STEORN DEMO LIVE & STREAM in Dublin, December 15th, 10 AM

[Omnibus]

Here are the two ways of measuring current, posted by @teslaalset: http://www.overunity.com/index.php?topic=8411.msg247902#msg247902

[gyulasun]

Thank you. 

Now it is clear that I misunderstood you, I had thought you were speaking of measuring the voltage across the 115pF capacitor with the 1:1 probe...  sorry.  (So in the schematic there should have been a second voltage probe on the right hand side, across the capacitor. The voltage probe on the left measures the input voltage and the other on the right measures the capacitor voltage...  that is how I thought. Earlier you had had such setup with at least 3 voltage probes.)

Now it is clear that my suggestion of using a 5pF cap instead of the 115pF is not good, unless the voltage across the C member is to be measured with the 1:1 probe, which is not case now.

Regarding the two measurement setups drawn by teslaalset, I agree his notices, and I think in the bottom case where the 1:1 probe is connected directly across the FG output, a remedy to avoid the phase shift of the probe's 110pF capacitance, you may wish to connect an inductance of the same reactance value to compensate for the 110pF.  Putting it otherwise, you neutralize the reactive effect of the cap with an equal but opposite reactance, constituting a parallel LC circuit at the FG output.
If your output freq is say, 800kHz, then X_c=1.8kOhm and the compensating inductance from X_L=1.8kOhm give about 360uH.
The output impedance of the FG will remain at 50 Ohm resistive as before because the reactances would cancel at the chosen frequency. The bonus is the 1:1 probe will not cause unwanted phase shift anywhere in the circuit.

rgds,  Gyula

[Omnibus]

@gyulasun,

That's a good idea. I'd appreciate it if you could sketch it to avoid any misunderstanding. Makes you wonder why this isn't a general practice with these passive probes -- it would avoid spending so much money for the active and many times of the differential probes. I guess, this techniques has to be applied with those probes too. Also, another thing -- this technique can be applied both to the 1:1 and to the 1:10 probe but then why use the 1:10 probe at all? Shouldn't I always stay with the 1:1 probe once that parasitic capacitance of the probe is taken care of. What do you think?

Do you think this method will be applicable n the measurements of transformers and various coils as well?

So, now, once I do that (add that compensating inductance), the only thing I have to worry about is the accuracy of the scope. Here is an example of a 14-bit scope: http://www.gage-products.com/14-bit-digitizers/?utm_source=google&utm_medium=cpc&utm_campaign=updated-north-america&gclid=CKP-2LfNo6MCFYdb2godLHIw5w. Do you think it would be better to carry out the measurements with this kind of scope rather than with a high end Tektronix scope which I don't even think has accuracy greater than 11-bit?

[gyulasun]

....
Makes you wonder why this isn't a general practice with these passive probes -- it would avoid spending so much money for the active and many times of the differential probes. I guess, this techniques has to be applied with those probes too. 

NO.  This compensation is frequency dependent, it is impossible to make it wideband for the full range of probes, valid only for one frequency for which you actually need compensation i.e. where you measure at.

Also, another thing -- this technique can be applied both to the 1:1 and to the 1:10 probe but then why use the 1:10 probe at all? Shouldn't I always stay with the 1:1 probe once that parasitic capacitance of the probe is taken care of. What do you think?

1:10 probes have 10MegaOhm input impedance @ 14-20pF parallel capacitance, this is much better than 1:1 probes 1MOhm @ 100-120pF parallel capacitances, this is why 1:10 probes are used, then come the active probes.  Parasitic capacitance can be taken care of at always one particular frequency only, in your case at 800-900kHz and you have to decide on frequency, then calculate inductance.

Do you think this method will be applicable n the measurements of transformers and various coils as well? 

ONLY in case you are fully aware of the effect of parallel resonance at a certain frequency, and also consider the terminating impedances of the transformers. For instance, capacitance compansation is a well known method in case of wide band transmission line transformers used for impedance transformations over some hundred MHz. In most cases the impedances are low (from some Ohms to some hundred Ohms), this means the parallel resonance effect cannot create unwanted amplitude increase at the resonance frequency, the loaded Q is low and controlled.

NOTICE that I added a capacitor in series with the compensating coil: this is needed when you switch on the DC offset of the FG for your further measurements, the cap prevents the coil's direct short circuit effect on the output DC voltage. The capacitor's value can be a 220-470nF, not critical in this high range with respect to the 110pF small value. (say you choose 220nF, this does create a series resonant frequency at 17-18kHz with the  360uH , well away from your 800-900kHz test frequency.)

So, now, once I do that (add that compensating inductance), the only thing I have to worry about is the accuracy of the scope. Here is an example of a 14-bit scope: http://www.gage-products.com/14-bit-digitizers/?utm_source=google&utm_medium=cpc&utm_campaign=updated-north-america&gclid=CKP-2LfNo6MCFYdb2godLHIw5w. Do you think it would be better to carry out the measurements with this kind of scope rather than with a high end Tektronix scope which I don't even think has accuracy greater than 11-bit?

I am not much familiar with those cards, a quick glimpse on their input impedance shows 1MOhm @ 40pF or 50 Ohm, selectable inputs. So there you would have input capacitance too at the 1MOhm input...  Otherwise the resolution and accuracy sounds very good. Maybe other members here can comment this too.

[Omnibus]

Thanks @gyulasun. I guess we should also take into account the parasitic capacitance of L. All in all probably it would be better to seek more advanced probes such as http://www.tek.com/products/accessories/differential.html although even then there may be questions in view of the extreme sensitivity of the OU effect especially regarding the I-V phase shift. It appears this is a major issue for all such measurements, in addition to errors from calculations based on thousands let alone hundreds of thousands of points.

Because the question for the reality of OU is a 'yes' or 'no' question there are then only two ways to prove OU indisputably -- through demonstrating a self-sustaining device or through proving  theoretically that OU is inherent in electric devices under certain conditions. I have shown that that's the case for certain voltage offsets. That's definitive. It can also be speculated that the core or other construction details (the mode of wire winding, for instance) would cause the I-V phase shift to be different from what's expected based on the components of the circuit and that would cause OU. This can be shown theoretically. Too bad that speculation cannot be checked experimentally. Like I said, no matter what you do to show it in an experiment, short of presenting a self-sustaining device, there will always be questions regarding the accuracy and precision of the measurements such as the ones we're discussing here.

Having seen no self-sustaining device in any of the attempts to demonstrate OU in electrical systems, the analysis involving the voltage offset is the only reason so far that gives categorical assurance that OU in electrical systems is real.