Re: Tesla Coil Blunderbusses

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "R.E.Burnett by way of Terry Fritz <twftesla-at-uswest-dot-net>" <R.E.Burnett-at-newcastle.ac.uk>


Hi Malcolm, all,

I ran a few simulations on Microsim today to investigate Malcolm's theory
that the conduction time of the static spark gap could influence the
resonant charging behaviour.  All the gory details follow...

I modelled a 10kv/100mA NST driving a 10nF capacitor from a 50Hz supply.
Gary Lau's "Static gap model" was used to fire at a fixed voltage and
conduct for an adjustable duration (quench time.)  I did not model the
oscillatory behaviour of the TC primary or secondary -  just dumped all
of the capacitor energy into the switch.  (The simulator doesn't mind ;-)

The aim was to see if a change in the quench time would influence how
smoothly the static gap would fire.  I specifically chose a small tank
capacitor to minimise the resonant voltage rise,  (otherwise the gap would
still fire at 80kV !)   I also believe that a smaller tank cap would have
its terminal voltage influenced to a greater extent for a given change in
the stored ballast energy.

The firing of the static gap was very erratic,  and there were certain
bands of breakdown voltages for which operation was regular and others
where it was totally chaotic.  I _eventually_ managed to set the breakdown
voltage so that the static gap would continue to fire if kicked with an
initial breakdown.  However if the gap was to stop firing,  it could not
restart by itself.  This is a condition I have observed in practice, and
I believe this is the precarious balance that Malcolm intended ???

I have posted two GIF files showing the results of two simulation results:
	www.staff.ncl.ac.uk/r.e.burnett/mw01.gif
	www.staff.ncl.ac.uk/r.e.burnett/mw02.gif

The mw01.gif graph shows the firing of the static gap with a breakdown
voltage of 28kV and a conduction time of 100us.  The gap is "triggered"
at 5ms and fires smoothly until t=42ms.  From t=42ms onwards the static
gap does not fire as there is insufficient voltage to cause breakdown in
this particular arrangement.  (In practice the transformer would just sit
there growling.)

The mw02.gif file shows exactly the same setup with the conduction time
increased to 150us.  The spark gap is triggered again at 5ms,  but this
time continues to fire smoothly until the end of the simulation !

Overlaying the two pictures does indeed show a slightly faster rate of
charging in the case with a prolonged gap conduction.  Exactly what
Malcolm predicted.   See: www.staff.ncl.ac.uk/r.e.burnett/mw03.gif

(I repeated this sim a few times with different component values and
breakdown voltage.  It was often possible to affect the stability of the
static gap by altering its quenching.)  If anyone would like to verify
this behaviour with different component values then I have put the
Microsim schematic on my web page:
	www.staff.ncl.ac.uk/r.e.burnett/mw01.sch

I also ran a simulation with a 200BPS sync rotary in place of the static
gap, because I'm not sure what is happening here !   In this case changing
the conduction time from 100us to 150us (or even 200us) made no
significant change to the power throughput.  The peak voltage actually
dropped by around 1% for the later quench.

Personally,  I would not like to say whether this behaviour is really
significant in terms of practical spark performance,  but the simulation
does prove that gap quenching can influence the stability of a widely set
static gap.

One additional consideration....  (Just to muddy the waters further ;-)
In practice I have seen the breakdown voltage of static gaps vary wildly
during repetitive firing.  I once used a storage scope to capture a half
second of gap voltage in my coil and the breakdowns occurred at a wide
range of voltages !  When tested with a one-shot method the breakdown
voltage was uniform and repeatable,  but dropped considerably when running
some real power.  I guess this effect would also act to maintain the
light in an already active gap.

I think the decaying performance often reported from overheated TCBOR gaps
may be due to this reduction in firing voltage rather than due to actual
loss of quenching ???  You can hear the gap scream as the repetition rate
increases ???

Well it's a big subject,  and I have rambled to much already.
I'll listen closely for others thoughts on this...

							Cheers,

							-Richie,

							(Newcastle, UK)