Testing the TK Tar Baby

[TinselKoala]

For another example, THIS execrable claim has not been corrected or retracted, and it is manifestly false, and since AINSLIE knows how to do power calculations correctly, it is A DELIBERATE LIE to leave this claim uncorrected.

NOW.  Let's look at your 'self-runner' demands.  We have never recharged those batteries - with one exception.  Two caught fire and BOTH were fully recharged.  We've had those batteries since January 2010.  We've been running them since August 2010.  I've now FINALLY checked their rated capacities.  They're 40 ampere hours each.  We've used 6 of them continually since that time.  According to this rating they are each able, theoretically to dissipate 12 volts x 40 amps x 60 seconds x 60 minutes x 1 hour x 6 batteries.  That gives a work potential - a total potential output of 10 368 000 JOULES.

According to what has been carefully established it takes 4.18 Joules to raise 1 gram of water by 1 degree centigrade.  We've taken a little under 900 grams of water to 82 degrees centigrade.  We ran that test for 90 minutes.  Then we upped the frequency and took that water up a further 20 degrees to 104.  We ran that part of the test for 10 minutes.  Ambient was at 16.  Joules = 1 watt per second.  So.  Do the math.  4.18 x 900 grams x (82 - 16) 66 degrees C = 248 292 joules per second x 90 minutes of the test period = 22 342 280 joules.  Then ADD the last 10 minutes where the water was taken to boil and now you have 4.18 x 900 grams x (104 - 16) 88 degrees C = 331 156 joules per second x 10 minutes = 3 310 560 Joules.  Then add those two values 22 342 280 + 3 310 560 = 25.6 Million Joules.  All 5 batteries maximum potential output - available for work - is 10.3 Million Joules. In that test alone the battery outperformed its watt hour rating.  And that was just one test.  Now.  Over the 10 month period that those batteries have been running at various outputs - which, when added to the output on just this one test - then I think its safe to say that the evidence is conclusive.  Those batteries have outperformed. They are still at OVER 12 volts EACH.  They are all of them still FULLY CHARGED.

If your lawyers have trouble with the math, I'm sure that I can recommend someone who can help them out. I've highlighted just some of the problem areas. There are more, to be discovered when you DO THE MATH (tm RA).

[TinselKoala]

Deliberate misrepresentation of the circuit used, or post hoc CYA action? One or the other OR BOTH.

1. Here's the "full disclosure" of the mosfet mixup that .99 asked her for ... this is the only mention on her blog where she announces the discovery of the error in wiring.

2. Followed by the admission of misrepresentation of the true state of affairs with respect to the Case of the Miswired Mosfets.

3. And this was immediately followed by a vicious attack against .99 where she accuses HIM of hiding something...  a clear attempt to divert attention from her amazing attempt to cover up the true state of affairs.

4. And finally, an earlier admission from her, inadvertent I'm sure, that she has misrepresented the data from her trials and has cherry-picked what she reports and has failed to report negative results and the conditions that produced them.

What do you think of THAT, imaginary counselors?

[TinselKoala]

Back on track... I noticed that .99 in the simulator checked on the effect of removing the Q2s one at a time.

My results in the Tar Baby hardware agree with what he found and described last year.

There is one interesting thing: with just one Q2 and the Q1, when I first ramp up the negative gate drive voltage from the elenco ps or another source, there is only a small, different, high freq oscillation that is nothing like the "real thing". Then... just a tiny touch with a finger to the heatsinks and WHHHHHHHH the oscillations emerge, and once they start they continue. That first non-osc stage is stable until perturbed.

I also found the same effect on voltage and frequency that .99 noted in the sim.

In other words, except for a current handling capacity, the single Q2 version worked better than the "gang of four" version.

[TinselKoala]

I've found so much of interest in reading through the old locked threads.

I'm glad I did this discovery process on the hardware myself, "blind" as it were, because my thoughts weren't influenced by the work that had gone before... and I was able to come to the same analysis and basic conclusions, now, as .99 and the other "dogs" as Ainslie calls her detractors, did then.

What is even more amazing is that he seems to have completely analyzed the real circuit in about 10 days or less after the true circuit schematic (the Case of the Mosfet Mistake) was traced out by him and finally acknowledged by Ainslie.

By two weeks after, .99, Humbugger (cHeeseburger), MH and a few others had fully explained everything...even the current path through the Q1 zener, the problem with the CVR location, the need (and the non-need) for a floating bias source.... all of it, from the sim and a few basic hardware tests and comparisons.

And as far as I can see, there really hasn't been much actual progress in any direction since then, over a year ago, in April 2011. Ainslie is no closer to performing actual tests of her central claim now, and in fact her attitude appears to have congealed into some kind of sticky mess that is no longer in the least cooperative. Rather, due to her extreme defensiveness in the face of the threat against her ego integrity posed by the falsification of her experimental hypotheses (ill formed though they are), she has become sealed off from real participatory activity. Now she can only engage in rants, insults, and threats rather than constructive cooperation in the effort to understand her errors. It's a shame, but it's also completely consistent with the general pathology she exhibits, and it illustrates the decompensation that has occurred over the past year -- and that will continue to worsen and interfere with her ability to perform constructive and valid work in the future.

But anyway, be that as it may, here we all are today, and at least my work with Tar Baby has shown that the measurements can be repeated in hardware as well as in .99's sim, that they are in fact easy to reproduce in many different circuit variants, and that fancy and expensive digital equipment is not even needed to show the same evidence that Ainslie claims supports her contentions. 

Meanwhile, Ainslie acknowledges that DC current flowing in the Q1 mosfet, which is turned on by a positive Gate HI signal, is important for the large load heating that occurs when... the Q1 mosfet is turned on by Gate HI.  (See post image below.)

There is no magic possible here, folks.... the Q1 ON situation is no different from straight DC supplied by some slightly resistive wires hooked up directly between the battery and the load.

Short duty cycles? OK fine.... the situation is still STRAIGHT DC during the Gate Hi portions of any duty cycle/period combination and whatever "effect" Ainslie claims must be happening during the oscillations, the Gate LO times. Consider the case with a 15 percent ON duty cycle..... and a 2 minute total period like some of the trials Ainslie has illustrated as being the "desired mode" of operation. For 18 seconds straight during each two minutes, there is simple plain ordinary DC current, and lots of it, flowing through the Q1 mosfet. No battery recharging could be happening, there is nothing preventing the battery from discharging, there is just a battery, wires and a closed "switch" in the fully ON Q1 mosfet.

Whatever the Ainslie effect is, IF it is, it must be happening during the oscillation portion....AND it must be able to affect or cancel out the REAL battery discharging effect of the STRAIGHT DC that is supplied by a simple closed switch during the Gate HI portion of the period. (Getting near the point of the whole SCRN0235 power discussion.)


Note that Ainslie acknowledges that high current through Q1 is necessary for her load heating "efficiencies".

[TinselKoala]

SCRN0150.

This shot shows another high-current trial.... this time with only 48 volts nominal in the battery, so things stay under control. There is substantial DC power shown during the Gate HI signal, and even though the duty cycle is about 16 percent ON (not even including most of the ramp-up to steady state) the mosfet still will heat substantially at the 2.4 Amps it is carrying during that ON time.

Of course all this assumes that the same 0.25R current-viewing resistor stack was used for these trials. Who really knows.

So what magic happens during the oscillations, to prevent the NON oscillating DC current that happens at a DIFFERENT TIME from discharging the batteries? It must be magic indeed if the oscillations are NOT RECHARGING the batteries, as Ainslie acknowledges they are not, but rather are preventing the batteries from discharging  -- during a time when the oscillations aren't even present.

I will include a graphic showing my estimations from the traces for the DC portion of the period.

(ETA: I calculated the frequency based on my rough measurement using calipers against the blowup trace of 675 microseconds for the period and got 1481.481481481 Hz ( a repeating decimal, all digits included here for the Lulz only). Note the frequency given for the gate signal in the parameters box: 1.462 kHz. My eyeball measurement is different from the scope's computation by 19 Hz.... and 19/1462 is a little over ONE PERCENT. )