Testing the TK Tar Baby

[poynt99]

@.99.... that scope shot is the signal from the mosfet Drain, with respect to the negative rail?? Or is that the signal "across the load", with the probe on the drain side and the reference on the battery side, or vice versa? I'm not getting it. And with a gate driver.... I am also surprised to see such a messy signal anywhere. Is it that the PG50 just can't keep up at the frequency chosen?

The scope shot is the trace across a 0.1 Ohm CSR.

[poynt99]

It is a pity that Greg does not understand his error. He is actually computing the power in RL with his method, not the battery power.

When computing the power from the battery, the battery voltage is not derated by the duty cycle.

Pin_AVG is simply VBat_AVG x IBat_AVG

[poynt99]

TK,

Pages 33-40.

[picowatt]

It is a pity that Greg does not understand his error. He is actually computing the power in RL with his method, not the battery power.

When computing the power from the battery, the battery voltage is not derated by the duty cycle.

Pin_AVG is simply VBat_AVG x IBat_AVG

It is even moreso a pity if Greg allows "her" to attempt to teach him anything at all related to electronics.

Prior to .99 throwing in the towel on the Basic FG thread, she was again disputing the ability of the MOSFET capacitances to carry significant current.  She was discussing the waveform at the gate of Q2, and from her posts, it appears that she was referring to the FG output waveform as being the waveform at the gate of Q2.  Looking at the schematic, it is obvious that the waveform at the gate of Q2 is essentially the waveform indicated by the CSR trace (the gate of Q2 is, afterall, connected directly to the CSR).

She continues to demonstrate her inability to read her own simple schematic, let alone comprehend basic electronic concepts or understand how Q2 is biased on.

As .99 states, power delivered to or from the battery is simply the product of Vbatt and Iavg.

A pity indeed...

PW 

ADDED:  She also stated that she has corrected the errors in her documents.  Did I miss her corrections regarding Q1 not turning on in FIG3, 6, and 7 even though there is sufficient gate drive indicated to do so?

[TinselKoala]

OK, let me see if I can follow this argument and convince myself.

First, I hope we can agree that the instantaneous input power is the product of the true instantaneous battery voltage times the true instantaneous current thru the CVR. This accounts for both phase differences and dutycycle and results in an instantaneous power curve, which can then be averaged in the usual manner, by integrating then dividing by the time interval.

However we are trying to find the input power without using scope math or spreadsheet operations on lots of tiny samples.

We know that the apparently fluctuating battery voltage really isn't, so that the instantaneous battery voltage can be replaced by the heavily filtered, steady DC average voltage. This then can be multiplied by the instantaneous current curve to give the correct instantaneous power curve.

And then when we compute the average of the instantaneous power by integrating and dividing by the time, we can pull the constant battery voltage out of the integration and just integrate and average the current curve and then multiply, since the sum of (a constant times some values) is the same as a constant times ( the sum of some values).

Therefore, the average current multiplied by the average DC voltage results in the true input power value, and all phase and duty cycle issues are taken care of by this process, since there is no phase issue (the voltage isn't really fluctuating) and the duty cycle contribution is taken care of by averaging the instantaneous current across the "dead" times by integrating and dividing by the time interval.

There is still some difficulty in my mind with the average current figure. I accept that .99 has shown that the DMMs give a usable and reasonably accurate reading of the average current in the filtered CVR. In the present case, are the average current value determined by this method, and the RMS current value determined by multiplying the peak current of the always-positive rectangular pulse by the sqrt of the duty cycle, the same? Perhaps that adds to my confusion. Is this always going to be the case? I need another cup of coffee.

OK, I'm convinced, and I see the difference between measuring the supply to the black box, and what the black box is supplying to the load. The supply is providing a steady DC that is "sipped" periodically by the circuit and so is correctly analyzed as above. The load is seeing a pulsating, oscillating supply coming from the circuit and so may be analyzed using rms values of the pulsating voltage and current.


I am now very suspicious though of using the rectangular pulse simplification for the load, if the CVR signal is as messy as all that, shown in the scope trace above. How are the RMS values of voltage and current at the load actually determined, for the average power (Pavg = Vrms x Irms as used by gmeast) at the load? Are these "numbers in boxes", computed by gmeast's scope?