Testing the TK Tar Baby

[TinselKoala]

TK,

Regarding the alternate bias using the first batt tap (+12v), you will have to either remove Q1 or tie its gate to it's source and Q2's gate will have to be isolated from the rest of the circuit.  I.e., pull Q1 out and then hook up the 12V via 100K or better to the Q2 gate.  The 50R can go across the current FG terminals (between Q2 source and the CSR).  That should give you similar DC bias conditions for Q2, but as I said, at AC things get a bit muddied.

OK that makes sense, and I'll try it that way after dinner.

Do you have a rough idea as to the frequency of the osc? 

A rough idea, yes. The scope says that there are nine full cycles PER 2 microseconds. Doing the math (tm RA), we find 9/0.000002 is equal to 4.5 MHz and the unit dimensions agree. (sorry, I couldn't resist).  But the Philips PM6676 counter reports a period of 254 nanoseconds or a frequency of 3923 kHz, give or take. Maybe my scope's timebase is a bit off at that freq, although it checks out with its own calibrator at 10 kHz. It's a miracle that it can resolve the fast oscs at all. I'm using the delayed trigger function to get them displayed stably. I trust the Philips implicitly.

I am not so sure the LED's in the video were lighting due to AC current through the CSR in the "traditional" sense.  Assuming 1.4V turn on for the LED's, that would require about 5.4 amps going thru the CSR, yet you say the load barely warms during the osc.

What happens to the LED's if you touch them with your fingers while the circuit is oscillating?

Touching it pretty much anywhere, or waving hands around, affects the oscs and can kill or start them. No oscs, no bidirectional current flow in the LEDs.

Have you looked at the CSR with a scope channel during the osc?  Does it give you any clues as to the actual voltage/current happening there?

PW 

Yes, and it's about what you'd expect to see on a HP180 scope at a gain of .5 v/div trying to get information from a 9 (or 4.5)  mHz signal. I'd hate to have to interpret it in a court of law.

But never fear, the Clarke-Hess power analyzer is near.

Wait till you see what 18 inches of wire inductance can do, in the latest video.

[MileHigh]

TK:

I knew it was high in frequency and I made a P.S. in my last posting about that.  I assume the RAT test was comparable in frequency.  It's not like you are in waveguide territory bit it's still quite high just the same - and not trivial.

Not easy stuff to deal with at all in terms of a serious analysis of an energy audit trail.  Like I already said, you have to assume way beyond the competencies of the RAT team.

You have probably looked at your share of "Dr. Stiffler" related clips and they are all just "playing" clips with people making neon lights light up and "sniffing around."

MileHigh

[Magluvin]

I agree that when pulsed, an led can take quite a bit more than continuous on current.
Remember the conversation we had a bit ago? There is an initial high output flash if you overdrive them. Say we put 10a through a laser diode that continuously wont handle as much, but for a very short amount of time. There will be an amount of time that it can take it.   After the flash dies, and current still flows, the led/laser diode begins to really die. err fry.

Just like diodes have peek amp/time ratings and continuous.

I have to find the laser circuit with the explanation. I was able to get leds to do it also. Not 10a though. ;]

If I can get it going again, with the led chip image on white paper, Ill make a vid.

Its pretty cool to see the 4 sides of the chip flash brighter 1 at a time as the current is increased till they are all brighter.

If this circuit is 1mhz or more, this could very well be a condition that the led will pass more current safely.


Instead of leds, what about shunted analog current meters with diodes on each. If the freq are very high,and rectified, the needle movement shouldnt have any visible vibration. Some meters are quicker and less damped. But if mhz, it should just show the average.
The meter that shows the most current, determines which direction more current is flowing. 

Mags

[TinselKoala]

I think the NERD oscs in the demo video were supposed to be at 1.2 MHz if I recall the presenter's narration correctly. Of course, I trust that figure absolutely. Don't you?

But it does seem plausible considering the larger capacitances involved with the PG50. I have a single one I can stick in in 2 different places, either as Q1 or as one of the "gang of four", and I predict that will lower the oscillation frequency but of course not all the way down to 1.2 MHz. We shall see. Tomorrow I may get the other PG50s I've ordered.

By the way, the whine heard on my stereo during the "just right oscs" is the "modulation frequency", that is, the FG's pulse freq or a harmonic of it, while the 4.5 MHz seems to be the "carrier".

[TinselKoala]

@Mags: You know I'd be using my old analog Simpson for current monitoring if I could. Unfortunately it's in another undisclosed location right now and I can't get to it. Nor a lot of my other gear either.

I do have a current sensing transformer of the Rogowski coil type that I can try, running several turns of the supply lead through it to increase its sensitivity. It will only respond to the AC component of the current and will indicate the magnitude of the current by the voltage trace of its output, and it's only connected magnetically to the circuit under test. But the one I have might not be very good at 4.5 MHz... we shall see.