity?

[TinselKoala]

Neither one, just simple electronics.
4 LEDs in series cause the voltage to rise higher before the LEDs turn on and cause it to drop again. This lowers the frequency of the JT, which in turn reduces the input current drain. But does it affect the light output? There is a difference between what the eye sees and what is actually happening. The eye sees a continuous light, but the LEDs are actually flashing on and off.
This phenomenon is why I changed my flashlight bank of 24 parallel LEDs into a bank of 2 x (12 parallel LEDs) in series, and why I put 4 x NE-2, 90-volt neons in series, in my HVJT.
Your Christmas gift to the world? Well, thank you LTseung for being so generous.

[TinselKoala]

Re-mystifying the Joule Thief: JT basic testbed (circuit identical to LTseung's) with addition of 70 nF poly film cap, switchable across 1k base resistor, and using the modified Loopstick as the inductor. Load is 24 LEDs, (12 parallel) + (12 parallel) in series. Note that I get pretty good "dim" brightness in the array, even at 1 V and 0.6 milliamps (600 microAmps) input. That is 25 _microwatts_ per LED. Even at the "full" brightness level (with cap) the system only draws 9 or 10 mA, or well under half a milliWatt per LED.
And....the system has several distinct operating modes, but a battery that is too full saturates it and makes it stop working.

http://www.youtube.com/watch?v=MeIWpkywGXs

Resonance or magic?!?

Or just electronics......

[e2matrix]

PFM      Interesting effect.

[TinselKoala]

More: Confirmation of DMM voltage drop across CVR reading, input current and output voltage waveforms, response to changing inductor setting.

http://www.youtube.com/watch?v=nqJPi35cRrE

[ltseung888]

Re-mystifying the Joule Thief: JT basic testbed (circuit identical to LTseung's) with addition of 70 nF poly film cap, switchable across 1k base resistor, and using the modified Loopstick as the inductor. Load is 24 LEDs, (12 parallel) + (12 parallel) in series. Note that I get pretty good "dim" brightness in the array, even at 1 V and 0.6 milliamps (600 microAmps) input. That is 25 _microwatts_ per LED. Even at the "full" brightness level (with cap) the system only draws 9 or 10 mA, or well under half a milliWatt per LED.
And....the system has several distinct operating modes, but a battery that is too full saturates it and makes it stop working.

http://www.youtube.com/watch?v=MeIWpkywGXs

Resonance or magic?!?

Or just electronics......

@TK,
Great job.  Now I appreciate my Atten Digital Oscilloscope much more.  If you can capture the instantaneous Voltage and the instanteneous Current on your scope and analyze them, you will get the instantaneous Power.  With the Input and Output Power waveforms, you can get the COP.

Your simple technique of varying the inductance is great.  I shall learn from that.  The tuning gets even better.  Resonance or magic???

If you have a higher capacitance (2.3V 10F), you can even take out the battery totally for say, 20 minutes.  You can then determine at what frequency range will get you the best COP.
If you use a DC Power Supply, you can find that a lower voltage may get you a higher COP which you already found with two different batteries.  But to find a commercial FLEET condidtion, the DC Power Supply is necessary.
By the way, my 4 LED set up was connected in parallel.  They gave more light, draw less current and gave higher COP..... Resonance or Magic???

Thanks for confirming the Christmas Gift in your different and brilliant way.