PhysicsProf Steven E. Jones circuit shows 8x overunity ?

[xee2]

hence Steven's guideline of a voltage range between 2.55V and 1.5V 

If you want to use 2.55 volts you can. But, as you noted, to get lowest power consumption the battery voltage should be as low as possible. The test voltages I use are what ever the battery voltage happens to be when I do the test.

[nul-points]

If you want to use 2.55 volts you can. But, as you noted, to get lowest power consumption the battery voltage should be as low as possible. The test voltages I use are what ever the battery voltage happens to be when I do the test.

you can use any voltage you like (within the transistor device operating limits) for your own testing

in that case, however, the results will have no meaning for comparing between different approaches and circuits - which is why Steven asked people to make their measurements within his guidelines

[xee2]

may be one can combine Per-Johan's idea (third coil on the toroid with full bridge rectifier) and your low power JT circuit in order to feed back power to the 10.000µF capacitor? This should reduce the power consumption even more.

Hi Conrad,
My testing has shown that there is no gain in efficiency by using an additional coil on the toroid. I only use the extra coil to get higher output voltages.

[xee2]

you can use any voltage you like (within the transistor device operating limits) for your own testing

in that case, however, the results will have no meaning for comparing between different approaches and circuits - which is why Steven asked people to make their measurements within his guidelines

I am sorry my tests have no meaning for you. To me the meaning is quite clear - if you want to reduce power consumption you should reduce the input voltage.

[JouleSeeker]

thanks for the kind words as always, Steven...

i'll just bask in those words a little longer - before having to point out that your quote about OU statistics must have been posted by someone other than me! 

here is my contribution to the list of reported OU development, hopefully with a reasonably mainstream emphasis:-

thanks
np


http://docsfreelunch.blogspot.com

   Oops, I thought you posted this, but apologize for the evident mistake... 
And thanks for your sense of humor, NP, gave me a hearty laugh this morning.

Xee2 -- yes, by increasing the cap feeding the base transistor, you decrease the pulse-rate on the LED (to 18 Hz) so that it flickers, and this decreases the input power.  It is remarkable to get the device to run for over 80 minutes on a 10,000 uF cap, but at the same time, let's note that decreasing the rep rate does not (almost certainly) change the Pout/Pin ratio.  Increasing that ratio, to a value > unity, is the major goal of this line of research.