Testing the TK Tar Baby

[poynt99]

To clarify somewhat the issue of the simulation result with 4 Ohms vs. 50 Ohms as the FG output resistance, I ran the two scenarios for comparison, and the difference is the following:

4 Ohms: peak to peak oscillation is from about 10V to 220V.

50 Ohms: peak to peak oscillation is from about 40V to 120V.

Otherwise the wave forms are pretty much the same. The 50 Ohm scenario has a much more gradual burst envelope, as may be expected.

[TinselKoala]

And I imagine the negative mean power product still appears as usual.

Does your sim model mosfet types individually or does it have just a generic mosfet model that you tweak with the exact part parameters? I'm wondering how hard it would be to run other mosfets and see the negative mean power for comparison purposes. The 2n7000 for example might make a mini-Tar Baby that could be built down into an Altoids box, with actual inductors instead of wire length, and thus be a pocket-sized, on demand, negative power demonstrator. Just pop a nine-volt battery in there, hook it to a soundcard scope program in your laptop, and amaze all your friends at Tupperware parties and church ice cream socials.
I'm serious, how hard is it to compare mosfets in the sim, all the way to the power product and determine the tuned inductance values to make it negative?

Also...   is it possible to trick the simulation into modelling the FG as having "infinite" output resistance, since that apparently is what the IsoTech "GFG324" has?

What kind of oscillations do you get if the FG somehow only applies a "voltage" with its pulsations or signal but cannot pass measurable current, or provide or source it, or however one puts it?

[poynt99]

And I imagine the negative mean power product still appears as usual.

Yes, but the negative power with the 50 Ohms is about 1/10th the power compared to using the 4 Ohm.

Does your sim model mosfet types individually or does it have just a generic mosfet model that you tweak with the exact part parameters?

There are specific models for specific type numbers. I believe there is also a generic model.

I'm wondering how hard it would be to run other mosfets and see the negative mean power for comparison purposes. The 2n7000 for example might make a mini-Tar Baby that could be built down into an Altoids box, with actual inductors instead of wire length, and thus be a pocket-sized, on demand, negative power demonstrator. Just pop a nine-volt battery in there, hook it to a soundcard scope program in your laptop, and amaze all your friends at Tupperware parties and church ice cream socials.

This would be relatively easy to do. I'm sure we could design a circuit that is compact and low power.

I'm serious, how hard is it to compare mosfets in the sim, all the way to the power product and determine the tuned inductance values to make it negative?

Not all that difficult.

Also...   is it possible to trick the simulation into modelling the FG as having "infinite" output resistance, since that apparently is what the IsoTech "GFG324" has?

I can place any resistance you want, except an infinite resistance. How about 100G Ohm?

What kind of oscillations do you get if the FG somehow only applies a "voltage" with its pulsations or signal but cannot pass measurable current, or provide or source it, or however one puts it?

I have not tried it, but I suspect either it will take a very long time for the DC bias to kick in, in which case you'd either have to go full osc mode, or use a very long period, or it will not oscillate at all.

[hartiberlin]

Hi Rosemary.
YOui don´t need expensive equipment to show if the the basic principle
works or not.

Just get a few 9 Volts NiMH rechargable batteries
a used scope from EBay and a few cheap digital multimeters and a few caps.

It will probably cost you less than 200 US$ including the scope...

You just need only to adjust the scope heads for calibration, but that you can do on your own
or let it do somebody from your team who has more electronic knowledge than you.

I guess you should take a break now and concentrate on your measurements with your team.

From your last postings I have seen, that you do not understand, what TK has posted in his
videos and I think you miss just the basic electronic knowledge to follow and understand it.

So just leave him alone and better concentrate on your own measurements.

And as I said, it is only important to show the basic principle which you should
be able also to demonstrate with small cheap batteries, if your design is not flawed
due to measurements errors....

As I said, you don´t need expensive equipment and maybe somebody else in your group
can educate you some more on the right measurement stuff.

Or just again watch all the 20 latest TK videos again and again and
see for yourself if you can make it better in your own circuit.

I don´t want to suppress any technology, but if it is just based on measurement errors
and there is no overunity with your circuit, then it also makes no sense to continue this
whole scenario.
Then you just have an unefficient underunity water heater and the batteries discharge
when using it... nothing more....

So show us, that we are wrong, but just quit ranting about TK´s circuit that you don´t seem to understand....

Regards, Stefan.

[poynt99]

OK TK,

I took up your challenge; I've come up with a circuit that operates on 9V (7V to 11V) and uses the 2N7000 MOSFET. Not only that, but I've "improved" on the obtained negative mean power that is computed (when using the LED vs. the 100 Ohm), by a factor of about 4x. 

Parts list:

1) 2N7000 (1)
2) 1N4007 (1)
3) 3.3uH inductor (4)
4) 1uF film cap (1)
5) 9V battery (1)
6) 1 Ohm CSR (1)
7) LED or 100 Ohm (1)
8] 30k resistor (10k to 50k) (1)
9) 1N5226 3.3V zener (1)

Schematic and scopes coming...