New NST-sync tests, was LTR Charging vs. Firing Time

Terry, Richie, Malcolm, all,

I figured it was time for me to return to an NST based sync gap 
system since my speculations about NST charging were not fitting 
real world data.  Now I'm seeing an NST LTR system mystery.

I decided to set up the system to be similar to Terry's, except I used
a 12kV, 30ma NST, and a .0147uF cap.  This cap is about 2.2 times
the reso-cap size, which is the same ratio that Terry uses, I think.
I figured this should give scope results that are similar to his.

When I first installed the NST and tested the system, the first thing
I noticed was that it ran very poorly.  At the usual gap phase setting,
the waveform dipped a lot (?), then fired while it was strongly rising,
well before the cap voltage peak.  The power factor was bad and the 
operation was inefficient.  It needed a radical shift in the
sync gap phase.  My PT system ran best at around 60 degrees
ASVP (After the Supply Voltage Peak).  Terry's works best at 75
degrees ASVP.  This new setup worked best at 135 degrees ASVP.
At this setting, the voltage gradually rises then peaks, and fires at
or slightly after the peak.  At this setting, the voltage rise, the efficiency
and the power factor (with PFC) seem as good as with the PT.  If I
set the phase a little earlier, the voltage rise is greater, but the gap
stops firing suddenly as the variac level is increased.  If i set the
phase a little later, the voltage rise is less.  I did not expect this
135 degree ASVP phasing requirement.

In any case, I tried adding some external ballast, but the more
I added, the more the output decreased.  This agrees with Terry's

I also noticed a wavyness in the charging waveform, I'm not yet sure
if this is real, or just some type of interference or meaningless

I'll do more tests to try to find a good gap phase setting near 60 or 75 
degrees ASVP that causes a small dip, and then a proper rise.
Maybe I missed it.  Here are more speculations:

1.  There may be two modes for LTR sync operation?  In this new mode,
the firing point is quite critical.  Imagine what this would do using a
static gap?  The efficiency and power factor would bounce all over
the place, maybe this is why this system doesn't work well with a
static gap.  There is really only one phase position where the coil
works well, the likelihood of the static gap firing at this point is poor, 
especially since the firing should occur after a voltage plateau.  The
sync gap permits this, the static gap won't.

2.  Why does Terry's system work more like the PT setup?  Does
the greater robustness (or the nature of the shunt) of a 15 - 60 NST
make it behave more like a PT?  Could this explain
why Gary's static gap system works well and mine does not?
Does the gap phase need to be moved to this new regime when 
the cap gets too large even using a 15 - 60 NST?

3.  Are these new results real or did I make a mistake in any of the
above work?  What causes the need for this new phasing regime? 
Do the computer simulations show effects such as I am seeing?

John Freau