Testing the TK Tar Baby

[TinselKoala]

Ainlsie speaks a language that sounds and looks a lot like English.

It even uses a lot of the same words, and has almost the same grammar as the English we all, more or less, speak and understand on this thread, even if it isn't our first language for some of us.

Texan, which is a rather strange dialect of English, uses words like "y'all" and "ice house" and "per" and "for" and "power" and "energy" and "Joule" and "Watt" in what I nevertheless thought was well-understood and proper, if slightly accented, sentences that have definite meanings. And if one has a Texan dictionary, one can see that these meanings don't change with the wind direction or the d(troll)/dt rate.

And mathematics, which is supposed to be some kind of "universal scientific language", appears not to have penetrated as far as South Africa yet. Here in Texas, by the time a child exits from the eighth grade and is about to matriculate at high school, she already has some competency in basic algebra and understands the subtopic "ratio and proportion" very well, or she will not be able to join her cohort in their freshman geometry and  pre-calculus classes.

But I've never been to South Africa. Things must be really different there.  Are there any aeroplane manufacturers building airliners in South Africa? I think I'd rather just walk.

A Turing Test entry? Hah. She's human, I am one hundred and ten percent (sic) positive. No machine could possibly have that large a vocabulary of words and yet be so very ignorant at the same time.

Not unless.... a lot of its mosfets are blown.

[TinselKoala]

I've been looking at some of the graphics I've put up and which Ainslie has buried or ignored.

Ainslie continues to accuse me of somehow "ignoring" inductance in my calculations, not presenting data, not even DOING calculations based on data, having insufficient sample counts, and all the rest of her clear "calumnies and traducements".... that are really just plain simple lies.

I did in fact compute the mean power values for Tar Baby at its oscillation frequency and I compared the power curve data using several different values for the CVR's AC reactance and total impedance. And of course since that value occurs in the math as a SCALING VALUE, it has no effect on the shape of the instantaneous power curve or its zero-crossings... only in the computed amplitudes.

Here's a graphic I posted some weeks back, a screengrab from the spreadsheet that she can't find, showing the result from using the two most extreme values I calculated with: the simple uncorrected DC resistance of 0.25 Ohms, and an extreme AC total impedance of 16 Ohms.

Note that the shape of the graph is identical in both cases, and the Negative Mean Power Product also survives, albeit in a rather debilitated state. But it remains negative. How could it be otherwise, considering how it is obtained?

Therefore, I conclude that the precise impedance of the "shunt", as long as it is low, will not affect the SIGN of the mean power product obtained by the Ainslie method.

In the spreadsheet, which has been posted here several times, I show _all_ the raw data samples I used to get this product. They are taken from a scopeshot on my analog scope: the shot is in the spreadsheet and all the data points are marked on it, so ANYONE can check to see if I did it right or not, and without wading through a half a million meaningless and unverifiable samples to do it.

Or tell me IF INCREASING THE SAMPLE COUNT COULD ALTER THE RESULT.

Anyone who can find and open the spreadsheet, that is.

When I posted it I asked to be notified if anyone caught an error, SO I COULD CORRECT IT (sorry to shout, it's a long way to SA). Nobody has pointed out anything wrong yet; maybe Ainslie buried it so successfully under Ainslie-piles that nobody noticed it.

(I have calculated Tar Baby's CVR total AC impedance at its oscillation frequency to be about 1.62 Ohms, but I am willing to be corrected by anyone who can show a more accurate calculation and support it logically.)

[Groundloop]

@Groundloop:
Good work, especially the confirmation of the RF radiation. Tar Baby of course howls in RF.
But Ainslie claims none.... we shall see, if she ever allows real testing.

The data sheet I have for the IRFPG50 says the Rdss is 2.0 Ohms, but that is using a 10 volt gate drive. How do you account for your 1.6 Ohm measurement here? Is it due to the 12 volt gate drive?

I also have questions about the bias current measured during the "on" or gate HI signal. Theoretically it should be quite low. I don't remember if I've checked that on Tar Baby or not. But... I will do so as soon as I'm able to, later today.

Data sheet attached.

TK,

I have the F and you have the G. :-)

GL.

[Groundloop]

TK,

I did measure 0,09 Ampere BIAS current @ 13,75 Volt when the circuit was in DC mode (not oscillating).
MileHigh did say that this was too high. What can the reason be for that relative high current? The DC input
voltage to my circuit was 24 Volt. I did calculate that in order to get that current the BIAS voltage input must
"see" a resistance of 91 Ohm.

ADDED: Drawing of the DC mode.

Does this mean that the internal diode in Q1 has a resistance of approx. 40 Ohm?

ADDED: I have solved why the bias current is 0,09 Ampere.

GL.

[fuzzytomcat]

TK,

I have the F and you have the G. :-)

GL.

Hi all,

Some strange things here from "Vishay" Siliconix on the n-channel mosfet .....

It appears we have two (2) PDF's with different manufacturing electrical data.

( 900 volt, 1.6 ohm, 6.7 amp ) IRFPF50.pdf

( 1000 volt, 2.0 ohm, 6.1 amp ) IRFPG50_91254.pdf 

The "International Rectifier" n-channel mosfet .....

( 1000 volt, 2.0 ohm, 6.1 amp ) irfpg50.pdf


NICE ....... 


FTC