Quantum Energy Generator (QEG) Open Sourced (by HopeGirl)

[TinselKoala]

There are also methods to use the coil itself as the sensor based on the coil resistance and inductance.  A lot of high current power supplies use that sort of method in order to avoid the power loss of a dedicated current sense resistor.  I have a colleague who is very good at implementing that method.  I'd rather avoid all the work and just use a dedicated current sense resistor of an appropriate value and power rating whenever I can.

That is exactly what I am trying to do, to use the primary coil itself as the sensor for the current magnitude. Its DC resistance is negligible and its inductance is about 4.25 microHenry so its reactance at 303 kHz is about 8.1 Ohms. Neglecting capacitance which, in spite of the TBF winding, is also probably negligible, and solving I_p-p=V_p-p/Z I arrive at the p-p current value of about 10 A in the coil. But phase information is lacking and I thought that the CT method might inform me of the phase shift in the current, and it apparently has done. I think I am measuring a true phase shift of about 87 degrees. So using the V_p-p measured across the coil as both straight V and also for the I by the previous logic, and then solving (Vrms)x(Irms)x(cos87) I arrive at about 5.4 Watts real power in the coil, which tallies well with the input power to the system, being about 80 percent of the input. At least it is of the same order! Please let me know what you all think of the "legitimacy" of this procedure.

Now, all this might just be hand-waving and coincidence of errors, and I'm definitely learning the trade here. But I may be able to make better measurements as I get more information and comments and help as we have been doing here!

Unfortunately the tank circuit is very sensitive and inserting a CSR in series must be done carefully and with the right components. Every time I've tried it, even using the very good Ohmite NI resistors of 0.25 ohm, the resistors heat up rapidly so I only have seconds to take a scopeshot, but when I do I get Vdrop values that are relatively in line with current calculations based on the coil and its inductance alone.

Now... the real issue I am trying to address and that I don't want to bury, is the "real" QEG current measurement using their Stangenes CT.  Note that in the image below, the CT is connected to a BNC-to-Banana adapter and a scope probe is connected by its spring clips to the adapter. (insert facepalm here).
What phase artifacts are being introduced here? In later pics they do appear to have the transformer connected with an actual patch cord instead of the kludge, but the issue of _their_ actual and measured phase shift remains to be answered.

ETA: even when they used the actual patch cord... did they remember to set the scope channel's input impedance to 50 ohms instead of the default 1 megohm?
Inquiring minds want to know. Was this item used as C-to-V as I believe, or as C-to-C ?

[TinselKoala]

Output of a web-based Z calculator, given the microQEG's 64 nanoFarad tank cap value and the (rounded) 4 microHenry primary inductance:

Ain't theory wonderful? Bada-bing!

ETA: Recall that I have measured the operating frequency as 303.59 kHz with the Philips counter, my most accurate instrument.

[MarkE]

That is exactly what I am trying to do, to use the primary coil itself as the sensor for the current magnitude. Its DC resistance is negligible and its inductance is about 4.25 microHenry so its reactance at 303 kHz is about 8.1 Ohms. Neglecting capacitance which, in spite of the TBF winding, is also probably negligible, and solving I_p-p=V_p-p/Z I arrive at the p-p current value of about 10 A in the coil. But phase information is lacking and I thought that the CT method might inform me of the phase shift in the current, and it apparently has done. I think I am measuring a true phase shift of about 87 degrees. So using the V_p-p measured across the coil as both straight V and also for the I by the previous logic, and then solving (Vrms)x(Irms)x(cos87) I arrive at about 5.4 Watts real power in the coil, which tallies well with the input power to the system, being about 80 percent of the input. At least it is of the same order! Please let me know what you all think of the "legitimacy" of this procedure.

Now, all this might just be hand-waving and coincidence of errors, and I'm definitely learning the trade here. But I may be able to make better measurements as I get more information and comments and help as we have been doing here!

Unfortunately the tank circuit is very sensitive and inserting a CSR in series must be done carefully and with the right components. Every time I've tried it, even using the very good Ohmite NI resistors of 0.25 ohm, the resistors heat up rapidly so I only have seconds to take a scopeshot, but when I do I get Vdrop values that are relatively in line with current calculations based on the coil and its inductance alone.

Now... the real issue I am trying to address and that I don't want to bury, is the "real" QEG current measurement using their Stangenes CT.  Note that in the image below, the CT is connected to a BNC-to-Banana adapter and a scope probe is connected by its spring clips to the adapter. (insert facepalm here).
What phase artifacts are being introduced here? In later pics they do appear to have the transformer connected with an actual patch cord instead of the kludge, but the issue of _their_ actual and measured phase shift remains to be answered.

ETA: even when they used the actual patch cord... did they remember to set the scope channel's input impedance to 50 ohms instead of the default 1 megohm?
Inquiring minds want to know. Was this item used as C-to-V as I believe, or as C-to-C ?

If you know the wire gauge and length then we can estimate the DC resistance.  But if jwL ~= 8.1 Ohms at the operating frequency and you've got 10A p-p ~= 7.1A rms ~= 50W/Ohm then as isim suggested, a 0.01Ohm 1W or larger non-inductive Kelvin resistor would do the job with no muss and no fuss.

The formula:  P_REAL = V_RMS * I_RMS * COS(theta) is valid for undistorted sine waves.  You're waveforms look  pretty clean, so you're estimate passes basic sanity checks.

A 6" clip lead adds very roughly 60-120nH.  What they did was poor practice but would hardly create a gross error against a 4-5uH UUT.  The bigger problem is stray pick-up.  Lots of current and voltage switching around makes it easy to pollute the instrumentation.  Eliminating antennas, and suppressing common mode currents gets important.

Something like this mounted to a PCB would work: http://www.digikey.com/product-detail/en/SR10-0.010-1%25/SR10-0.010-FCT-ND/2138980

[TinselKoala]

The microQEG primary coil wire is about 171 cm of solid copper wire with diameter 1.59 mm. The coil is wound on a 3.25 inch OD cardboard form, 6 turns "Tesla BiFilar" closewound spaced by insulation, with 3 inch leadin legs. My inductance meter reads it between 4 and 5 microHenry and calculations seem to indicate it is right about 4.25 uH.

Unfortunately I am up against the limitations of my kit, I fear. I can't easily measure fractional-ohm resistances precisely so making a custom shunt for 0.01 ohms would be a matter of guesswork.

[TinselKoala]

OK, here is some more data. I was able to mount up the 0.25R precision Ohmite non-inductive resistor stack, and the insertion loss and heating wasn't as bad as I thought. (The traces of overheat on the resistors are from an earlier unfortunate event, but the resistors are still OK.) The data is noted on the scopeshot below. The CSR gives me the honest phase relationship as wired in, I think, but I am only measuring about 70-72 degrees. This is embarrassing because if it is correct, the real power in the coil is over 30 Watts, but I'm only putting in about 6 Watts DC power.



The shot is taken with the TKTransverter in place and the light bulb glowing brightly.

ETA: The measured frequency is now 340.46 kHz, due to the insertion of the resistor, so there is some added inductance there.