Testing the TK Tar Baby

[poynt99]

And just to clarify one thing:

You don't need any inductance added to the CSR.

@all: Keep in mind that my schematics (when simulating) include parasitic components, and these components are not meant to be added to your actual build on purpose.

parasitic components that I may use include:

- the series resistance and parallel capacitance of inductors
- the series resistance of CSR's
- the series resistance and inductance of long wires
- the series resistance (ESR) of capacitors
- the internal resistance and inductance of batteries

I try to make it obvious which components are "parasitic", but most of you will (or should) know this already.

[poynt99]

I'm a little surprised that the polarity of the LEDs can't be reversed and have them still light up. Since the circuit works and oscillates with just a 100R in there, the diode action isn't required, I think...

The diode action isn't required, but it does enhance the effect and provide for higher "VBAT" spikes. It also gives a nice visual indication and confirmation of the direction of current flow in the circuit.

The fact that the polarity of the LED can't be reversed shouldn't be surprising, since this simply indicates that the average current through the circuit is from the battery, and not the reverse.

[TinselKoala]

Here's what the traces look like with a single red LED as the load. A lot more amplitude but not as clean waveforms. The Vbatt trace goes negative, which is something I've not seen in Tar Baby yet, that I can recall. This still looks to me like it will give a negative mean power product, but I don't know what that negative batt excursion will do to the final result.

[TinselKoala]

Oh, yes.. the use of the single LED raises the osc frequency to around 3.6 MHz, and it is easily picked up as a strong carrier on my little SW radio at that frequency, and it's pretty narrow band too. There is no modulation of course.... I wonder how hard it would be to modulate at say 440 Hz, by putting a simple interrupting transistor in series with the main loop, like another 2n7000 run off a 555 or even a simpler RC osc config.

[TinselKoala]

Argh. I got confused with probe attens now that I have this new one with the switch. The Vcvr trace in the last scopeshot is at 0.1 V/div, not 1 V/div as I had it at first, since I am using the 1x setting on the probe.
I've corrected the scopeshot. A peak current of 180 mA negative.... that looked bad, and it is bad, due to the atten error. The true peak is about -18 mA, I think.
Sorry....