PhysicsProf Steven E. Jones circuit shows 8x overunity ?

[nul-points]

 

[...]
NP   Hope you keep experimenting... you might get to a self-runner yet   .
[...]

good morning Steven

i can't see me either getting a self-runner or even winning a competition with my test circuit - i don't think that the LED is visible much below 1.2V!!


i've rerun the cap/time tests for the reduced voltage as requested and it's looking like my LED/circuit combination is already below its 'critical voltage', because i can't state categorically that the LED is still visibly lit at 1.10V

maybe there's a basic performance difference between the usual JT oscillator config (as used by Xee2) which is a 'Common Emitter' type oscillator, and your SJ1 config which is a 'Common Collector' type oscillator?

i've used your SJ1 config (my PNP variant) for all these tests so far - at 1.10V supply the oscillator is still generating positive peaks of approx 2V, even though it's very difficult to see if the LED is still lit

anyway, the cap voltage/timing results for this test are as follows:

the circuit takes 191 seconds (3min 11sec) to discharge a nominal 1000uF cap from 1.37V to 1.10V

C2 1000uF (nominal)
1.37V => 0.938mJ
1.10V => 0.605mJ
         -------
    Ein: 0.333mJ

Pav: 0.333/191 = 1.74uW


cap leakage for 191 seconds from 1.37V:
1.370V => 0.938mJ
1.367V => 0.934mJ
          --------
    Ein: 0.004mJ

the cap leakage is less than 0.25%, which i think can be ignored (both here and in the previous 2.55V -> 1.5V test)

at least i have a better cap now for subsequent tests, even if my present LED/circuit is not so good!

i will try and get a more accurate capacity value for it, to use in future


i haven't found my 1N4148s yet (so will buy some more) - but i was thinking about your aim to standardise LEDs and had an idea for a possible solution:

Opto coupler devices contain an LED and these kinds of parts are more likely to be available as a standard distributed item than a particular LED part

they would only need to be used for the standard reference measurement - the usual visible LED could be used to experiment with the circuit for optimum visual results

an opto coupler would also open up possibilities to take some measurements direct from the transistor o/p

- either using a filtered DVM resistance range reading, like i used a few weeks back as a comparison with the internal LED DC power consumption;

- or making a simple amplifier circuit to provide a voltage o/p representing the DC average of the LED continuous power which could either be measured by DVM for 'spot readings' or be logged for graphs or accummulated data


just an idea


thanks
np



http://docsfreelunch.blogspot.com

[JouleSeeker]

Just saw your post, NP -- good morning! 

i don't think that the LED is visible much below 1.2V!!


i've rerun the cap/time tests for the reduced voltage as requested and it's looking like my LED/circuit combination is already below its 'critical voltage', because i can't state categorically that the LED is still visibly lit at 1.10V  -- NP

As noted this morning above, I've relaxed the request to run at a "common" voltage-range (also the request for using a 1N4148) since the device itself is affected significantly by the choice of LED (or diode) and by the voltage range. 

LED-off, then back-on effect:  I have re-confirmed the effect with my build of Xee2's circuit, as the voltage from the cap drops, the LED goes OUT then comes back on at a lower voltage -- and at much lower energy usage.  With a 5-ohm resistor in series with the green LED in my circuit, the LED goes out at 1.665 V and then, after several seconds as the cap loses voltage very slowly, the LED comes back on (very visibly) at 1.631 V.  I was glad to learn about this effect.

  I am also pleased that also that Xee's circuit allows resistance in series with the LED and works fine.  I get roughly the same power usage (Pinput) with 1, 5, and 21 ohms in series with the LED.    I did burn out two MPSA06 transistors yesterday in the process of playing with the circuit...       Not sure how I did that.

[NickZ]

  @ JS:
   You had mentioned that your voltage test with the single AA showed hardly no discharge rate.  Did that change now???
   My BwJt are still discharging the AA battery in two days.  So, I have not really seen the benefit of this circuit yet.
   I'll bet the capacitor test will only show discharging. While your first days test with the AA showed no discharge of the battery. 

[xee2]

I made some small changes to the very low power Joule thief video.
The new link is >>>  http://www.youtube.com/watch?v=qHdhMZzwQ_g

[NickZ]

   If the 5 of a buck 1" Goldmine ferrite toroids are not available, maybe the original Hartley type of air coil can be used instead. It was wound on a straw using 50 turns one way and 25 turns in the opposite direction.
   There may be loses due to the ferrite cores.  An air core may provide less losses, may be worth testing.
   The different types of cores will effect the result even when the other components are all the same.  I have several different cores going going 24/7, but they are all discharging the battery.  So, there must be a trick or two to learn still.