PhysicsProf Steven E. Jones circuit shows 8x overunity ?

[NickZ]

  @JS:
   Your device source capacitor drained a volt in 12 seconds,  so where is the gain??? Or am I missing something? 

[JouleSeeker]

OK, I saw that you stated the following:

Is that in reference to the start or end of the run?

Super-brights! 

.99

Clearly, we would need to quantify a MINIMUM brightness for the LED to make useful comparisons between circuits...   

I have another idea -- experimenter will replace his LED with a common diode, 1N4148, and then use the stop-watch/Cap method to determine INPUT power.

In this way, we approach a reliable STANDARD METHOD for evaluating the INPUT POWER for various devices.

And the input voltage needs to average out to 2 volts, at present.  Hmmm...  thinking of a standard, we may want to go from 3.24 V (fresh AA's to charge the input-energy Cap) to 1.50V on the cap.

[poynt99]

Clearly, we would need to quantify a MINIMUM brightness for the LED to make useful comparisons between circuits...

This only works if everyone uses the same type of LED, and everyone has a means of measuring the LED's intensity.

.99

[JouleSeeker]

  @JS:
   Your device source capacitor drained a volt in 12 seconds,  so where is the gain??? Or am I missing something?

Nick -- right now, we're working on a STANDARD means of measuring the INPUT POWER, to compare various circuits reliably, repeatable by anyone with a voltmeter and a stop-watch.  Easy.
Does your circuit draw less power than mine?  we can soon know with a repeatable method.

I replaced my LED with a 1N4148 diode, from 2.55 to 1.5 V came out at 12.6 seconds -- still at 170 uW (OK, 169 by the calculator).

GAIN is of course the next question, and requires a method to measure the output POWER reliably.  Do you have a good way to measure Poutput so we can get the efficiency? 

I'm proposing a Thermal Wattmeter, given the strong AC components in the output typically... Still working on that.

[JouleSeeker]

This only works if everyone uses the same type of LED.

.99

AS I SAID,  test requires that you replace the LED with a 1N4148 for the "final" test, to permit comparison.  Maybe our posts are crossing in ether-space?