Testing the TK Tar Baby

[picowatt]

Yup... I should be able to do the comparison between one battery decoupled with caps and one battery naked.

So this will be done with and without decoupling caps right at the battery, but the battery probe at the board itself?
If I put the battery probe on the battery, then I'll have the issue with probe grounds at different points; if I put the battery probe at the power entry points at the board, then I'll have those long battery leads....
Well, I guess I'll just do it everywhichaway and see what happens.

At this point it looks like the logistics are good to go, anyhow... I mean my car is back together and except for my housemates being in the hospital... all is good for tomorrow's play date.

Meanwhile here's the video of what we've been discussing:

http://www.youtube.com/watch?v=EPESB5-5iTA

TK,

Have you decided not to try the test with the entire battery string decoupled and added Batt+ wire?

PW

[picowatt]

So across points 1 & 2, correct?

.99,

Yes, decouple across 1&2 and lump all the battery string L together by increasing one of the L's in the batt+ lead wire to the right of 1.

Then do what you do... Using 1 as Vbatt and then comparing results using the top of Rload as Vbatt for your pwr calcs.

PY

[MileHigh]

Poynt:

I just noticed those new simulations and the how the addition of the filtering switches the battery power calculation over.  Very impressive!

TK:

I just saw your clip demoing the CVR voltage change and I will share some thoughts with you.

Let's assume that wire impedance at the operating frequency + harmonics and other capacitive/inductive effects mean that the whole setup is like a slinky festival in a mosh pit.

So, you really need to define a "home base."  Let's call the negative terminal of the lowest battery in the totem pole your home base.  I would move your CVR off of the circuit board and solder it to the negative battery post.  This is not going to change the replication and Rosie has no argument about this not being a replication because of this.

So in theory your returning current will be viewed between the negative battery post and the CVR.  It's unlikely that you will see dramatic changes in the voltage across the CVR with this setup.

Now here is the key:  Needless to say your second channel will use the same negative battery post for the probe ground.  Now your second channel signal probe can "go off on an adventure" to see what the rest of the circuit looks like relative to "home base," the negative battery terminal.   Even if you have to add a longer length to the ground clip lead to make measurements that are far away from the battery post, stick with the "home base" philosophy.

So you can imagine "snaking" your second channel probe through the rest of the circuit, where you look at the potentials every three or four inches of wire length.  You will probably find that your whole circuit board and load resistor setup is "bouncing up and down" relative to your home base.

Once you get a feel for how the CVR voltage is behaving then you can disconnect that channel from the CVR and "send that scope probe out on an adventure" also.  So you can have two channels on your scope tied to your "home base" master reference ground point while you examine the "bouncing" going on at different points in the circuit.

Notice that if you line up the ground references for each channel of your scope, and have the gain for each scope channel the same, you will get a live visual display of the differential potential with respect to time between any two points in the circuit.

With enough interactive probing of the whole setup, you should be able to get a feel for how the major parts of the circuit are fluctuating in potential with respect to home base, i.e.; the slinky fest in a mosh pit.

Also note that if you do this measurement setup it makes adding capacitive decoupling to the bottom-most battery quite simple and elegant.  You are just a step away from measuring the current and the capacitively-decoupled potential across that single battery in order to measure how much power is being output by that single battery.

MileHigh

P.S.:  It may be worth it to strip off some wire insulation as needed in order to see what is going on as your probes snake their way away from home base.

[fuzzytomcat]

Hey TK,

As always your doing a nice job 

I'm curious what would happen if you leave your oscilloscope(s) grounded from the 120 volt utility power connection and isolate the 120 volt utility power utility ground connection from your function generator by using a old style receptacle ground adapter.  I'm not sure of a huge testing equipment safety issue with any large peak to peak voltages to harm anything which all seem low.

This might eliminate your testing equipment ground loops if any exist. I do remember somewhere in all the hundreds of NERD RAT postings that Rosemary did , she had removed the plug ground tab on some piece of 240 volt European testing equipment ... I'm still looking.

FTC

[MileHigh]

Just one question for the true oscilloscope experts here:

One thing I was never certain about comes into play for my previous post.  The question is do you need to use the ground leads on the probes themselves or can you just make a robust connection between the common scope ground terminal (which may be on the face of the scope) and the negative battery terminal?  That way you can remove the scope channel ground leads.

This may tie into the scope probe compensation.  At the frequencies in question at least you are still not really in transmission line territory for the shielded probe cables.

Honestly it's just one of those things that I have never been sure of, I must have used scope probes with the ground leads and with a common ground connection and no ground leads.

Perhaps the experts would comment.   If you can get away with using just the common ground connection it certainly would make it easier to probe the circuit to your heart's delight.   With enough probing it should be possible to visualize what the circuit is doing in your head.

If my skills were really honed and I had real experience in analog RF and I refreshed myself on how amplified negative-feedback oscillators work, I would be tempted to figure out the true mechanism for the current reversal though the CVR and draw up a timing diagram for the whole shebang.  I still believe that PW got it right when he mentioned that AC was being capacitively coupled through the Q2 array gates and that is one of the main agents responasble for the current reversal through the CVR.  The heart of the oscillator is in the Q2 array, and there may be a wire inductance and a MOSFET capacitance coupled with the amplification that turns the Q2 array into an oscillator that is broadcasting AC throughout the circuit.

MileHigh

P.S.:  TK, off topic:  I am on a Win 7 computer, ASUS motherboard, Core2Quad and my keyboard periodically "nods out" and skips typed characters.  I see the lights flashing like it is doing a soft reset or something.  It drives me crazy sometimes.  Some times whole words are lost and sometimes words I type are chopped but are still proper words so the spell check does not highlight them.  Combine that with how it is difficult to spot your own freshly written prose for errors and it can drive me nuts sometimes.  Besides switching to Linux  do you have any uber-nerd suggestions?