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

[Omnibus]

Let me say a little more. It is unclear what Vin, if nothing else is known, really includes and therefore it is incorrect to assign it to Ohms law only. In order to understand what it really means, you have to find an independent way of measuring power. So, on the one hand you measure power as Iin*Vin and that's indisputably the input power. On the other hand, if CoE holds, that Iin*Vin must equal the power calculated independently by using Joule's first law (Iin^2*R). Any discrepancy in favor of (Iin^2*R) is overunity beyond a doubt, no matter whether we observe transformers, separate coils or any electrical devices for that matter.

[Omega_0]

This plot is of 1Hz data that you posted above.
I have plotted Ein and Eo for two values of Rin. (Original Rin and a value of 0.0085 ohm where Ein==Eo)
Now a measurement error of 0.004741 ohm is predicted, as any other engineer will say surely.

This value is tiny and error can not be ruled out. There are many possibilities how this error may occur:

- the resistance meter is not calibrated
- the resistor changes in value when heated due to current
- while measuring the connection resistance is less and gives low value (loose connections in working circuit)
- meter is not very very accurate (we need a meter with 1/1000 ohm or more precision here to measure such a low resistance)
- many other factors

So I still think more accurate measurements are needed. The setup is not within error limits.
I suggest you get the industrial/lab quality 1 ohm +/- 1% shunt resistor with secure connections and measure again.

Ein and Eo are also equal at Rc = 28 ohm, but thats too big a variation to worry. Problem is mostly with Rin

[Omnibus]

@Omega_0,

Of course, all these concerns will vanish when current is measured with a proper current probe and LCP0030 appears to be such: http://www.testpath.com/Items/AC-DC-Split-Core-Current-Oscilloscope-Probe-120MHz-30A-117-121.htm. Also, first thing in the morning will be to order a shunt resistor. @LarryC gave a very good link for that: http://www.rc-electronics-usa.com/buy-current-shunt.html . The resistance of the shunt, however, should not be 1Ohm because, as @LarryC pointed out, it will draw power. I tried various resistances and found out that their value is crucial for the effect to be observed because thay become unneeded part of the circuit and interfere with the effect. The lower the resistance of the shunt resistor, the more the effect shows up. Of course, there's a limit set up by the scope resolution. And, to say it again, the least such parasitic intereference with the studied circuit will be induced by this so much mentioned Hall effect current probe.

Nevertheless, the points you raise can be addressed even at this time:

- the resistance meter is not calibrated -- I'm using a Keithley 2000 multimeter as is, haven't sent it for calibration so that is a legitimate point. Hope the precision shunt resistor I'm going to order may shed some light on that.

- the resistor changes in value when heated due to current. Current is on the order of several hundred mA and, indeed, one may think that approx. 2 inch long RadioShack magnet wire may heat up. It didn't feel, though. Besides, the effect shows up with a resistor made of about 40cm connector wire (that standard connector RadioShack wire) as well as with even longer such wire (in fact a whole spool of those standard spools RadioShack sells).

- while measuring the connection resistance is less and gives low value (loose connections in working circuit). That's out because, having in mind such possibility, I soldered all the connections, save those of the probes, of course.

- meter is not very very accurate (we need a meter with 1/1000 Ohm or more precision here to measure such a low resistance). The meter I'm using (Keithley 2000) does have such precision. So, that's out too.

- many other factors -- can't think of any that would substantially alter the result, other than the above.

To address the measurement error, the used value 0.003759Ohm is the average (as Excel gives it; when minding the significant figures it should be 0.0038Ohm) of 22 separate measurements at the interval of 1min. The standard deviation is 0.000105Ohm. Therefore, the value of 0.0085Ohm as an upper error limit ou give is more than two orders of magnitude higher than the real error limit.


P.S. Although the spreadsheet contains Eo values, such are not measured. Therefore, arguments based on Eo should be ignored. Recall, only the input voltage and the voltage across the resisor Rin are being measured.

[IceStorm]

@Omega_0,

Of course, all these concerns will vanish when current is measured with a proper current probe and LCP0030 appears to be such: http://www.testpath.com/Items/AC-DC-Split-Core-Current-Oscilloscope-Probe-120MHz-30A-117-121.htm. Also, first thing in the morning will be to order a shunt resistor. @LarryC gave a very good link for that: http://www.rc-electronics-usa.com/buy-current-shunt.html . The resistance of the shunt, however, should not be 1Ohm because, as @LarryC pointed out, it will draw power. I tried various resistances and found out that their value is crucial for the effect to be observed because thay become unneeded part of the circuit and interfere with the effect. The lower the resistance of the shunt resistor, the more the effect shows up. Of course, there's a limit set up by the scope resolution. And, to say it again, the least such parasitic intereference with the studied circuit will be induced by this so much mentioned Hall effect current probe.

Nevertheless, the points you raise can be addressed even at this time:

- the resistance meter is not calibrated -- I'm using a Keithley 2000 multimeter as is, haven't sent it for calibration so that is a legitimate point. Hope the precision shunt resistor I'm going to order may shed some light on that.

- the resistor changes in value when heated due to current. Current is on the order of several hundred mA and, indeed, one may think that approx. 2 inch long RadioShack magnet wire may heat up. It didn't feel, though. Besides, the effect shows up with a resistor made of about 40cm connector wire (that standard connector RadioShack wire) as well as with even longer such wire (in fact a whole spool of those standard spools RadioShack sells).

- while measuring the connection resistance is less and gives low value (loose connections in working circuit). That's out because, having in mind such possibility, I soldered all the connections, save those of the probes, of course.

- meter is not very very accurate (we need a meter with 1/1000 Ohm or more precision here to measure such a low resistance). The meter I'm using (Keithley 2000) does have such precision. So, that's out too.

- many other factors -- can't think of any that would substantially alter the result, other than the above.

To address the measurement error, the used value 0.003759Ohm is the average (as Excel gives it; when minding the significant figures it should be 0.0038Ohm) of 22 separate measurements at the interval of 1min. The standard deviation is 0.000105Ohm. Therefore, the value of 0.0085Ohm as an upper error limit ou give is more than two orders of magnitude higher than the real error limit.


P.S. Although the spreadsheet contains Eo values, such are not measured. Therefore, arguments based on Eo should be ignored. Recall, only the input voltage and the voltage across the resisor Rin are being measured.

The problem with your setup is not the way you measure it, its how you understand impedance with a inductor. You expect Rprimary to be a static value when its not, the value change as the frequency change, the formula is : 2 * Pi * F * L . F is the frequency in Hertz and L is the inductance value in Henry.

So suppose you have a coil , 250mH, DC resistance is 5 Ohm , what mean 5 ohm at 1KHz ? nothing , here why
2 * 3.1415926535897932384626433832795 * 1000 * 0.250 = 1570.7963 Ohm for Rprimary and not 5 ohm.

Best Regards,
IceStorm

[Omnibus]

@IceStorm,

While I agree with you about the inductance of the coil, I do not agree with respect to the active resistance of a coil. The active resistance of a coil is not a function of frequency (inductance is). The fact that impedance of the coil changes with frequency (because of inductance changing with frequency) in no way affects the results I'm presenting. Once the active resistance of the coil is correctly measured all remains as already presented.