Testing the TK Tar Baby

[picowatt]

TK&MH,

What do you think, wire under compass perpendicular to magnetic north?  That way Earth's pull would would put the needle perpendicular to the wire and E-W would be Idirection.  Earth can be the return to null "spring".

Also, if necessary, a few turns of sufficiently sized for the current magnet wire over and under the face of the compass for more sensitivity?

TK, you would have to empty your pockets of all those magnets you carry around.  Possibly a low iron diet as well?

PW

[TinselKoala]

You do realise that I have a Clarke-Hess 2330 power analyzer sitting here on the shelf, right? I'm really not sure what a compass needle will tell me that the 2330 won't. It will tell me the power in and out, and it won't be fooled by complex stuff like HF AC sitting on top of DC. At least I don't think it will.

Besides, all those "nice compasses" are two thousand miles away from where I am right now, and I'm at the limit of my expenditures for this little time-waster. I think I have one or two good navigational compasses here, but except for the visual effect I can't see them being any different than a moving coil meter and they'd be less sensitive too. I'm not rejecting the idea, I'm just not sure what the payoff might be. I mean, we have several good instrumental measures of what's happening, and when I scope the CVR with a proper connection to a fast DSO, that will show what's happening well enough, won't it?

I am even getting PMs from True Believers telling me, over and over, that I am wasting my time doing this, and that I should be doing Real Research (tm), like looking at the Steven Marks hoaxes or Bedini motors or real stuff like gravity wheels and such like that.

Regardless, the LEDs have told us what we were looking for, have given us a point of comparison and have scared the knickers off the RATS. But perhaps they interfere in the battery charging effect, because  my batteries certainly do run down at this point. So I've removed all the LEDs, and I've also removed my "internal load" on the board, since it's no longer in use either.  So the only "extra" component is the 0.33 ohm  resistor that I can use, or not use, in series with the gate signal, as was shown on one or the other of the diagrams she's claimed to use. And instead of 4 x 1 ohm resistors parallelled as the CVR with one extra mystery resistor, I am using 2 x 0.51 ohm in parallel. Other than that, and using sockets instead of clipleads and threaded rods for my IRF830a mosfets... and the mosfets themselves.... it's getting hard to tell the difference between Tar Baby and the Nerd Rat device.

Of course, I'm still not using white pegboard either..... which proves that I'm not doing an exact replication, so of course none of my results can be applied to the analysis of the NERD RAT device.

My idea about using the Zeners for the DimBulb test was simple and naive... won't it work? At 11.5 volts, the batteries will be nearly depleted. I don't have the exact discharge curve available but these are standard sealed lead-acid. So I thought simply to put a reverse-biased 11.5 volt, 5 Watt zener in series with the bulb. I don't care about the absolute brightness of the bulb, all I care about is which battery reaches the Zener voltage cutoff first. I've never tried this and I don't know if I'll get a clean cutoff with just a simple series zener or not, in fact I doubt it, but would it be suitable for the test to do it that way? One could certainly design a slightly more complex circuit and have a zener switch a transistor or something like that....

Or I can just monitor the voltage with an Arduino and have an alarm wake me up when it approaches a threshold, I suppose.

PG50s are supposed to be in yesterday, but weren't. Maybe today....

And I do have on hand some op amps, 741s and TL082s and some others, and I have a bunch of 4000 and 74 series chips stashed in metal cans in anticipation of the Great EMP Event, and I know there are some flipflops, inverters, and Schmitt triggers in there.

[TinselKoala]

Oh yes... operating mode.

It seems that I will either be able to heat the load strongly, OR have a unipolar driving pulse with only Q2 active at all and minimal load heating. Both modes produce the oscillations, of course, with the only apparent difference being where the baseline is of the drain signal during the NON_oscillating phase of the signal. And of course the DC current draw is different, which reads about 100 - 150 mA with negative gate drive only, or 1.5 amps or more during bipolar driving with Q1 carrying most or all of the heavier load.

With the 1.5 amp drain, heating the load strongly, a four-hour run time should pretty much flatten my 5 A-H batts, right? But with only 100 mA drain, I'd have to go a day and a half or two days for the same charge depletion, right?

Seriously, I'm asking what would be the expected runtime on the TarBaby, letting it charge or discharge as it will,  to prepare the batteries for the DimBulb test, in the two operating modes.


I'd really really like to see  some _good evidence_ that the NERD RAT device can produce substantial load heating from a negative-going gate drive pulse and oscillatory waveforms and low load currents as shown in their scope shot. Did they talk about load heating at all in the video(s) or present evidence that the load was hot during the demos?

(I sure don't remember any dramatic moments when a cup of steaming boiling hot tea is poured out of the load chamber, live on camera. Either they have no sense of theater, or they had no hot water.)

I mean, when my load heats up I can at least show it doing so on a thermometer reading on a video, and I can illustrate the circuit electrical parameters when it is doing so. And I am so short I can barely reach my tools !

(It might be interesting to put the DC ammeter DMM in series with the Function Generator, and set it for very slow pulses so the DMM can get a decent sample.)

[TinselKoala]

A brief summary of results so far:

With the tight layout on the perfboard, using 2n7000 mosfets and a heavier inductive load at low voltage, the oscillations were easy to make all over the place, but not very informative. This testing did show that I at least didn't screw up the wiring and allowed me to get some idea of orientation of the circuit and its dynamics. Here I showed that using a series capacitor would still switch the mosfets but killed the oscillations. I smoked 2 ea. 2n7000s during this phase.

With the tight layout and IRF830a mosfets I couldn't make the "real" oscillations consistently at full amplitude and fully across the cycle. I thought I had identified some feedback oscs in there but I wasn't seeing the massive noise that seem to be the genuine article. I could switch the mosfets fairly cleanly and get lots of load heating even with the heavy inductive load. Some nice spikes, but evidently those aren't needed now. I managed to blow one mosfet from stupidity (wire confusion on the bench) and two more in an overheat runaway event.

Then I read the humbugger work on the other forum. This, plus looking at the NERD video again, made me realize that it was the layout, not the mosfets themselves, that was likely responsible for the oscs.... so I cut random wires and soldered them onto my mosfets, and placed them on some larger heatsinks, not so much for heat transfer but for capacitance. Bingo et voila! Massive Robust Feedback caused by inductance and capacitive coupling in the leads. Sensitive to motion and exact placement.
I tried a simple optoisolator arrangement to get the FG out of the current path, which worked very well as far as switching goes but damped the oscillations. I still think a modification of this idea might work. I fried 2 ea. H11D1 optoisolators during these experiments, one by blowing its input LED and the other by somehow blowing the phototransistor.

Lots of scoposcopy determined the correlation between which mosfets were on and what the waveform looks like; the presence of the AC component; the frequency of the oscillations; the effect of replacing one 830 with a PG50 as either Q1 or one of the Q2 stack; the effect of varying offset and bipolar pulsing on mosfet switching, current and load heating; the LEDs of Doom test... Let the smoke out of one green LED just for fun to rattle MH's cage....
(But also I hope I was able to illustrate a little bit about the quantitative use of an oscilloscope, for the Watchers from Beyond..)

Load heating using the water heater load was inefficient in my setup because of all the metal that wasn't submerged; the 10.3 ohm, 60-watt parallel wirewound ceramic resistor load can be totally submerged in the mineral oil and therefore works much better as a heater at the same power levels in my system than the commercial product does. The little inductive transformer choke also worked well as a heater until one winding went open when it got too hot, but it was lousy for oscs. I wondered why then.... (but not any more....) since at one time I thought I recalled the NERDs specifying a 1.5 Henry inductance. (A meter reading without sanity checking, I think that might have been.)

Load heating worked just fine using a bipolar pulse to turn on at least one mosfet fully during the pulse cycle, during the antiphase from the oscillations. By overdriving the gate I could produce oscillations in both phases but this reduced the current, as the oscillating mosfets do not turn on fully. With unipolar drive pulses going strictly negative with no positive baseline offset, very little load heating happens and the device current is low, since Q1 is off (and can even be removed entirely) and the Q2 mosfets are only oscillating, not switching.

During the testing, my batteries have discharged. I have no control comparison so I can't tell if they discharged faster or more slowly than "normal" for the kinds of drains I have been putting on them, and during these pilot experiments I haven't been monitoring their state of charge, other than by looking occasionally at their no-load voltage. When they dropped below 11.5 volts yesterday evening I stopped testing and started recharging, so they will all be nice and fresh when they meet Mr. Clarke-Hess for luncheon.

Still pending while I've still got the 830s in place are the Capacitor Battery test, in and out power measurements with the Clarke-Hess, some explorations with compasses, and one or two other things that I want to try like the DMM inline with the FG. Then, assuming my other PG50s haven't been confiscated by the Forces of Suppression, aka the MIB mob, I'll repeat everything with them in the slots. I'm not going to do any drawdown or Dim Bulb tests until I have the PG50s and have run with them for a while, though.

So the score is: 2 minimosfets, 3 little brother mosfets, one choke winding, two optoisolators and a green LED. Not bad for a week's work !

[conradelektro]

@ TinselKoala: 

Your careful experiments demonstrate how difficult it is to understand what a weird circuit is doing, not to speak of making meaningful measurements. (That explains the many misconceptions put forward by overunity experimenters.)

Question: From the discussion just above this post I gather that you try to measure tiny deviations from the expected energy output or input of your big batteries (a tiny little bit less flows into the batteries than comes out to run the circuit)? But the claim of the "inventors" of this weird circuit is that the effect is massive. So, even a crude measurement should proof or refute it?

Assumption 1: You want to show that the weird circuit consumes a tiny bit more than it feeds back to the batteries (as common theory would suggest)? Which is difficult, because the power consumtion of the circuit is very small in comparison to the huge capacity of the batteries?

Assumption 2: You want to show that the transistors are not even switching and that only some low power high frequency AC current (in fact a malfunction or misoperation of the circuit) is flowing through the circuit in an unexpected way?

Assumption 3: You want to show that only during a neglected and short adjustment phase (when the transistors are indeed switching) some really strong current is flowing through the circuit doing the claimed heating?

Sorry for the stupid questions, for a layman it is easy to loose track of your objectives. The subject matter is rather complicated (and you are doing a good job to dissect it in order to show the real issues at hand.)

Conrad