Testing the TK Tar Baby

[hartiberlin]

Is the 200 mA really a DC current ?
What about the AC component flowing while the oscillations are happening ?

If we would use just a 9 Volts battery with a pot and put on the oscillations,
then we would pronbly only have an AC current, right ?
Or will we have an AC current superimposed on a DC 200 mA current ?

Maybe TK can measure this and show it on a new video ?

Many thanks.

Regards, Stefan.

[picowatt]

Thanks... I think I get it..... maybe..... So the bias power is "mixed in " with the main power from the batteries and must be dissipated in the load and mosfets (I suspect in the mosfets mostly). Hence the load heating and current viewing resistors are giving an inaccurate picture of the power flows in the circuit.... is that right?
(I know this has been covered before; I just want to hammer it home somehow if it is correct.)

I'm running so late that there's no point in making the drive. So here I am, stuck in Mobile with the Memphis blues again.

Anyway.... I tried the 50r/pot/cap/main battery thing and I can't get it to work.  And yes, I've confirmed that the pot is good and is adjusting the voltage properly, tried with three different valued pots.

This _does_ work just fine with the external battery (no decoupling cap), but using the running battery I can't get it to work. External 9v gives beautiful nearly perfect sinusoidal oscillations on the common drain when the 10 k (not 1 meg; I also tried 100 K, no luck) pot is set right, and the "voltage floor" still shows up: the negative voltage on the gate drive signal floors at about 4 volts negative indicated on the scope trace.

I think you really do need to provide a potential that is more negative than the negative battery terminal....

I just made another video... chaotic but the points come through I hope. It's uploading now.

TK,

I would decouple the wiper of the pot, or just use a 10K resistor from the first Bat+ to a cap and tie the other end of the cap to real ground (not the CSR, althogh you might want to try both ground points).  Then, to the junction of the resistor and cap, connect another resistor, 1Meg or so, and connect the other end of that high value resistor to the MOSFET gate.  You may have to add a bit of wire length between the end of the 1Meg and the gate to add some inductance to get the osc.  Possibly you will have to add some inductance (wire) in series with the 50R at the source as well.

PW

[picowatt]

Is the 200 mA really a DC current ?
What about the AC component flowing while the oscillations are happening ?

If we would use just a 9 Volts battery with a pot and put on the oscillations,
then we would pronbly only have an AC current, right ?
Or will we have an AC current superimposed on a DC 200 mA current ?

Maybe TK can measure this and show it on a new video ?

Many thanks.

Regards, Stefan.

Stefan,

If a 9 volt battery was placed in the source leg of Q2 with a series resistor, let's say 50R, then the DC bias current would be around 100ma plus or minus 30-40ma or so.  When oscillating, there would be this DC bias current passing through the 50R and 9volt battery and a separate AC current.  My assertion is that the AC current would follow two paths, one being the through 9 volt battery and 50R in this instance, and the other AC path being the gate to source capacitance of the MOSFETs.

PW

[poynt99]

I don't think we want 200mA of AC current flowing through our bias battery, that's why I isolated it.

With isolation, the battery supplies only a tiny DC current to slightly bias the MOSFET ON, and the AC path is provided capacitively, which bypasses the DC bias network.

[picowatt]

I don't think we want 200mA of AC current flowing through our bias battery, that's why I isolated it.

With isolation, the battery supplies only a tiny DC current to slightly bias the MOSFET ON, and the AC path is provided capacitively, which bypasses the DC bias network.

.99,

Yes, but in TK's set up using the 555, and in the NERD circuit using the FG, the 555 or FG must provide that 200ma of current at whatever negative voltage is used to keep Q2 biased on.  Only by by applying the bias voltage to the gate can a "tiny DC current" be needed from the bias source.

Also, I am only estimating the Q2 Ibias.  In the RA paper it was mentioned the FG was set to full negative offset in at least one test.  I am assuming a maximum of -14.5 from the FG open circuit and an Rgen=50R.  The voltage at the source of Q2 will always be Vgs from the gate, which is ground in this case.  At -14.5V open circut at the FG, then there would be around 10V across the internal Rgen of 50R.  10V divided by the 50R is 200ma.  This will vary a bit due to Vgs characteristics and device to device differences, so I believe I said some time ago I would "guess" Ibias to be around 150ma plus or minus 30ma or so.  Likely the circuit could be "played with" to get it to osc at a lower bias current.  Without knowing the open circuit voltage versus in circuit voltage of RA's FG, we cannot claculate Q2's Ibias.

PW