Testing the TK Tar Baby

[TinselKoala]

The instrumental measurements that I have shown using the Clarke-Hess integrating power analyzer are better measurements of circuit power performance than any Ainslie has shown. The Dim Bulb tests are better basic measurements of battery charge performance than any Ainslie has shown. When Ainslie claims that I have not shown any measurements.... she again makes false claims which are easily refuted.  For Stefan's benefit, since he is unlikely to want to prowl back through the thread, I will repost the Clarke-Hess measurements here. And there are videos full of measurements on the YT channel... just not a lot of colored numbers in boxes.

Conditions of testing: Tar Baby in negative-going gate drive pulse mode, making the oscillations, driven from 555 timer, heating a load, running on 4 batteries at a nominal 48 volts. In other words, just the same mode as shown in the first part of Ainslie's demo video but with one less battery. (I have six batteries, of course, but I am reserving two, chosen randomly, as "setasides" for Dim Bulb test comparisons.)

The first shot shows the "input" power: the Clarke-Hess is inserted between the battery and the rest of the circuit:
Battery >>> CH2330 >>> Tar Baby >>> Load

The second shot shows the "output" power to the load; the only change made was to swap the CH2330 over to the output side, a matter of switching cables that takes about 30 seconds to do.
Battery >>> Tar Baby  >>> CH2330 >>> Load

The third shot is the scope during the CH testing, negative pulse gate drive signal on the top at 10 V/div and the battery voltage at the battery terminals on the bottom trace at 20 V/div, zero references indicated by the graticle line nearest the grey dots on the right side of the bezel. The feedback oscillations have a slight envelope distortion (amplitude increase) caused by the CH's cable harness inductance.

The CH current readings are consistently about 10 percent higher than the readings on my cheapo inline ammeter. This current difference could reflect the fact that the CH is better at reading odd signals than the DMM is. And this current is adjusted by fine tuning the gate drive to the circuit and looking at the form of the "oscillations", which refutes yet another false claim that Ainslie has made about Tar Baby's performance. The differences in the input and output power readings represent power lost in the circuit that doesn't reach the load: mosfet heating, RF radiation, and so on.

The Clarke-Hess might be missing some power in higher frequency oscillations, though. That's why I also used an RF wattmeter on the Tar Baby, as if it were a radio transmitter putting power at 1.6 or 2.4 MHz into a transmission line or antenna represented by the load and batteries. The SWR sucks which means that the RF power is bouncing around and being dissipated within the circuit; not much of it is radiating, even though I can pick it up on my FM radio. So there is an issue with respect to calibration of the CH2330 to see if it's accounting for the RF power in its measurements. One thing seems fairly certain: Tar Baby isn't recharging batteries with RF-- since it is AC, and since TB's batteries do run down. Some is evidently heating the load which doesn't care about polarity reversals, and some load heating is coming from the "dc" that the partially on mosfets can pass, separate from the oscillations. The power levels are so small though that I might not be able to tease them apart with my crude calorimetry.


(Note that the CH is _not_ simply multiplying the average current shown by the average voltage shown. It is taking into account the phase differences involved in the signal it's measuring. I could select "power factor" on the display to indicate what the CH is detecting here, but for this measurement I'm just showing the "true power" as determined by the CH's integrating function.... combining the functions performed by the NERD RATs scope dumps and spreadsheet analysis and doing it properly... within its frequency limitations of course.)

[Rosemary Ainslie]

The instrumental measurements that I have shown using the Clarke-Hess integrating power analyzer are better measurements of circuit power performance than any Ainslie has shown. The Dim Bulb tests are better basic measurements of battery charge performance than any Ainslie has shown. When Ainslie claims that I have not shown any measurements.... she again makes false claims which are easily refuted.  For Stefan's benefit, since he is unlikely to want to prowl back through the thread, I will repost the Clarke-Hess measurements here. And there are videos full of measurements on the YT channel... just not a lot of colored numbers in boxes.

Conditions of testing: Tar Baby in negative-going gate drive pulse mode, making the oscillations, driven from 555 timer, heating a load, running on 4 batteries at a nominal 48 volts. In other words, just the same mode as shown in the first part of Ainslie's demo video but with one less battery. (I have six batteries, of course, but I am reserving two, chosen randomly, as "setasides" for Dim Bulb test comparisons.)

The first shot shows the "input" power: the Clarke-Hess is inserted between the battery and the rest of the circuit:
Battery >>> CH2330 >>> Tar Baby >>> Load

The second shot shows the "output" power to the load; the only change made was to swap the CH2330 over to the output side, a matter of switching cables that takes about 30 seconds to do.
Battery >>> Tar Baby  >>> CH2330 >>> Load

This is not clear.  Where on the circuit have you put the CH2330?  In the NERD circuit the load is clearly indicated in a schematic.  You have given us a multiple and optional reference with no clarity at all.  Have you put the CH2330 on either side of the battery?  At its positive terminal and then its negative terminal?  Or have you put the CH2330 on the near and then far side of the load in series with the positive terminal?  If the former - then the readings should be co-incident.  If the latter then the readings bear no relevance to the 'input' and 'output' as you claim.  And if you are drawing a distinction between the input and the output then exactly what are you distinguishing?  To me that amperage value looks like the 'sum' of both the input and the output.  In which case?   On our NERD circuit, the sum of our voltages gives us a current flow that shows considerably more back to the battery than delivered by the battery.  Which is a negative voltage value.  If your Ch2330 is not showing a negative current flow resulting from that negative voltage sum then you have not replicated our values. Nor have you replicated our waveform across the batteries. In which case I would expect your batteries would discharge.  And our range of battery oscillation is considerably greater than that shown on your circuit.  Which gives our circuit considerably more advantage over both cycles of each oscillation.  I suspect that your lack of voltage may be because your load is not sufficiently inductive.  And there is no consistency between each oscillation period - the one varying from the other.  Therefore is there no consistency in the claimed results.  That's the pivotal requirement related to any claimed measurement.   

Regards
Rosemary

ADDED

[Rosemary Ainslie]

The CH current readings are consistently about 10 percent higher than the readings on my cheapo inline ammeter. This current difference could reflect the fact that the CH is better at reading odd signals than the DMM is. And this current is adjusted by fine tuning the gate drive to the circuit and looking at the form of the "oscillations", which refutes yet another false claim that Ainslie has made about Tar Baby's performance. The differences in the input and output power readings represent power lost in the circuit that doesn't reach the load: mosfet heating, RF radiation, and so on.

The Clarke-Hess might be missing some power in higher frequency oscillations, though. That's why I also used an RF wattmeter on the Tar Baby, as if it were a radio transmitter putting power at 1.6 or 2.4 MHz into a transmission line or antenna represented by the load and batteries. The SWR sucks which means that the RF power is bouncing around and being dissipated within the circuit; not much of it is radiating, even though I can pick it up on my FM radio. So there is an issue with respect to calibration of the CH2330 to see if it's accounting for the RF power in its measurements. One thing seems fairly certain: Tar Baby isn't recharging batteries with RF-- since it is AC, and since TB's batteries do run down. Some is evidently heating the load which doesn't care about polarity reversals, and some load heating is coming from the "dc" that the partially on mosfets can pass, separate from the oscillations. The power levels are so small though that I might not be able to tease them apart with my crude calorimetry.

The rest of this post is simply an admission of the inadequacy of your measuring instruments.  Why are you not simply analysing the waveform across a shunt resistor to determine amperage?  That should be possible even with your scope.  And even if you can't finely integrate your values.  At least then we'd see the 'trend' of that voltage - whether or not it is greater above or below zero?  Right now the evidence is that you've partially replicated our battery waveform - but without the required amplitude.  And that you have simply NOT replicated our current flow as you give no evidence of it's voltage one way or the other.  The signal from the switch is not relevant to this power analysis.  It will only be relevant when you need to prove that the battery supplying that switch may or may not be adding energy to the system.  And that's another thing.  Are you using a separate battery?  Or are you powering the switch from one of the four batteries you use for your input.

If you are going to claim that you've debunked - I'd have thought that all these points are critical.  Why are you not addressing them?

Regards,
Rosemary 

[TinselKoala]

The noise will be ignored.

[Rosemary Ainslie]

And TK - don't give me that 'the NERD team have shown no measurements.  We give copious measurements.  The problem is that we dare not compute the wattage as we do not know how to resolve the anomaly of a negative product related to this.  It has no relevance to any known power measurement.

And our measurements are clear and have been conducted on highly sophisticated machines.  What is true is that we have not tested our apparatus switched by a 555.  That's due.

Again,
Rosemary