Testing the TK Tar Baby

[TinselKoala]

I picked up a 70 000 microFarad, 30 volt cap today at the surplus store. I'm not going to say how much I spent for it, it's embarrassing.

I also got a handful of H11D1 optocouplers and a bunch of sand and wirewound resistors to make loads and shunts with. I can even reproduce the 4+1 group of one-Ohm ten watt "shunts" or CVRs.

Ten Watts here? Four paralleled to give 40 Watts power handling in the CVR?  OK, fine. I guess that is just in case they do run the mosfets fully on by mistake and manage to expose the 10 ohm load to the full 72 volts along a low-resistance pathway for more than a brief pulse. But if they are averaging only 12 Watts at the load, as claimed.... they aren't running fully on, or they are on for a very short duty cycle only, or both, and don't need a 40 Watt CVR. All of that is conjecture, of course, since we have no reliable data from the NERD RATS.

I also got 5 nice ceramic tube , wirewound, 50 Ohm 12 Watt resistors, Clarostat VPR12F50 kind. I will be parallelling these to make a nice load that will fit better into my insulated Fleaker system than the water heater load does.  So I'll have a 10 ohm, 250 Watt equivalent load in there, which should be beefy enough to handle the anticipated maximum continuous current from the Tar Baby at 36 volts. Which is, neglecting the Rdss of the mosfets, by Ohm's law, I=V/R, so the current will be 3.6 amps, and by Power = I^2 x R, I get about 130 Watts dissipated by the stack at 36 volts and 3.6 amps ... so we be cool. Cooking with Hot Grease now, momma. With enough "oomph" if the mosfets do turn on to make tea or oxtail soup.

Now, if I could just find an old ox to chop the tail off of for my soup.....

ETA: Isn't anyone checking my math? OOPS..... five x 50 Ohm, 12 watt resistors gives us 10 Ohms all right, but only 60 Watts power handling capacity, not 250 like I figured above. So it's a good thing I've got them immersed in oil.... because I could exceed their power rating if I'm lucky.

[evolvingape]

Hi TK,

Yes it's an interesting quote.

There is the issue that there were originally 9 batteries, and no data logging on hours run on which battery under what conditions, so the number you calculated for total energy available has unknown energy input from an additional 2 batteries included in the total for the 7 batteries.

Then you have the issue of surface charge on the plates. As we all know even a completely discharged battery if left alone overnight may recover enough to show a Voltage reading above 12V across the terminals under no load. This is why when people try to start the car and the battery dies, if you leave it 10 - 15 minutes and try again sometimes it will have enough energy to turn the starter and off you go. Never trust a battery is the rule of thumb.

Then as you mentioned the function generator is run from the wall socket and is capable of providing a float charge to the battery, which would probably keep it above 12V for a very long time while under a relatively tiny load.

Then you have the issue of potential desulfation of the plates from the oscillations:

http://www.reuk.co.uk/Battery-Desulfation.htm

Voltage spikes are the method of choice for reconditioning sulfated lead plates, restoring the charge holding capacity of the battery, not normally done while running a load at the same time though. The battery can even desulfate itself from it's own power if you hook the circuit up to the terminals, and recover charge storage capacity that way. You will still have to charge the battery though for a fully charged status of 13.2V, which is calculated as 2.2V PER cell, and 6 cells PER 12V battery, which equals 2.2V MULTIPLIED BY 6 cells = 13.2V fully charged.

RM

[picowatt]

TK,

Regarding your proposed optocoupler setup...  That would be fine if the MOSFETS were being switched on hard.  But they are not.

Q1 may be switching on fairly hard (and only briefly as per the RA waveforms) but the Q2 "quad array" is being biased into a somewhat linear region.  I believe .99 has touched on this.  Even with the FG at  -15 volts, the Q2 drain current at DC wll be limited to 300 ma or less due to the Vdrop across Rgen.  More likely it is between 100-250 ma.  Without knowing the FG open circuit voltage versus in circuit voltage, it is difficult to know the actual bias current used.  The variability of the gate threshold voltage also presents an unknown.

For DC and at 300 ma. bias, the FG 50 ohm would dissipate 4.5 watts, the 10 ohm (?) load .9 watts and the Q2 ""array" would dissipate around 17 watts (assuming a 60 volt battery set).    The parallel pair of 100 ohm output resistors in RA's FG (as per the schematic .99 posted) may not even be 100 ohms anymore, as 4.5 watts would be a bit much for them.  Opening the FG and inspecting/measuring these resistors is in order.  Possibly, if they have increased in value, the bias current is way less than one would expect from the Rgen of 50 ohms.  If they are intact, the FG can only do just above -15 and Ibias would therefore be under 300 ma.

For AC current, the FG output is bypassed by the 12,500 pf (or more) of the total Q1 and Q2 gate to source capacitance, hence you better stick with the IRFPG50's, as all that capacitance sets the AC gain of the Q2 common gate amplifier and passes the AC current to the battery negative (via the CSR).

Picowatt

[MileHigh]

TK:

If you do the capacitor test then it would be worth it to measure the capacitance of your bank of capacitors.  Many people reply on capacitance and inductance meters.  I will go out on a limb and assume that 100,000 uF is to large a capacitance for a capacitance meter to measure.  What do you do???  Plus you mentioned tolerance, an almost taboo term on the free energy forums.  And the tolerance of big electrolytic caps is what, +/-10%?  Perhaps +/-20%?

So you might get some oohs and aahs if you made a clip where you measured the capacitance via the RC time constant using your best multimeter.

Then you hope your your cap-based power measurement will be in agreement with the Clarke-Hess 2330 and with the Tek scope.  Convergence!  That's another important lesson for all.  Hasn't a whole year's worth of discussion all hinged on a single measurement method that many people considered highly suspect?

You asked about the inductance of the the inductive resistor.  The off-the-shelf RAT heating element had minuscule inductance.  So small it might be too low for an inductance meter.  That's a reasonable guess, I have never played with one.  I am going to guess in the tens to hundreds of nano-Henries.  Same thing for the heating element that you bought.  I would guess that you can just take the rule of thumb for inductance per inch of wire and apply it to both of those commercial heater elements.  I forget the rule of thumb, it' something like 10 nano-Henries per inch.  In both cases it looks like the inductance in the interconnect wires would be comparable to the inductance of the heating element.

MileHigh

[TinselKoala]

TK,

Regarding your proposed optocoupler setup...  That would be fine if the MOSFETS were being switched on hard.  But they are not.

Q1 may be switching on fairly hard (and only briefly as per the RA waveforms) but the Q2 "quad array" is being biased into a somewhat linear region.  I believe .99 has touched on this.  Even with the FG at  -15 volts, the Q2 drain current at DC wll be limited to 300 ma or less due to the Vdrop across Rgen.  More likely it is between 100-250 ma.  Without knowing the FG open circuit voltage versus in circuit voltage, it is difficult to know the actual bias current used.  The variability of the gate threshold voltage also presents an unknown.

For DC and at 300 ma. bias, the FG 50 ohm would dissipate 4.5 watts, the 10 ohm (?) load .9 watts and the Q2 ""array" would dissipate around 17 watts (assuming a 60 volt battery set).    The parallel pair of 100 ohm output resistors in RA's FG (as per the schematic .99 posted) may not even be 100 ohms anymore, as 4.5 watts would be a bit much for them.  Opening the FG and inspecting/measuring these resistors is in order.  Possibly, if they have increased in value, the bias current is way less than one would expect from the Rgen of 50 ohms.  If they are intact, the FG can only do just above -15 and Ibias would therefore be under 300 ma.

For AC current, the FG output is bypassed by the 12,500 pf (or more) of the total Q1 and Q2 gate to source capacitance, hence you better stick with the IRFPG50's, as all that capacitance sets the AC gain of the Q2 common gate amplifier and passes the AC current to the battery negative (via the CSR).

Picowatt

Thanks for your analysis, I really appreciate it. Your knowledge appears to exceed mine somewhat in these matters, so I'll just say that your estimates of maximum drain current attainable agree with my measurements. In fact without really cranking up the output of my F43 FG, which is capable of 40 V p-p into 50 ohms, I mostly stay under 200 mA as long as I am using a strict negative going pulse so that only the Q2 "gang of four" is active. However, I can operate my circuit in a mode, by varying offset on the FG, so that both mosfet sets turn on, and the oscillations can be seen on both phases of the signal, and the current can go up to 3 amps or more, mostly going through Q1 I think, from comparing temperatures. I'm not properly heatsunk so I only operate at those levels briefly, but now that I have the better load setup with the wirewound resistors, giving more inductance than the water heater element, I'll be pushing it a bit more.

I don't think I'm willing to buy an Instek FG of the same model as theirs. They cost about 220 bucks and I'm not rich. Besides.... I feel that the Interstate F43 is doing just fine, so far. It's an oldie but a goodie, and I keep it in good condition. Still.... the Instek is available locally if I do need to get one.

I'm not sure if I follow your reasoning as to why I should stick to the PG50. Are you saying that the PG50 may be necessary for some kind of battery charging from the FG to occur, and it might not if I use the 830a? Anyway, I'll be switching to the PG50 as soon as they arrive, and I'll be doing side-by-side comparisons like I did those many moons ago with the 2sk1548 in her COP>17 circuit.

The damn RF from the thing messes with both my little digital thermometers... I've had to start using an old analog bimetal type to get a reliable reading, with loss of precision and accuracy.

And of course, since I'm exploring these other modes as well, I'll try the optocouplers anyway. As long as I don't kill the oscillations....
because without oscillations there is no way that certain people will think I'm really doing what I'm doing, really.