Testing the TK Tar Baby

[MileHigh]

A little side comment about the whole capacitor test was that I wanted to emulate only a single battery with a large cap in the middle of the array of batteries to "keep it battery-like" if that makes sense.  This was to keep Rosemary happy, she would see exactly the same oscillations on her scope and the power source was still predominantly batteries so any imagined "battery effects" could still take place.

Anyway, I think that the test is an interesting and worthwhile exercise.  One of the important concepts is that if you have five batteries in series, each battery will contribute 20% of the power required to run the circuit.  Alternatively, if the bank is being recharged, each battery will take 20% of the recharging power.

So by "slipping a capacitor" into the battery array and replacing one of the batteries, whatever the capacitor is indicating will also be happening within each of the individual batteries.

And certainly the capacitor will go down in voltage as the circuit runs, even if your DSO is spitting out garbage data that says that the battery bank is being recharged.  That's because the capacitor is more accurate than the DSO and can't be fooled.

Just one little tweak:  Assuming 100,000 uF and 25-volt rated electrolytics, and assuming the batteries are actually 12.6 volts, then why not time how long it takes to go from 15.6 volts to 9.6 volts for high power dissipation.   For low power dissipation, time how long it takes to go from 13.1 volts to 12.1 volts.  That way you are quasi-centered around the true battery voltage and you get a slightly more accurate average power dissipation measurement for the circuit.

I also have a caveat:  In thinking about the process in my head I am quite certain that as the cap discharges the rate of voltage drop will actually start to accelerate, not slow down.  We are so conditioned to expect that the rate of the voltage drop on a cap will slow down as the cap voltage decreases.  So it may seem counter-intuitive but indeed, as the voltage on the cap decreases, the rate of the voltage drop will increase.  Kind of a fun brain twister. (** WRONG ** - I failed my own brain teaser.  The rate of voltage drop will NOT increase, it will decrease but very slowly.  Typically it will be so slow so that it will almost look like the voltage is dropping linearly.  That was my Doh! moment.)

MileHigh

[MileHigh]

TK:

I went back and reread your posting.  You have only 40,000 uF to work with.  So I recommend that you scope the cap voltage like I said and watch very carefully.  Do not use a digital multimeter with a 1/2 second display update.  You might get squirted in the eye with some primordial ooze if you are not careful.

The fact that you are typically running with 30 volts instead of 60 volts may save you here because that's 1/4 the power dissipation rate as compared to 60 volts.

I have this gut feel that those 25,000 uF 25-volt "Coke can" electrolytic capacitors are relatively cheap.  Waaaay back I used to go to these 'electronics depot' type places that had recycled and Grey market and who-knows-what parts.  You may find some big caps in a place like that on the cheap.

MileHigh

[TinselKoala]

Thanks, MH, I'm with you all the way. I've also got 16 x 4700 uF @ 40 v surge caps but that would be too much wiring for my liking. Unfortunately, I live in an intellectual "black hole" and there is only one electronic surplus place that will have big cap pulls from obsolete gear and they will be mixandmatch (but I'm going there today to take a look). The only decent "new" parts place is 100 miles away, the Fry's in Austin. Of course we have RS, if you need a battery or a common resistor or a cellphone... but for electronic components it's a weird wasteland. Not like TO where I could just walk down to the hardware store and browse thru their component section where they had _every_ Japanese 2sk transistor and _every_ 74 series logic chip IN STOCK to support the EE department at UofT, or go into one of three different surplus stores to get the "weird stuff". I even found a medical electroshock machine in Active Surplus there one time... nice bank of big caps in that one !!
But around here, electronics means car stereos and not much else, and the big caps they sell for those -- really impressive big caps -- are just too expensive for me to mess with. But sure, some of these car stereos around here will have FARADS of capacitance stuck into the trunk somewhere to pump that drumandbass into their blown speakers.



Meanwhile, I found this hour-long seminar on.... power measurements in AC circuits.

http://www.youtube.com/watch?feature=endscreen&NR=1&v=MHa18mC2ZLc

Your capacitor test will be a good easy one to do to test for battery charging vs. battery discharging by the circuit. But when I get down to actual power measurements, _assuming_ that the performance of the circuit can be derived from electrical parameters at all -- I'll be using the Clarke-Hess 2330 sampling V-A-W meter as a power analyzer. No scope traces to interpret, just digital readouts of nice red numbers. I'll measure the input for a while with the load at equilibrium temperature, then I'll swap leads quickly and measure the output for a while at the same equilibrium temperature, then re-charge lather rinse repeat.

Later on, I'll subject the Tar Baby to the fancy Tek scope with internal math and we'll see what kind of nice colored wiggly lines it will make. You don't want to skip ahead to the "money shot" already, do you ?


--TK

(Thanks for your help, I really appreciate it.)

[TinselKoala]

There are still a few points about the open-source, help-to-replicate NERD RAT device I would like to understand.

First there's the value of the inductive-resistive load. I've seen inductances cited that are all over the map. I've seen 1.5 HENRY posted somewhere as the inductance of the load. This seems implausible to me based on the waveforms I've seen and what I know about the construction of common water heater elements. Also I have little confidence in instrumental measurements, especially difficult ones like inductance, that have been reported concerning that device.

Then there's the question of the battery capacity of the batteries that team used. I can't figure it out from the company's website catalog listing of the battery they used. The most reliable outside source I could find (humbugger) has it at 50 A-H, but the "official" reports of the NERD RAT device have it listed as 40 A-H. If battery draw-down tests are being considered, one would normally like to know the actual rated battery capacity, just in case someone cared enough to wonder if a 25 percent difference in actual capacity could affect such a test at all.
( 40 plus (25 PER cent of 40) = 50 ).

[TinselKoala]

How do we know the state of charge of a battery? It is not easily and directly related to the simple, no-load or lightlyloaded terminal voltage measurements that are so often cited as "evidence" for the battery's charge state. The battery's " discharge curve" must be considered, along with the load's demands. That's why drawdown tests or simple side-by-side tests like performance under a _heavy_ load are preferred to simple voltage measurements.

What do we expect to see, as a measurement of the terminal voltage of such a battery?

1) when it's being recharged by an automatic battery charger designed or approved for the battery type
2) immediately after it's been fully charged by the charger, still no-load
3) one hour after 2), unused, no load
4) when first hooked up to the operating circuit, under load
5) after running the circuit for long enough to "theoretically" (using conventional theory) to have transferred half of its stored capacity to the circuit and the load, tested both under load, and unloaded?

For my 12-volt, 5 A-H, sealed lead-acid batteries, I can answer some of the questions.

1) is about 14 volts. I am using an automotive "automatic" charger that charges at up to 2 amps and that goes into a "float" mode when it senses that the battery is full. How it knows, I dunno... but I am assuming it does and is consistent between batteries. (actually I do know, but it's not too important I hope)
2) is about 13.8 volts
3) is about 13.5 or 13.6 volts
4) is about 13.0 volts (so I get about 39 volts from my stack when first running the TarBaby circuit)
5) is still unknown

The most important point here is that simply citing "the battery is still over 12 volts" is NOT evidence that it is fully charged. In fact, if a large 12 volt silver-calcium-lead-acid battery IS fully charged, its terminal voltage should be something over THIRTEEN volts when measured under no or light load.