Testing the TK Tar Baby

[TinselKoala]

Figure 1: Overview, 5 each IRFPG50s on cute little heatsinks of aluminium, anodized..... fan blowing on Q1 mosfet, load heating at +-700 mA. Load is the 10.3 ohm 60 Watt resistor stack in 250 mL mineral oil.

Figure 2: Inline cheapo DMM readings at battery.

Figure 3: Typical scope trace at 700 mA. Top is common drains at 100 V/div, bottom is battery at 20 V/div, normal timebase 0.2 ms/div, expanded TB  1 us/div. Osc frequency about 2.4 Mhz, FG at 1 kHz, gate drive (not shown) offset and amplitude fiddled with to obtain 700 mA on DMM.

Figure 4: Last recorded time and temp... temp still rising, batt voltage still over 36 volts.

[MileHigh]

TK:

If I may make a few back to basics comments.  I can't remember if in the last NERD report there was a definitive thermal profiling of the heating element so that they could then extrapolate how much wattage negative oscillation mode was dissipating in the element.  This of course implies the same ambient thermal environment for both the negative oscillation mode and the thermal profiling tests.

If the thermal profiling was done, then the RAT team could say something like, "We measured battery power consumption indicating that 2.5 watts of power was being returned to the battery while our thermal profiling showed that 4.5 watts of real power was being dissipated in the load resistor."  (Hence COP infinity because of the returned power.)  I somehow doubt it because it seems to me that every report that I ever read from Rosemary and her various RAT teams never showed a simple table with (negative) power in vs. power out.

Anyway, unless some kind of thermal profiling is done then just quoting an absolute temperature for a heating element doesn't say too much.  i.e.; the same output power could give you 50, 60, 70 degrees. The variable in play is the thermal resistance to the outside world vis-a-vis the power being dissipated.

I know that you are also heating up mineral oil in a thermally isolated container. That's a different way of doing it with a different set of procedures.

MileHigh

[TinselKoala]

TK:

If I may make a few back to basics comments.  I can't remember if in the last NERD report there was a definitive thermal profiling of the heating element so that they could then extrapolate how much wattage negative oscillation mode was dissipating in the element.  This of course implies the same ambient thermal environment for both the negative oscillation mode and the thermal profiling tests.

If the thermal profiling was done, then the RAT team could say something like, "We measured battery power consumption indicating that 2.5 watts of power was being returned to the battery while our thermal profiling showed that 4.5 watts of real power was being dissipated in the load resistor."  (Hence COP infinity because of the returned power.)  I somehow doubt it because it seems to me that every report that I ever read from Rosemary and her various RAT teams never showed a simple table with (negative) power in vs. power out.

Anyway, unless some kind of thermal profiling is done then just quoting an absolute temperature for a heating element doesn't say too much.  i.e.; the same output power could give you 50, 60, 70 degrees. The variable in play is the thermal resistance to the outside world vis-a-vis the power being dissipated.

I know that you are also heating up mineral oil in a thermally isolated container. That's a different way of doing it with a different set of procedures.

MileHigh

They claim to have heated water, at a specific rate of heating which allows for no leakage or extra heat: 4.18 Joules per gram per degree C, the specific heat of water. My insulated container is heating mineral oil which requires 1.67 Joules per gram per degree, and it's insulated so it might not take much more than that in actuality.
You are right, though... doing it properly is different from doing it shoddily, and with a different set of procedures.

Meanwhile, in the second part of the video linked above, here is the proof that they are in fact turning on the Q1 mosfet STRONGLY and WITHOUT OSCILLATIONS to produce the higher heating in the second temperature demonstration. AND they are plainly measuring the temperature of a bare element. Those 190 degrees are Celsius degrees, down there !

Note the FG offset. They have ingeniously put the TOP of the trace right on the scope's center graticle scale, and the oscillations on the gate trace below the center, just like all the other gate signal traces they have shown. But look where the offset and the trace zero level marker is (little blue symbol on the left.) They are using a POSITIVE GOING PULSE in this part of the demonstration, and the Q1 mosfet is carrying the load, if they still have the same hookup happening. No wonder they removed one battery. Without cooling, that mosfet at 60 volts and whatever current it's now carrying, it wouldn't last long. So to reduce the current they pulled one battery, without explanation.

Everything to this point could have been an error, a misinterpretation. But THIS is the real smoking gun. Displaying the trace like this, even though the narrator says "positive offst", borders on deliberate conscious deception. And where is the tell-tale mosfet common drain trace? It reveals TOO MUCH so cannot be shown. These people are consciously lying through their oscilloscopes, and MH is worried about my mineral oil load cell.

ETA: In the second shot, of the LeCroy, I think the drain trace is the green trace, #4. Note the nice saaag in the phase NOT oscillating.. it sags towards the baseline. This indicates the Q1 mosfet is at least partially turning on during the NOT oscillating phase.

[MileHigh]

TK:

Thanks for the pictures, really appreciated to see the setup.

The AC superimposed on the battery voltage is interesting.  Interesting because I think the NERDs always took it literally and used it in their DSO calculations.

Poynt believed that Rosemary had her battery probe in the wrong place and the AC would not be manifested if your probe was right on the battery terminal itself.  Now I am not so convinced of that in looking at your scope traces because I am assuming that you are measuring the battery voltage properly.  Have you tried a direct connection with your scope probe to the negative terminal of bat 1 and a direct connection to the positive terminal of bat 3?

I would not be surprised if you would see the oscillations like that.  What is likely happening is that the Q2 MOSFET array, acting like an RF waveform generator, is simply superimposing AC on everything.   The batteries probably appear like just another load (not sure of the impedance) to the AC source coming from the Q2 array.  I am pretty sure that the RAT team was completely faked out by that superimposed AC and interpreted it as the literal battery voltage for the DSO crunching.

It's pretty obvious that a big fat battery, kind of like a slow electronic sloth, is not going to react to the buzzing fly of the high frequency AC that is superimposed on it's output.

So the moral of the story is the suggestion that if indeed there is high frequency AC literally on the terminal posts of the batteries, that is a fake-out, and you would have to divine the true battery voltage for purposes of DSO number crunching by filtering it out.  It's garbage data that needs to be filtered out.

MileHigh

[TinselKoala]

MH asked, "Have you tried a direct connection with your scope probe to the negative terminal of bat 1 and a direct connection to the positive terminal of bat 3?"

Yes, that is exactly how those battery traces were obtained, with another compensated P6047 10x atten. probe. The FG is not connected to the scope any more, except for the FG's "synch" output to the scope's external trigger input.

The rest, I agree with too. The RF/AC/Whatever is _feedback noise_, takes a tiny bit of energy to sustain, is actually _turned off_ by the FG (this is the main role of the FG it seems) and their large heating comes from the same place TarBaby's does: current through an on-state mosfet during the NON oscillating phase. Garbage in, garbage out. And yet, if those scopes had been used properly, one could indeed have resolved the power issues.

ETA: Wait, are we counting the batteries the same way? I assumed you meant from the most negative pole, to the most positive pole, of the whole stack. That's where I have the probe, across the whole 36 nominal volts.