PhysicsProf Steven E. Jones circuit shows 8x overunity ?

[JouleSeeker]

Having received a question via email about measurements --  my reply:

How I made the power measurements is as follows, see attached schematic for the numbers of the points on the circuit where scope connections were made.

Pinput:  Connect 2 probe grounds to point 1.  Then one probe to point 2 (for current measurement, voltage across 1-ohm resistor) and one probe to point 3 (for voltage measurement).  Probes x10.
  Then Pin = the math product of the two probe voltages.  I then take numerous cycles of the waveform and have the Tek 3032 calculate the MEAN input power.

Poutput:  Connect 2 probe grounds to point 4.  Then one probe to point 5 (for current measurement, voltage across 1-ohm resistor) and one probe to point 6 (for voltage measurement).  Probes x10.
  Then Pin = the math product of the two probe voltages.  I then take numerous cycles of the waveform and have the Tek 3032 calculate the MEAN output power.
Note that when Rr is zero/removed, then point 4 is connected to/becomes point 1 and one does not have to move the two ground probe-connections when going from Pin to Pout measurements.

I have also measured Pin (as a check) using the Cap/stop-watch method described on the thread.  Ein = 1/2CV**2 and Pin = Ein/time.  This gave me Pinput ~ 0.23 to 1.1 mW, on two separately-built devices.  This measurement is in reasonable agreement with the Tek 3032 method described above, but appears to disagree with the DMM measurement method.  I trust the Cap/time method the best as this relies simply on input Energy and the time for the partial discharge of the cap -- and the voltages on the Cap before and after the run are done with the cap disconnected from the circuit.

Poutput could be measured via temperature rise on Ro, eliminating the need for CSRout.  In that case, I would replace LEDout with a diode.  I have not done that test yet.

In these ways, one could measure Pin and Pout without the use of oscilloscopes.

But I think that a self-running device would demonstrate super-efficiency more compellingly than such measurements.

Best wishes,
Steve 
On the road again in minutes

http://www.youtube.com/watch?v=FLRqM0nDT_M

[TinselKoala]

Hi Bruce,

Thanks for the advise. I am always open to constructive suggestions, TinselKoala has not replied and I don't expect one now. Maybe he has looked at my data again and thought other wise?
(snip)

P.S. If anyone See's a mistake please let me know. Sometimes I work on this under tired wary conditions so like anyone I do make mistakes. I do try to always double check things however. A better way to do something is always a better way!

I'm trying, I really am.
If you want to see some of what I've been working on in the past, you can look at my YT channel, and put "Joule Thief" into the search window. I have posted a couple of videos concerning measurement of energy using oscilloscopes in the Joule Thief and in the Rosemary Ainslie circuit. Please note the dates, they are pretty old.
Also, you might look at the Rosemary Ainslie thread here, and look at what poynt99 is trying to illustrate.
The main problem that I see here is that some intelligent and creative people are wasting a lot of time making meaningless measurements, when they have the apparatus at hand to attack the problem properly. They might not get results they like, though, and THAT is really the problem.
I am willing to do "power analyses" the same way you are doing them. Let's start with my TinselKoil and see what kind of COP we get. I am also willing to do them properly and share the techniques for doing so.
Now... are you willing to do the measurements the way I suggest, just for comparison's sake? That is, use the amount of energy in a cap bank, applied over a period of time, for input energy, and use the integrated instantaneous power curve over the same time, for output energy, and compare the two, for a true COP efficiency value?

(ETA: Once we are taking proper energy measurements, we can start talking about probe placement, stray inductances and circuit layout. These are so critical that they can actually have large effects on measurements and the calculated values derived from them. Very very large.....)

[TinselKoala]

Perhaps the difficulty is that it isn't being understood, that the energy is calculated by multiplying the Power times the Time over which the power is dissipated.
In the DC case, the "instantaneous" multiplication of voltage and current yields a straight line. So you just take some time period... like a scope screen .... and multiply the "average power" in Watts, which is also the value of the straight line power, by the time interval in Seconds, and that gives you the energy in Joules passing your measurement point in that time period.
It's just the same problem as finding the area of a rectangle, with the height equal to the wattage and the width equal to the time.
The "Integration" of any curve, geometrically, corresponds to the area under it, just like in the rectangular case above. Find the area and you've done the integration. For a rectangle DC waveform that's easy, it's height times width. But for a real AC HF power curve you can't do that.
So... for a complicated signal like an instantaneous power waveform, you need some way of approximating the area under it during a time interval, to get to the Joules of energy. I've described a simple manual way of doing it using tracing paper, but it's fairly easy --if relatively expensive, ha ha -- nowadays to get the oscilloscope to do it for you.

[Omega_0]

@hyiq

I agree with TK. Setup a measurement protocol first.
Since you are in a milli and micro range, nothing is reliable. Even touching the probe or keeping a PC on besides it will alter everything.

Since the waveforms don't allow direct looping back, the only way is proper measurements. (For those who don't have expensive calorimeters)

I'd also advise you to go back to the original circuit of Prof. We all have a tendency to forget that we are doing a "Replication" and start twisting the original to suit our own convenience. I see that you are falling in the same trap.

All the best.

[yfree]

@hyiq

Chris,
You should watch this video:
http://www.youtube.com/watch?v=bcftGrBEaL0&feature=player_embedded#at=25