Testing the TK Tar Baby

[MileHigh]

Before I post some more I just want to give Itsu a plug, because he makes great clips.

Check out his fantastic clip where he compares the performance of three different multimeters for measuring a sine wave at different frequencies.  You can see how the multimeters respond as the frequency increases.

DMM's AC RMS test.mpeg http://www.youtube.com/watch?v=iHcl-85baHs&feature=plcp

[MileHigh]

TK:

I am just going to expand on the theme that I started.  Please anyone correct me if I am making a mistake or comment.

Here is a simple model for the instant when the MOSFET switches off.  Note that I am saying a generic "MOSFET" and not Q1 or the Q2 array.  That's because it doesn't matter, the concept will apply in either case.

You model the wire loop as two separate distributed inductances.  The "top" inductance goes from the battery positive terminal to the MOSFET drain.  The "bottom" inductance goes from the battery negative terminal to the MOSFET source.  You can split them like this because the MOSFET is in the process of switching off and becoming an open circuit.

The top inductance is anchored to the positive potential of the battery positive terminal.  The bottom inductance is anchored to the ground potential of the battery negative terminal.

Both of these inductances become "EMF whips" when the MOSFET shuts off.  The further you travel along the length of the whip (keeping the scope channel ground clip on the battery negative (or positive) terminal), the higher the measurable EMF you observe in the whip.

The top inductance "whips up" in EMF as the MOSFET shuts off.

The bottom inductance "whips down" in EMF as the MOSFET shuts off.

So it's likely that the top inductance could explain the apparent spike up in battery voltage.  When the top inductance is discharging it's stored energy it is charging the MOSFET drain-source capacitance.  So that could be reflected back at the battery terminal as an extra EMF source on top of the battery EMF for a brief amount of time.  That would have to be investigated further.

When the bottom inductance whips down in potential, we see the fake negative voltage across the CVR.   That implies that as you move your scope probe counter-clockwise along the length of the wire (again,keeping the scope channel ground clip on the battery negative terminal), you see a higher and higher negative voltage.

When you look at the scope shot from Itsu's clip, you can see a single definite negative spike when the MOSFETs in his circuit switch off.  In his case you have the luxury of operating at a low frequency so you can isolate the single separate event of the MOSFETs switching off.   For the NERD circuit you are running at a much much higher frequency so you don't have that luxury.  Everything seems to blend into a full regular waveform.  However, the same effects are still taking place.  The apparent negative current is simply the "EMF whip" effect taking place in the bottom inductance.

MileHigh

[MileHigh]

A picture is worth a thousand words:

[TinselKoala]

@MileHigh:
I believe your analysis is largely correct. However, the current in the CVR itself is real and is reversing and can have a real negative-going component; this is shown by the various LED tests, where it is possible to get an LED glowing in either orientation depending on where inductances are placed, as well as by the instrumental measurements. Whether the currents are caused by probe pickup, inductive shunts, or zipons, they are indicated on the scope and so they are in some sense "real", I will at least agree with the head NERD on that point. In the total circuit, though, the power flows from the battery and is dissipated in the load and the other circuit components; some of it sloshes around and depending on where your measurements are taken you can watch it doing all kinds of crazy stuff, like differentiating spikes, bouncing energy from electric to magnetic fields and back again, or making incredibly high powers for really short times. These isolated measurements can be called artefacts or data, depending on what you are trying to prove. The real issue, however, is easily answered and no oscilloscope is required to answer it.
I believe that cHeeseburger/humbugger actually illustrated what you are describing, using his simulator and showing how a simple circuit can respond to the rate of change of a voltage that never reverses, indicating negative currents when there really are none. I can't find it now but I think it's in the old, locked thread, from about a year ago, or perhaps on OUR.  The voltages shown are real, detected voltages; ascribing them to a real, significant change in current direction is where the mistake is made. In the case of Altoid, Tar Baby, and NERD, the matter is the same: if you make the measurements as shown you will get the voltages shown, and if you interpret them as showing real currents, you will be tempted to make some conclusions. And if you install components that also perform the same differentiation (LEDs and inductors) you will get the same results in hardware: a current that reverses direction _in that part of the circuit_. Oddly enough.... this current and the power it dissipates in a load are coming from the only power source(s) in the circuit: the main battery and the bias source.
And of course all our work here is bullshit, our equipment junk and our procedures worse than amateurish ....except when it appears on casual inspection that the work supports the Ainslie conjectures. Then the praise is voluminous, the equipment excellent and the procedures golden and bulletproof -- and the praise is still wrong, because it fails properly to assign credit where it is due.

But of course..... Altoid's battery still discharges. Doesn't it?

[MileHigh]

TK:

Thanks for your comments.  They are appreciated and in taking them into account it's indeed possible that what I am describing is taking place but it's only part of the big picture.  Some people might not be aware that multiple effects can be happening at the same time, and they all get added together. It also sounds about right what you said about Humbugger.  My memory is fading for what's taken place in this long drawn-out drama.

Please keep that current pickup transformer test in mind.  It will not be faked out my any inductive EMF effects and only responds to the current flow.  You just have to look at the current pickup transformer waveform and do the integration in your head to derive the real current flow.  I would not be surprised if you don't see that second positive spike like I mentioned.  It's a poor man's real current probe.

MileHigh