Testing the TK Tar Baby

[TinselKoala]

Do you think that high-amplitude oscillations on the battery trace are going to be necessary for the NERD Effect to occur? If so... then there is no hope of success with a three-battery test, as you had been discussing with RA earlier.

But their Paper 2, Figure 8 shows oscillations on the battery trace with a battery evidently at 60 volts... and only a 100 volt p-p amplitude. This variable is strongly dependent on the bias current which in the NERD case will be determined by the FG's exact offset setting. At higher drives the waveform becomes increasingly distorted, so it is possible that they turned down the drive to get cleaner looking waveforms in that figure.

[poynt99]

Do you think that high-amplitude oscillations on the battery trace are going to be necessary for the NERD Effect to occur? If so... then there is no hope of success with a three-battery test, as you had been discussing with RA earlier.

No. All you need is to show the two wave form traces as you have, and show their product p(t) as negative, all the while heating the load.

You're nearly there. Once you get use of the Tek scope back and show that negative power, you'll be able to claim your prize too. 

[TinselKoala]

No. All you need is to show the two wave form traces as you have, and show their product p(t) as negative, all the while heating the load.

You're nearly there. Once you get use of the Tek scope back and show that negative power, you'll be able to claim your prize too. 

But wait... is this going to be another story of needing to "replicate" an error in order to prove a point? That is, when I at first used a tight layout with mosfets mounted on very short leads right at the motherboard, and twisted pairs to connect the external components like the load, I got no oscillations, apparently. It was only when I read humbugger's report on OUR, combined with a comment from fuzzy, that I decided to "sloppify" the layout. Then I got the feedback, no problem.

So... if I do the things that I consider proper in making a power measurement (probes deskewed and calibrated, using proper sampling intervals and proper integral forms and constants, minimal inductance in leads, and so forth.... will it even be possible to show that negative power?

I don't recall just now exactly what changes you made in the sim to go from the "negative power calculation" mode to the more correct mode.

As far as I'm concerned the negative power measurement that the NERDs reported is an instrumental artefact, until THEY demonstrate otherwise. It's clear from the scope shots that they didn't do it right, and I have no idea what they did with the spreadsheet data to come up with their cited figures.

But of course I'll try it, when I have a chance.

[MileHigh]

TK and Poynt:

Looking good!

TK, your comments are well understood.

Negative computed power while the load resistor is producing heat followed by a battery dim-bulb test would be 100% conclusive.  No need for the detailed energy audit trail.

TK can then claim another "Mylow Buster" prize!  And of course we know he is the original Mylow buster!  <cue the Star Wars music..... again!>

Now everybody's heard about a NERD!

MileHigh

[poynt99]

TK,

I am fairly certain that if you just take those two traces and multiply them with the Tek math function, you will show a negative VV. Then you're pretty much done. You'll have replicated most if not all the major points raised by Rosemary to equal their claim.