From: Paul
Date: Sun, 05 May 2002 14:42:59 +0100
Subject: Re: [TSSP] Secondary voltage stress factor
I wrote:
> One very odd thing is that the resulting step is so small.
I've fixed this. The model was only working with the lowest 22
resonant modes - not enough of a basis to reconstruct the response
to such a sharp transient. The model was just doing the best it
could with the available modes.
I've re-run the response using 60 modes, and the plots below
start at the foot of the final 1/4 cycle rise before the discharge,
so we've skipped a couple of cycles from the start of the bang.
http://www.abelian.demon.co.uk/tssp/pn040502/tfsm1-h1d2.wave.gif
The transient is now almost of the expected size. The animation
is
http://www.abelian.demon.co.uk/tssp/pn040502/tfsm1-h1d2.anim.gif
which is long enough to see the pulse travel down the coil, reflect
from the bottom, travel back up and reflect from the top.
The model is predicting that the pulse maintains a substantial
amplitude all the way down the coil. Dispersion due to different
velocities of its components is scattering the pulse, so that by
the time is has travelled the length a couple of times, the coil is
awash with small wavelets reverberating back and forth.
Note the initial drop in voltage near the top of the coil in the
frame straight after the discharge - this is due to the coil-toroid
distributed capacitance which is able to couple the transient
directly to all parts of the coil, although the most effect is near
the top.
The question is - are the discharge transients really this big? If
so, why don't we see sparks popping out of the secondary whenever
the topload discharges?
This is one of those situations that calls for an experimental test.
Something like a solid state gap, modest bang voltage, earthed
electrode brought up to the toroid for a low voltage discharge.
Base current CT, and something to pickup the top volts. We can look
at the size of the base current transient which occurs 1/4 cycle
after the discharge to see if it's consistent with the model.
One possibility is that the nature of the simulated discharge is
unreal. The program simply removes the stored charge from the top-
load in between two time steps of the model, ie within around 30nS.
Thereafter the top voltage is no longer clamped to earth and is free
to rise again - perhaps in the real coil the discharge can continue
for much longer, clamping the topvolts down, and altering the
characteristics of the transient.
--
Paul Nicholson,
--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.