Re: Racing Spark Prediction

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Paul Nicholson <paul@xxxxxxxxxxxxxxxxxxx>

Dest wrote:  (I still think Dmitry sounds better :))
> this night i found the simpliest explanation from John
> Freau:

> http://www.pupman.com/listarchives/1998/July/msg00099.html

Yes, a good explanation.

John wrote:
> There is a TC operation mode that can produce a great number
> of racing sparks all over the secondary.  I don't remember
> anyone mentioning this but they may have.  One time I was
> testing a static series gap under relatively high power.  The
> gap was not cooled so it heated up fairly quickly causing it
> to either power-arc and/or quench poorly.  When the gap was
> still cool, there were no racing sparks.  As soon as the gap
> heated, a great number of racing sparks appeared.

We seem to be hearing stuff that suggests the primary gap
behaviour has something to do with racing arcs.  Couple of
questions:-
a) Did the racing arcs above cause any damage to the secondary,
   or where they faint, low energy sparks?
b) Can the phenomena be reproduced on some other coil?
c) Could the phenomena be made to go away by a small change to
   the primary tapping point?

> Any thoughts or speculations about why this might happen?

A few.  A tie-up between racing arcs and gap behaviour is a
rich source of possibilities.  We'd better go read Terry's old
paper mentioned in

> http://www.pupman.com/listarchives/1998/July/msg00019.html

Under certain conditions then, is the gap (nonlinearity)
generating lots of unwanted RF at some high frequency?  Is
there a -ve V/I slope involved here?  Driven by power arcing,
is the primary gap trying to be a Poulsen arc?

> I didn't actually scope the firing of this coil.  I think I've
> seen other folks' coils which did the same thing but it seems
> to be something that hasn't been discussed much or at all on
> this list.

John may have given us a handle on the problem here.  Too high k
will surely cause quenching problems, in theory and in practice,
which in turn may cause racing arcs.

Dest wrote:
> what are you talkin about - which other factors?

Factors other than reduction of k.  Reduction of k fails to
significantly reduce the HF modes excited by virtue of the
non-uniform coupling geometry.  I'm pretty sure now that
these 'geometrically' excited overtones are neither strong
enough, nor sharply enough affected by k changes, nor of
high enough frequency, to be responsible for racing arcs.
We must look at other things which can generate HF - things
which happen to be alleviated by low k.  Primary overtones
excited by gap non-linearity might be one such 'other factor'.
HF transients from topload discharge might be another, etc.

I'm keen on this 'pathalogical primary behaviour' idea,
associated with quenching difficulty which in turn is brought
on by 'over-coupling'.  It fits - k is not directly the cause.

> i see, that all 10 years starting from appearance of D.C.
> on this list, he is suffering from old misconceptions about
> coupling, standing waves, e.t.c.

Yes, true, but he's in good company.  I think most coilers will
only offer some partially faulty explanation of coil resonance.
After all, not many study theory in much detail, their enthusiasm
and expertise is for construction and operation.  Similarly, ask
ten average radio hams how their antenna works and you may get ten
different explanations.   That's no worry so long as someone
wanting a clear, detailed, and correct explanation can actually
find one if he wants.

> i think that you guys must help him to understand the very
> basic things about how resonance transformers works, coz this
> is the one of the purposes of this list, isn`t it so?

And we do - we all help fill in the gaps in one another's
understanding of the subject.  But nobody can take everything
in - we all have our specialities.

Okay also your reservations about info imported from ham radio
sources.  Often the context is quite different so care and
frequent referral to first principles is needed.  You hit the
nail on the head with

> there is no info about pulse mode of operation
> with limited amount of energy

Yes, a finite energy system, not a steady state forced response
situation, is the root of many confusions with respect to
things like standing waves, impedance matching, etc.

Peter Terren wrote:
> I can contribute with a picture of racing sparks.
> http://tesladownunder.com/TeslaPhysicsRun5secs.jpg

This looks like a point-to-point breakdown of the secondary,
which I think is probably a different thing.   I'm assuming
that the typical case of racing arcs would be described as
small arcs racing at random all over the secondary, oriented
vertically, short, not very bright, and not pinned to any
particular breakout points on the secondary.  Maybe someone
has a photo of something like that?

Bob wrote:
> Better attenuation relative to the fundamental can probably be
> achieved with L in parallel with R. The idea being to bypass
> the R at low frequencies. The practical problem is probably
> constructing the L.

Yes L in parallel with R would do nicely wouldn't it.  I don't
think the voltage across this parasitic suppressor would be very
high, not enough to give problems with breakdown.  If the
primary gap is generating HF RF during power arcing, there may
be quite a lot of power dissipating in R, much more than if
just a small percentage of bang energy was being scattered
into HF.  Maybe the primary circuit is the best place for the
RL suppressor - the resistor placed across suitably chosen
turns, say, to dampen specific unwanted modes.   But, discussing
the cure is premature...

Gerry wrote:
> What measurements do you have in mind and what type and quality
> equipment is needed for the measurement.

Capture of a secondary base current waveform on a digital scope,
transferred to a computer file for analysis.  Only rough amplitude
calibration needed.  Wide bandwidth, though, say up to 30 times the
TC operating frequency.   One sample of the genuine phenomena is
all that's needed, so no long damaging runs required.  We would
analyse the captured waveform to determine what signal components
are present - chances are the offending signal(s) will be quite
prominent.  We would work out their voltage and current
distributions and approximate gradients, and by modelling,
determine which parts of the structure are involved in generating
them.   The cross check at that point would be detection of the
offending resonances during low power sweeps and showing by probing
and loading that the nodes and antinodes are in the predicted
places.  The tricky bit is getting a coil to display classic racing
arcs on demand.  Maybe John has pointed the way there.

Variation on the above:  Just fire the primary under various gap
operating conditions, without the secondary (is that possible
or would that damage something?).  See what frequencies are excited
and at what levels.  Knowing that, we can calculate how the
secondary would cope with them.

Dest wrote:
> i`m reading archives from the begining

I hope you're making a list of all the 'good' posts!

Thanks everyone for all the interesting comments in this thread.
--
Paul Nicholson
--