From: Paul
Date: Fri, 18 Oct 2002 14:35:47 +0100
Subject: Re: [TSSP] Top Voltage
Thanks Bert, I thought maybe I was making another mistake
there. Can't seem to engage brain today (:.
With this correction to fig 7.4 to redefine V, the tssp
sphere-sphere breakdowns agree with North, very well at
long range, deteriorating for short gaps.
When I compare these figures with
http://hot-streamer.com/TeslaCoils/Misc/SGapVolt.jpg
the north/tssp values for breakout voltage become quite
a bit less than the breakout given in this table.
Eg 2.5cm diam electrodes, 4.05cm gap
SGapVolt.jpg: 70 kV
North: 52.05 kV
tssp: 51.2 kV
this is about the worst case discrepancy, and things improve
as the two spheres become closer. I take it that the AIEE tables
are derived from measured values? If so, does this indicate
a systematic departure from the values expected purely on the
basis of a fixed surface field (north & tssp). Are we seeing
a long range effect that holds off breakout despite the surface
field reaching 30kV/cm.
I'm wondering if north/tssp predicts the onset of corona,
whereas AIEE tables show discharge voltages. For small
gaps the discharge follows immediately after corona starts,
but when the gap is wider, the initial corona 'softens' the
field gradient just above the electrodes, effectively increasing
the ROC a little, and allowing some further voltage rise before
discharge. Is that what we're seeing with this difference
between north and AIEE?
Extrapolating to the case of the small sphere on OLTC, this
could easily account for considerably higher breakout voltage
measured.
Is this the space charge effect that others have mentioned? If
so we seem to be acquiring some sort of numerical handle on it.
Terry,
Does the AIEE have spark-gap voltages for sphere-plane discharges?
For our purposes these would be much better. Any other good
sources of data?
The amount of space charge forming along a line of E-flux must be
some function of the field gradient along that line of flux.
Just that we don't know that function. If we could take a guess
at this, we could calculate the correction to the breakout voltage.
We could refine our guess by adjusting it first to match published
sphere-plane and sphere-sphere discharge voltages, and then against
measured TC breakouts.
Is this on the right lines? Perhaps we can make an initial
guess of how much space charge is formed per unit volume for
a given background E-field uniform in that volume? Maybe this is
in a book somewhere.
Then we just work out the location of peak surface field. Follow
a flux line out of that point, calculating the computed space charge
along the way, and subtracting the field that it induces at
the surface from the initial surface field. Well something along
those lines anyway.
Of course, it would be nice to have some evidence that space charge
is holding off breakout. It would have to show up as an increase
in the apparent C of the terminal. And maybe it would also leave
behind a negative static charge on the terminal, as we've seen.
So if any of this is realistic, we should be seeing signs of an
apparent reactance change and/or a DC residual, in the voltage
regime between the north/tssp predicted breakout and the observed
visible breakout.
Does this approach consolidate everyone's thoughts on the subject?
--
Paul Nicholson,
--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.