Testing the TK Tar Baby

[polln8r]

I know, right?

"The batteries are shown to be not discharging. If it is found that they really are it won't matter, because the measurements say they aren't."  (not a direct quote). (Indirect quote continues...) "Therefore, we can assume that the measurements are an anomaly, which means that we have found something truly incredible.

Incredibly anomalous measurements. Indeed.

Edited: Forgot to say thanks for the link. Expletive? I didn't hear shit.

[TinselKoala]

Hmmm.... there seems to be some confusion all around about just how to calculate equivalent heating power in a circuit that is receiving a DC pulse train or an AC rectangular wave or oscillations.

Perhaps these links will help. The first document has been posted before, I think, and the Wiki seems to agree with it.
http://www.eznec.com/Amateur/RMS_Power.pdf
http://en.wikipedia.org/wiki/Root_mean_square


At this point I think the main discrepancy to be accounted for is the accuracy and false precision in gmeast's quoted data. There is no way that anyone can accurately measure heating power down to the tenth of a milliWatt using tabletop hobby equipment, sorry. The latest actual results, calculated using average power = V(rms)^2/R or V(rms) x I(rms) as is proper according to the links, give a result consistent with unity performance, within the limits of measurement accuracy. And without seeing calibrations against real standards, the accuracy in the third or fourth least significant digit of any tabletop instrument is questionable and should be confirmed by multiple measurements with different instruments if possible.

[TinselKoala]

I sure wish I could see gmeast's waveform shots. I'm kind of concerned still about his use of the Drain voltage and how it's being interpreted. I think the drain voltage is important if used right. But it has proven to be tricky for Ainslie fans to interpret. In the Ainslie circuit the drain voltage will be HIGH at battery voltage when the mosfet is OFF and the load is not carrying current, and will drop to a voltage that reflects the total resistance of the circuit when the mosfet is ON and the load... and the mosfet....are carrying current.
Also, is gmeast aware that the Rdss ON of an  IRFPG50 is at least 2 ohms? That's not chopped liver in a circuit with a total on-state DC resistance on the order of 12-14 ohms. The mosfet will be dissipating a significant part of the total circuit power dissipation.

[TinselKoala]

. checking math -- I agree with the Pavg = Vrms x Irms , and that for a rectangular, all-positive waveform Vrms = Vpeak x sqrt(dutycycle) , so I don't know what .99 is meaning in this case... but...
--
Actually, running gmeast's numbers using a voltage drop across the mosfet of 1.6 volts, so the load only sees 24 volts peak, 12 rms, I get a better result of 1.09 or 109 percent efficiency compared to the DC case.
Close enough to unity to demand more precise and accurate measurements.

[picowatt]

TK,

It appears that Greg has measured/calculated Iavg and Irms to both be around 400ma.  So, average power from the battery would be Vbatt times Iavg, which is how .99 arrived at the 10.5 watts from the battery figure (25.6x.410).

Greg also measured/calculated the Vrms across Rload and then multiplied that number by the Iavg (or Irms, both of which were in close agreement) of ca. 400ma, and arrived at the power dissipated in Rload.

I am actually quite impressed that his measured/calculated power in Rload and the equivalent power he measured for Rload using his DC power supply and the measured temp rises are in such close agreement.  My hat is off to him for that.

However, to measure the power delivered by the battery, I agree with .99 in that Iavg should be multiplied by Vbatt to arrive at Pavg from the battery. 

It is somewhat difficult to follow along seeing only the text in the posts.  As you say, too bad the schematics and 'scope shots are unavailable.

Oh well...


PW