Testing the TK Tar Baby

[poynt99]

Here are the results when all the batteries are replaced with a single 8uF capacitor charged to the equivalent 72V.

[poynt99]

Here from the same test with the single 8uF capacitor source, are the results showing the instantaneous and average power delivered by the capacitor, along with the capacitor voltage over the 3ms time period shown.

Keep in mind that the "capacitor voltage" measurements for the power computation are taken from the same measurement point as always, namely at the high side of the load resistor, NOT directly across the capacitor itself. The red capacitor voltage trace IS directly across the capacitor in each case.

[MileHigh]

Poynt:

Thank you very much for those simulations.

I will make a few comments for information purposes.

You note that the capacitance values used are quite small and the time frames are short.  If you used a much larger capacitor then the same observed behaviour would take place but over a much longer time frame.  Poynt wisely used a small capacitor so that you could get a snapshot of the entire discharge process.

In the first test where a single capacitor replaces one of the batteries you can see how the voltage will reverse on the capacitor.  I actually warned about that a few times when I posted the suggested test.

You notice in the last test run that Poynt is showing the instantaneous power supplied by the capacitor.  You can see how the instantaneous power is negative.  Rosemary has had a few freakouts over the concept of negative power but it's all perfectly normal.  All that the negative instantaneous power means is that power is being exported out of the capacitor to the outside world - the capacitor is discharging.  Positive power relative to the capacitor means that it is charging.

There are what appear to be "gaps" or "missing teeth" in some of the waveforms.  In all likelihood that is because of aliasing effects.  That means that the waveform is still there but when you compress the resolution down to display it on the screen the waveform disappears because of the compression.  There is a possibility that the waveforms actually do disappear or get reduced in amplitude also.  This could be because of the way the circuit behaves at certain lower voltage ranges.

Note that Poynt got his simulation running with some tweaking so that it was a very close approximation of the waveforms that the NERDs captured with the DSO.  The capacitor test simply confirms what we have all known all along - that the batteries are discharging.

MileHigh

[poynt99]

MH,

There are no significant aliasing effects there. The "missing" sections are simply showing the circuit beginning to "sputter" due to the lower supply voltage.

The negative power computation is a result of the voltage measurement point. If the voltage measurement used for the power computation was taken directly across the capacitor (as it should be), the resulting power would be positive, as shown in my detailed analysis and various other times subsequent to that.

[MileHigh]

Poynt:

Thanks for the clarification on the power.  I thought that you were using a power probe, but I now see that you are using a voltage probe at the load times the CVR current.  I should read next time!

So for the blue negative instantaneous power trace, do negative values imply power being returned to the capacitor?  In other words your simulation is so good it even shows how you can be "faked out" and made to believe that power is being returned to the battery/capacitor when in fact the decreasing capacitor voltage is clearly showing that power is being output by the battery/capacitor.

MileHigh