[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: the cure for racing sparks



Original poster: "Paul Nicholson by way of Terry Fritz <twftesla-at-qwest-dot-net>" <paul-at-abelian.demon.co.uk>

Bert Hickman wrote:

> What model did you use for the arc to ground?

Just a 50 ohm resistance.  Thus the modeled discharge is an RC
decay taking around 100nS. This is only a couple of the time steps
used in the animation (50nS) so essentially it's a step transient.

In a real coil the collapsing topload charge might oscillate
back and forth through the arc at some HF or VHF frequency
determined by the resonance of the top capacitance with the total
path inductance of the arc, which includes the ground return 
inductance.

This may have the effect of softening the transient, by stretching
it out in the time domain, but, on the other hand, if the discharge
frequency just happened to coincide with a secondary mode...

> the full pre-arc voltage voltage, plus a bit of VHF oscillatory
> overshoot, is now present across the top ~3% of the winding

Yes, although the peak voltage of the transient doesn't seem to 
exceed the topvolts anywhere, it does rise and fall steeply over
a short coil distance, giving rise to a high vertical surface
gradient moving right along the coil.  Dispersion seems to deal
with this after a cycle or two, but as you say, the initial
transient tries to place the entire top volts across just a few
turns.  Just how many turns depends upon the steepness of the
voltage fall, and this in turn depends on the top-C * discharge-L
time constant.

This is a case where close proximity of the toroid to the winding
top helps a lot.  The factor involved here is the toroid-coil
distributed capacitance.  If this is large enough between the
toroid and the top few turns, the transient is coupled through this
C and the turns don't see the full voltage.  You can see this
happening with frame 6 of thor.anim2.gif, which is the first frame
after the discharge begins and we see that the top voltage is about
half way through its RC decay.  Note that the coil voltage is
immediately depressed for about 20-30% of its length.

I'm pretty sure the 50 ohms is much too low - Malcolm has described
the secondary ringing decaying in a cycle or two following a
topload discharge, so the effective resistance is more likely to
be many K ohms.  I left it low deliberately so that we could see
more clearly the post-discharge ringing of the secondary.

With these comments in mind, it looks as if the modeled transient
is a worst-case in that the simulated discharge is steeper than
those we might expect to occur in practice.

Thus we now have two suggestions why lots of C between toroid and
coil is a Good Thing(tm).
a) It increases the Les of the secondary (unfortunately also the
   Rac too, by the same proportion). 
b) It allows the transient to be spread out to some extent along a
   good length of the coil.

Bert, what should we expect from of a typical topload discharge?  Does the R
dominate to give mainly an RC decay with perhaps a bit of overshoot, or does
the L make a big appearance to give a sinusoidal ringdown?
--
Paul Nicholson
--