Rosemary Ainslie circuit demonstration on Saturday March 12th 2011

[Rosemary Ainslie]

In the hope to avoid any confusion, I'd like to re-emphasize how the power traces were obtained in scope shot Q1_scope05.png previously posted and reproduced here.

When taking a power measurement with an oscilloscope and two probes, we use one probe for voltage across the device of interest, and one probe for current through the device of interest. The latter measurement is obtained (with the Ainslie apparatus) by using a current sense resistor (CSR) placed in series somewhere with the battery, and the resulting current is obtained by using the following equation for Ibattery:

Ibattery = Vcsr/Rcsr

In the case of both probes, these are instantaneous measurements.

In the scope, a MATH function is used to multiply these two values together, i.e. Vbat(t) * ibat(t), in order to obtain the instantaneous power p(t) in that device. Then in order to determine the average (REAL) power, we use a measurement function in the scope to perform an averaging of that p(t) trace, and the scope displays the MEAN numerical value it computes on that trace. Many of you know this already, and this is simply review.

Still wrong Poynty Point.  You need to compute those instantaneous wattage values and then average those values over time.  NOTHING ELSE. You absolutely CANNOT take a mean numerical value of voltage and mean numerical value of amperage and give us that product as being representative of the power from or to that battery.  Your Spice sim would need to do multiple sampling analysis of multiple points in each waveform.  Self-evidently it cannot - as it represents the voltage sample range as a mean average 'ripple'.  If it was doing what it was meant to do the resultant waveform resulting from that power analysis would 'FOLLOW THE SHAPE' of the voltages across the shunt and the battery.  That is why the LeCroy and the Tektronix are so excellent.  It holds up to 500 000 points sampled on any one screen shot.  Your simulator can NEVER get even close if it can do it at all. When you can show a TRACE on that PSpice simulation of yours that actually does this - then ONLY will I be convinced.  The wattage values resulting from those products NEVER defaults to anything that looks like a steady line ANYWHERE AT ALL.  YOUR ANALYSIS IS FLAWED POYNT.  GROSSLY.

In PSpice v10.5, there is a wattage probe that can be placed onto any device. After doing so, the scope shows the p(t) of that device. As discussed, this is the instantaneous power in that device, and is the instantaneous voltage across it times the instantaneous current through it. The wattage (W) probe in PSpice does this multiplication and displays the results automatically.

Then the trace resulting from that product would look VERY DIFFERENT.  I can pull up the waveform from the wattage over any part of our own data dumps and it never, never, never defaults into a straight line with a small ripple.  Therefore whatever you CLAIM PSpice is doing has no bearing on what it is actually doing.

After the p(t) trace is on the scope screen, we can perform the same averaging function we do with the scope in order to obtain the average (REAL) power in that device. In PSPice this function is called "AVG", and you will often see this included in the trace statement at the bottom of my scope shots, esp. when power is being examined. In PSpice, rather than displaying the average of p(t) numerically as on the scope, it shows you the running average of the p(t) trace, and you can see and measure with a cursor what the final value is that it converges on. This is how I determine the numerical values I place on the scope shots and write in the posts. See the scope shot below.

If you can see the wattage with a cursor - AS YOU CLAIM - then show us that wattage shape - drawn as it is - in real time - from the traces of the current and the voltages.  What you are SHOWING US is something that is averaged.  And with respect, I think this is because PSpice CAN'T do that instantaneous wattage.  And IF IT CAN - then, I say it again.  SHOW US THAT PRODUCT. 

Look again at the PSpice traces that you've shown us.  It has flatlined the oscillation.  A product of the amperage and voltage WOULD NEVER FLATLINE.  Why don't you realise this?  It has INDEED averaged everything.

Rosemary

[Rosemary Ainslie]

Look at the shape of the math trace here Poynty.  That's what I expect from PSpice.  Anything less and we are not looking at anything relevant.

I'll see if I can find a sample.

[Rosemary Ainslie]

And Poynty.  That highlighted bit of nonsense in your download of the battery voltage.  WHAT A JOKE.  You write this.

negative power of the battery = -33.3 watts.  Then you add (a negative battery power is normal and indicates a net loss of energy over time) - in paranthesis.

WHAT A LOAD OF COBBLERS.  IF ANYONE EVER - ANYWHERE IN THE WHOLE WORLD - EVER FOUND A NEGATIVE POWER VALUE - THEN THERE WOULD BE NOTHING BUT AMAZEMENT.  There is absolutely NOTHING normal in a negative wattage value.  In fact - so EXTRAORDINARY is this that it would be an entirely meaningless term.  Wattage is NEVER NEGATIVE.  Nor is it EVER EXPECTED TO COMPUTE TO A NEGATIVE VALUE.  IF AND WHEN IT DOES THEN IT IS EXTRAORDINARY.  This truth is so enshrined that I was ASSURED - by every expert that I have ever spoken to - that it is ENTIRELY IMPOSSIBLE TO GET A NEGATIVE WATTAGE VALUE because it is ENTIRELY IMPOSSIBLE TO RETURN MORE ENERGY THAN DELIVERED.  And a NEGATIVE WATTAGE ABSOLUTELY HAS NOTHING TO DO WITH THE ENERGY LOST OVER TIME.

You really do take us all for fools.  God help us if the nonsense that you expound is believed by anyone ever.  It is a travesty of science.

Rosemary 

[cHeeseburger]

Guys - please note.  There is absolutely NO WAY that one can do any power analysis on any circuit that has a waveform that is more complex than a direct current.  Certainly it cannot be applied to the waveforms we're generating here.  I do hope you realise this.  And in as much as no-one has challenged all this nonsense I'm concerned that perhaps the most of you do not know this.  If you doubt it - then look it up.

Where do we "look it up"?

WTF?  Poynt is doing the exact same set of measurements at the exact same points in the exact same circuit you have been doing ever since day 1.  And showing the exact same results.  His use of averaging (being done AFTER the real-time point by point multiplying) is exactly the same as your own use of averaging: MEAN AFTER MULTIPLY. 

That full set of measurement done the same exact way you have always done it serves only to show that the sim agrees rather precisely with your bench testing when done in exactly the same way.

Then, after doing the work to replicate your circuit, your tests and your results almost to the millivolt, showing the validity of the simulation rather convincingly, Poynt goes on to move the scope probes and the shunt to the proper locations to allow for a true direct unpolluted measure of the battery voltage and the battery current.

He continues to use the same real-time point-by-point sample multiplying technique just as you always have done and then takes the mean of the resulting power trace, just as you have always done.  The results clearly show a large net drain from the battery and a very low efficiency of power delivery to the heating element through the circuitry. 

HE CAREFULLY POINTS OUT AND AS TESLASET HAS CLEARLY VERIFIED (AND I VERIFY AS WELL) THAT THE PSPICE W PROBE USES THE CONVENTION OF STATING A DRAIN FROM THE BATTERY AS A NEGATIVE POWER. 

Why don't you just quickly put your shunt right at the battery negative terminal itself and take a good look?  It's not hard to do.  It will show you immediately that the net current is draining the battery and not charging it.  No need for lengthy battery runs at all.  No need for further confusion.

Rosemary, your arguments are so thoroughly "straw-man" based and so poorly stated at that, that you really should take a break from your incoherent ranting and just try to absorb what has been clearly shown:

The simulation behaves exactly like your circuit.  When measured wrongly, exactly as you measure, the results agree with your results.  When measured properly, the very low COP and gross inefficiency of powering a heating element using MOSFET parasitic oscillations is revealed clearly.  End of story.

No amount of rude insults, calling people dogs and vermin and stupid, will change these well-demonstrated facts..

Humbugger

[poynt99]

Look at the shape of the math trace here Poynty.  That's what I expect from PSpice.  Anything less and we are not looking at anything relevant.

I'll see if I can find a sample.

I see now one thing that is tripping you up; it's the p(t) trace. I usually fore go displaying that, and I go directly to the application of the AVG function on that trace.

When you see the trace statement as" "W(R1)" for example, that means the trace is of p(t) for R1.

When you see the traces statement as: "AVG(W(R1))" using the same example, that indicates MEAN[p(t)]. This latter trace statement is what is required to obtain a numerical value of the average/real power in the device of interest.

Here is what p(t) looks like before applying the AVG function.


.99