Testing the TK Tar Baby

[Groundloop]

GL,

From your measurements, the MOSFET RdsON isn't quite1.6 Ohms. The equivalent RdsON is actually 6.46V/1.83A = 3.53 Ohms.

Take out Q2 and re-measure V_DS. Then re-calculate the Rds using 1.94A.

.99,

I'm currently in bed with a flu and fever.

I will do new measurements when I can think straight again.
Then we can see if the theoretical analysis agrees with measurements.
Will be back in a couple of days.................

GL.

[poynt99]

.99,

I'm currently in bed with a flu and fever.

I will do new measurements when I can think straight again.
Then we can see if the theoretical analysis agrees with measurements.
Will be back in a couple of days.................

GL.

Take it easy, and hope you're feeling better soon

[polln8r]

Hi all,

I used a different method from TK to measure the pixels in his pixel puzzle. I also used GIMP, but, used the 'magic wand' selection tool to get these results. I also get the same results if I try to manually trace the outline, but in all honesty I don't trust the accuracy of any of these methods when using this image as its resolution makes it difficult to determine where the edges of the green and red really are.

Cheers,
I hope this is helpful, considering I'm not even remotely qualified to discuss the actual business in this thread.

Polln8r

[fuzzytomcat]

Hi all,

I used a different method from TK to measure the pixels in his pixel puzzle. I also used GIMP, but, used the 'magic wand' selection tool to get these results. I also get the same results if I try to manually trace the outline, but in all honesty I don't trust the accuracy of any of these methods when using this image as its resolution makes it difficult to determine where the edges of the green and red really are.

Cheers,
I hope this is helpful, considering I'm not even remotely qualified to discuss the actual business in this thread.

Polln8r

Hi Polln8r,

Thanks for the work.

It may seem odd doing areas this way but prior to the fancy oscilloscope being able to calculate this with a push of a button, there was someone who painstakenly cut out the wave form printed on paper possibly from a photographic image taken, using the area above and the one below zero and actually put them on a scale and weighed them.

We have now came to the pixel puzzle method ... not bad if you don't have a fancy scale 

Fuzzy

[TinselKoala]

@MH:

That, but more than just that.

Examine the claims made for this shot and the ones immediately surrounding it on her blog.

For one thing this is the "high heat" trial that resulted in the claims of 44 or 80 Watts dissipated at the load.... definitely enough to warm up even that big chunk of thermal mass she is using-- which, incidentally, will stay hot for a while after the current is removed. Duh.

However... _and this is the key claim_  ... she claims that this high heating is achieved without drawing power from the battery.

This is why I want that scopeshot fully analyzed in public by somebody other than me.  I've listed the points that need to be covered in the analysis in an earlier post. The interpretation of those points will follow after the complete analysis. So please.... humor me. I've done plenty of things that I've been asked to do... so now I'm asking for somebody else with calipers and a calculator and knowledge of scopeshots and Ohm's Law to analyse SCRN0235 fully and "do the math" and post the results here.

There are two phases to the period of pulsation: One with NO oscillations, the Gate HI portion, and one with copious oscillations, the Gate LO portion. Notice how we lately have been subtly steered towards the idea that the oscillations are the important features of the claim and the action of Q1 is actually unimportant in the desired mode, so we shouldn't even bother to turn it on.

But the EVIDENCE that Ainslie constantly cites of strong load heating invariably comes from trials in which a bipolar pulse is used AND A WEAKER BATTERY PACK of 48 volts is used generally.  In other words, Q1 is put to use, indeed. You will note that this SCRN0235 is one of the RARE trials in which the full 72 volts "AND" a substantial Q1 ON time are both allowed. And I know why, and you do too.

Now.... there IS manifestly a lot of power shown in the SCRN0235 trial. Is it 44 Watts? Or even the "impossible" 80?

Come, let us calculate.
Batt voltage is 72. Shunt resistance to DC is 0.25 Ohms. Voltage drop across shunt DURING GATE HI is 1 Volt. Current through shunt is 1/0.25 Amps. Instantaneous power is therefore ((1/0.25) x 72) Watts. Right? Period is 31 microseconds, and ON time, not even including the ramp up, is 14 microseconds. Duty cycle ON time, no oscillations, then is (14/31). AVERAGE POWER then, IGNORING OSCILLATIONS ALTOGETHER, is
 
Pavg == ((1/0.25) x 72) x (14/31)

Right?

Note that I am not even counting the "ramp up" time where the current is still rising to its steady level within the ON periods.

This is why I need somebody else to do this.  WHAT POWER LEVEL is shown by this scopeshot, even disregarding the oscillations?

And it is easy to calculate from the scopeshot, and it has NOTHING TO DO WITH OSCILLATIONS.

And this is the point. The large power levels that are heating the load in this trial are essentially DC, even though they are chopped up into little timeslices. Ainslie has used long duty cycles with 17-second ON times (no oscs, when a functioning Q1 would be passing current). Surely nobody could argue that , during those 17 seconds, the current is NOT DC. So how does shortening the _period_ turn DC into AC during the Gate HI, Q1 ON timeslices? IT DOESN'T.

Regardless of that, just compute the AVERAGE POWER shown, using the duty cycle and whatever definition of AC-DC you like, but ignoring FOR NOW the contributions of the oscillations. They can only "add" to the total average power anyway, right? So by leaving them out for now we should be UNDER estimating the "true" power in the circuit. Right?
(But for Anslie's claim to be true, the oscillations must actually SUBTRACT at least as much power as is added during the non-osc portions in order for the overall average to come out negative. Right?)

Once we have a full analysis of what the scopeshot is telling us, from somebody other than me, we can start talking about How those measurements were generated by the circuit, and What they mean for the Ainslie claim set.

We are getting closer and closer to calorimetry, but first we need to understand this issue fully: how important is the STRICTLY DC component of the circuit's total average power to the high-heat trials and the claims arising from them? And this is what a comprehensive analysis of SCRN0235 is meant to accomplish.

So not only is this a forensic investigation into the torture slaying of an innocent, hard working mosfet, an unnamed victim of oscilloscope abuse, but it is also an important step in understanding and comparing calorimetric data.