Testing the TK Tar Baby

[picowatt]

All these zeners test good, the magic smoke still inside, and in the mosfets too. I got 4 of the 3.3 volt Zeners, cut out the first one just on spec and replaced it with another new one. No change... it still just doesn't want to oscillate with the 3.3 Zener in there. I suppose it's possible that all 4 of these are bad... but they test the same on my diode checker. The circuit definitely oscillates with the 5.1 V Zener in there... quite a bit lower f0, big spikes, definite ringing but the waveform looks less like it will make the Negative mean power product. The only lowvoltage Zeners I have in stock are the 3.3 and the 5.1, but I found some different 3.3's in the box that I can try from another manufacturer.

The box... Right now I've soldered a 3" cliplead on the circuit's common ground terminal and can use that to clip to the box or not. I've also soldered a 0.01 uF cap to the box itself, so I can clip to nothing, to the cap, or to the side of the box directly. There seems to be an issue with the metal case of the battery. When I have the battery sitting in the box and I clip the circuit common to either the box or the little cap, the waveform gets really complicated and my scope can't trigger cleanly on it. But if I remove the battery from the box and just set it alongside, then clipping the circuit common to the box slightly smooths the ripples and I think gives the nicest looking and most stable waveforms. So there is some issue with coupling to the metal case of the battery. When I built the thing I put a sheet of thin plastic underneath the circuit board itself but didn't think that the battery area would need insulation... so it is too tight in there now to add any easily. But I'll tear it apart and get something lining the battery compartment to insulate the battery case, and maybe that will allow "grounded" operation with stability.

TK,

Rather than insulate the battery, maybe work the problem from a different angle.  Consider decoupling across the battery so that only smooth DC is across the battery terminals.  You could do this with a cap across the terminals where the batt wires are connected to the board, or, if the top of the batt connector can be removed, solder a cap at the batt connector.  Adding the cap at the board will allow a bit more flexibility regarding cap selection, but a single tant will likely suffice.

If osc dies, add more inductance on the board with inductors or coiled wire.

This should allow you to install the batt in the gnd'ed case and provide a probe point for PMP to compare to the NMP.

PW

[ReFried]

TK

Here in the jungles of central america your circuit wouldn't be complete without black electrical tape, somewhere. So, consider a tight wrap on the battery. It should still fit in the casing. The only caps in these parts are worn on peoples heads. The black tape might fit into the TarBaby theme as well. Altoid mini is a real breath of fresh air

RF

[Magluvin]

TK

Here in the jungles of central america your circuit wouldn't be complete without black electrical tape, somewhere. So, consider a tight wrap on the battery. It should still fit in the casing. The only caps in these parts are worn on peoples heads. The black tape might fit into the TarBaby theme as well. Altoid mini is a real breath of fresh air

RF

The battery and the case are probably introducing a capacitance across the circuit.
Like an ignition spark coil, the can of the coil can be used as a capacitive connection to the coil. Like if you use an AV plug to charge a cap for discharge into the spark coils primary, the AV plug is usually sourcing from a 1 ended HV secondary. But if you take that open end of the HV feeding the AV plug, and connect it to the can, the discharge cap wil fill much faster than with the AV plug alone, as the can capacitance completes the HV circuit.

Being you are in the Mhz, the capacitance could be minute and affect things. Taping it up will reduce it, but not eliminate it completely.

Mags

[Magluvin]

To test my theory, insulate the battery case, but solder a short wire to the case first near the terminal end. Now take that wire and attach it to either the + or - of the battery then close the altoids case and see if there is a difference in whether it is connected + or -

Simple to try.

Mags

[ReFried]

Thanks Mags. Your post helped me understand PicoWatt's suggestion better also. I will add this to the list of the many things I have learned connected with this thread.

RF