From: Paul
Date: Fri, 17 May 2002 18:59:40 +0100
Subject: Re: [TSSP] Racing arcs
...still catching up on emails...
John wrote:
> when the frequencies split due to overcoupling, the lower primary
> split frequency currents tend to be in phase with the secondary,
> but the higher split frequencies tend to be out of phase with the
> secondary.
Ok, the single resonance of the isolated secondary splits into two
with even the tiniest coupling - it doesn't require overcoupling.
Those are the two resonant modes which make up the beat envelope.
The two modes differ in phase in the sign of their primary currents with
respect to their secondary currents. You can't say in an absolute sense
that one is in-phase and the other out of phase - it depends how you
choose to label the current directions in each winding. If you assign a
'dot' to the correct end of each winding, and label the currents as
flowing into the dotted terminals, then the relative phases of the
currents are as above.
> This could fit in with what Malcolm said perhaps, about
> tuning lower to emphasis the lower frequency. This would
> reduce the out of phase higher frequency component, and
> perhaps lessen voltage stresses that may cause racing
> sparks?
If the primary is tuned to the sec Fres then the bang energy is
almost equally shared between the two modes, giving them equal
amplitudes so that almost complete cancellation of the primary
current and voltage occurs every 1/(f2-f1) seconds, thus creating
a sharp notch and placing all the energy for a moment into the
secondary.
When tuned low or high, the two modes have different amplitudes and
no longer achieve complete cancellation of the primary current or
voltage at the notch. Thus there is always a proportion of the total
energy retained in the primary. Therefore generally speaking the
secondary voltages are going to be lower everywhere when the primary
is detuned either way. This may not be the case with the pri-sec
voltage. The detuning (either way) alters the relative phase of
the primary and secondary voltages during the risetime of the beat
envelope.
See
http://www.abelian.demon.co.uk/tssp/pn040502/tfsm1-h1.wave.gif
for an in-tune set of k=0.2 waveforms. Compare the primary volts
and secondary volts over the first 10uS - they are nicely 90 deg out
of phase. Then turn to
http://www.abelian.demon.co.uk/tssp/pn040502/tfsm1-hb.wave.gif
for the -20% tuning. Notice how in the first 5uS both the primary
and secondary voltages are simultaneously large. Given a bad luck
choice of winding directions you could end up with quite a large pri-
sec voltage difference in the detuned coil. Consider this in the
context of Boris's potato shaped profiles.
Note that when the same coil is tuned high, as in
http://www.abelian.demon.co.uk/tssp/pn040502/tfsm1-h9.wave.gif
the relative phase of primary and secondary voltages have altered. So
if you weren't getting pri-sec breakdown when tuned -ve, then you may
do when tuned +ve.
> I was thinking that maybe these out of phase conditions are causing
> an extra voltage stress between primary and secondary which is
> promoting racing sparks near the bottom of the secondary?
I think you are right there. I guess some waveform plots of the
pri-sec difference voltage for the various tunings would do for a
start. My next job is to make a graph comparing those red-lines,
and to do some dV/dx animations, so I'll calculate the pri-sec also.
--
Paul Nicholson,
--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.