TSSP: List Archives

From: Paul
Date: Sun, 22 Oct 2000 17:18:39 +0100
Subject: Re: [TSSP] NSVPI - Latter Results

Some notes and comments on Terry's NSVPI results.

One can see from the pictures what a delicate job it must be
to place the 'Terryphorus' against the coil. This must have
been a time consuming and tedius measurement run requiring
considerable care, patience and determination.

The results are of sufficient stability and precision to reveal
and confirm all the salient features of the predicted V profile,
in particular the slight concavity near the bottom, followed just
above by the gradient maximum, presumably associated with a current
maximum at that point. Above, the rise continues convex.

How does the integrated top volts of 69.72 match with tsim predictions?
Well first of all, the top voltage is actually 69.72 + 0.6606 since
the measurements are relative to the base potential and tsim predictions
are relative to ground. Now tsim predicts a transimpedance value 55.898K
ohms, so with an input current of 0.0465/50 amps, we would expect

   Vtop = 0.0465/50 * sqrt(2) * 55898
        = 73.5 volts.

against the 70.4v measured, so the prediction is within 5%. The
transimpedance of 55.898K came from simulations of the coil
which we carried out in august, and at the time the resonant frequency
was a little higher than for the present measurements. I'd suggest
therefore that we are entitled (on the basis of pn1310 equ 9.2) to
adjust Zft in the ratio of the two Fres. This adjustment leads to a
revised Zft of 55.898 * 146.34/148.37 = 55.133 K ohms. This leads
to a revised Vtop of 72.5v wrt ground, now a 3% error.

I'd say that since the measured top voltage is potentialy subject
to an accumulation of systematic errors in the individual gradient
readings, this error of 3% in Vtop is remarkably good. It remains
to be seen whether this result survives when checked against
an absolute Vtop measurement - I suspect it will.

I've plotted the comparison between Terry's readings and tsim
predictions, having referred the measurements to ground, and adjusted
the predictions for the Fres shift and the actual coil input current.
The results are in

  http://www.abelian.demon.co.uk/tssp-data/nsvpi2210a.gif

The readings fall short of predictions by about 3% over most of
the coil length, as discussed above. Below about 2 inches the
readings exceed predictions due to the high driving voltage
necessary to overcome the actual Rs of the coil, compared with
the comparitively small Rs and driving voltage used in the simulator.

We can get a closer look by examining the voltage gradient dV/dx,
the comparison chart is in

  http://www.abelian.demon.co.uk/tssp-data/nsvpi2210b.gif

Again the predictions have been adjusted for the slight shift in
Fres, and the measured gradients have been aligned with the centers
of the probe positions. Clearly the predicted gradient peak (the
point of inflection in the V profile) at around 8" coincides nicely
with measurements in both position and amplitude. Below the gradient
peak the voltage rise is concave and in this region the measured
gradients are a little below the predictions. This is probably due to
a reduced external capacitance resulting from the partial ground plane
used in this test setup - the predictions were based on a continuous
groundplane. The extra Cext in the latter case would predict a slightly
higher dV/dx in this region. This seems to have a knock on effect all
the way up the coil, giving a slight skew to the measurements
wrt the predicted curve, and the measured gradient peak is probably
slightly above the predicted peak as a result, say 9".

All in all this is a very satisfying set of results. From my point of
view as maintainer of the tsim program it provides a much needed
confirmation that both the determination and the application of the
capacitance matrices is not grossly in error. The character of this
voltage profile lies somewhere between that of pn1710 examples 1 and 9,
and the most delightful part is perhaps the clear evidence of
the internal capacitance circulating currents at work here.

Terry, the overall simplicity and effectiveness of this method is
impressive, and I believe your results represent a significant
milestone in making visible the inner workings of the secondary.

More to follow later,

Regards,
--
Paul Nicholson,
Manchester, UK.
--


Maintainer Paul Nicholson, paul@abelian.demon.co.uk.