PhysicsProf Steven E. Jones circuit shows 8x overunity ?

[hartiberlin]

Here is some critique of the circuit I found on the Peswiki page by the user Motor Guy:

Motor Guy:
The circuit has huge stray inductances, and the transistor is not decoupled. That causes the transistor to oscillate horribly each time it starts to conduct substantial current. In the video the scope shows oscillations in the 100MHz range that good layout and decoupling of the transistor would eliminate. That 100MHz easily couples into the high impedance passive oscilloscope probe making the current readings completely erroneous. You can see those oscillations begin to disappear when he adjusts the rheostat he inserted into the low side of the left hand circuit loop. 

In the video Dr. Jones says that he has had a version running delivering 900mW out for 4mW in. 900mW is calculated by the oscilloscope. 900mW does not seem possible with these components. For his load that is no less than 5K Ohms, 900mW would mean more than 60V RMS at the transistor emitter, and peak voltages of around 100V. The MPS2222A CE breakdown voltage is only 40V. 900mW would also make the rheostat he has in series with his LED very hot. 

Dr. Jones needs to clean-up his circuit and his probes. For the circuit, using a PCB with a solid ground layer would be best. If he doesn't want to do that, he can probably do adequately by moving the transistor very close to the V+/V- strip of that EZ Circuit proto board, and adding a 0.1uF capacitor from the 2222A collector to V- using leads cut as short as possible. Once he cleans the circuit and the instrumentation up, he will find it is an ordinary oscillator that gets all of its power from the battery.

Motor Guy:
This is a nice demonstration of measurement error based delusion. Stray circuit and scope probe inductance cause invalid measurements. Clean-up the measurements and the illusion of over-unity will disappear. 

First, get rid of the huge pick-up loop formed by the scope probes' 6" ground clips. This can be done by placing a 0.1uF capacitor across the battery leads where they connect to the board, and using a coaxial probe connection at that point. The coaxial connection can be arranged by either cutting the probe off an old scope probe, or using a coaxial cable with a BNC at both ends and a BNC connector in series with a 50 Ohm resistor soldered right at the capacitor that is across the battery connection to the rest of the circuit. The 50 Ohm resistor is needed to suppress ringing in the coaxial cable. Second, suppress HF current flowing between the scope body and the circuit by clipping a bunch of those clamp-on ferrite EMC filters over each of the scope probe cables. Professor Jones can buy the clamp-on ferrites at Radio Shack for a few dollars each. 

The last problem that I see is that his circuit common should be defined as the negative terminal of the battery, not the bottom of the current viewing resistor. The reason for this is that the stray inductance of the resistor and wiring to the battery creates spikes that throw the measurements off. By setting the common at the bottom of the battery a coaxial probe can be soldered across the resistor right at the resistor body. Lead length between the resistor body and the negative side of the battery pack connection where it is picked up by the capacitor and voltage monitoring probe common must be kept to a minimum. 

If Professor Jones is sincere, he will clean-up his measurements and report the results. He can do so without spending more than $100. and a few hours of time.

Motor Guy:
Just to add that iit is important to keep the 0.1uF capacitor leads as short as possible. If Professor Jones has a good soldering iron, he can buy 1206 size surface mount parts for both the capacitor and the current viewing resistor. A 1206 resistor will handle 1/4 W, and while reasonably small, 1206 parts are still reasonably easy to solder with a fine soldering tip without using a magnifier.

Motor Guy:
One other minor thing I forgot to say: When the common is defined as the negative terminal of the battery the polarity of the sensed voltage will be opposite the current flow. Be sure to invert the channel to get the right polarity. The Tek scope can do that, and I'm pretty sure the ATTEN scope can as well. 

Also as with the voltage probe coax the coax from the current viewing resistor should have a series 50 Ohm resistor right at the end.

[hartiberlin]

Here are again the circuit and a few other pics about it:

[Omega_0]

Never trust the spiky waveforms, they can confuse even the most sophisticated instruments. Best way to measure them is to rectify them and measure the DC instead. Of course there will be some loss; but at 8x output it will not be an issue.
In this circuit even the input is spiky, which means double trouble.Right now I can't think of any way to measure the input reliably.

To protect the probes from radiation, shield the circuit by placing it in a metal box and running long thick wire to the rectifier placed far away.

Then there is the issue of scope ground. The scope probes have common ground and when you connect them at the same time to an ungrounded circuit, results become unpredictable.

If you get a good DC power out of it, its best to pulse it back into the input and get rid of scopes and meters. That will be the final test.....

[xee2]

@ JouleSeeker

I am sorry if I misunderstood how you are measuring the output power. But how can you measure the whole cycle using a scope? The voltage on the scope is only valid at one instant of time and changes over the cycle. This is how I measure efficiency > http://www.youtube.com/watch?v=smOiVmKv9f8

[JouleSeeker]

  Appreciate the comments and questions.  We have some family activities this weekend, but will have more time to respond later today and tomorrow. 

Good Morning Dr. Jones,

Following with interest, up to my eyeballs in RomeroUK motor (bearings and magnets!), but a question.  Have you tested or do you think this circuit can operate at AF frequencies, in the range of 1 to 3 kHz with larger inductors/cap., with the same ratio of input to output? Going on vacation for a week, soon as I get back, can build no problem and have a good scope to check it also.

Respectfully,
Ben K4ZEP

Right -- as you increase the Lb and Cb, the frequency of the tank circuit will go down.  I have not gone below about 500 KHz with this circuit, but I think your idea is a good one.    Please do try this, and let us know your results.