TSSP: List Archives

From: Paul
Date: Fri, 01 Jun 2001 20:03:39 +0100
Subject: Re: [TSSP] Racing arc clues

Hi John and all,

Probably most on this list are accustomed to reading your frequent and
considered postings on pupman,  and I've certainly learned a lot from
reading them, so its great to see you pop up on here!

Your comments have helped me to clarify things a little, and perhaps
you can check my reasoning below.

First, I think we need to recognise that racing arcs are simply a case
of the secondary breaking down before anything else. Then the advice we
might give to cure a case could be one or more of the following,

 a) Reduce the drive power.  (Not what you really want to hear).

 b) Use a smaller toroid. (Its lower breakout potential will ensure
    the topload breaks out before the secondary.)

 c) Add a breakout point to the toroid. (Again, ensuring that the
    topload breaks out first).

 d) Reduce the coupling. (The topvolts still reaches much the same
    ultimate peak value, just takes a little longer to get there, and
    we have a smoother voltage gradient on the secondary).

 e) Arrange a strike ring or some other earthed fitting to act as a
    'spark gap' limiter for voltage rise. (The toroid breaks out to
    this object before the secondary breaks down.)

 f) Increase the BPS. (So that accumulation of streamer breakout from
    the topload over successive bangs helps to keep the voltage down).

Have I missed anything out?

Item (f) is a little weak, because what happens on the first bang?
Initially there is no benefit from previous bangs so the secondary
gets the full punishment. Does (f) only work if you ramp up the power
with a variac?

I suppose that racing arcs are the inevitable consequence of taking a
well built and tuned coil on a quest for spark length by fitting larger
and larger toroids.

Referring to Richard Hull's coil, you could argue that he was using a
larger than 'allowable' toroid, in the sense that it was so big that
some other part of the system broke down first.  We could revise our
definition of 'maximum tolerable topload', or whatever,  to be the
largest toroid that the system can stand and still guarantee it will
break out before any other part of the system breaks down.

John wrote:

> Some folks say that large toroids actually prevent racing sparks

Wouldn't that be the case if the input bang size is not also raised?
ie if the larger top C was *not* matched by a corresponding increase
in primary C and instead the tuning was maintained by increasing
primary tap or turns.  Then the final topvolts and the coil voltages
would be reduced by this change, making the racing arcs go away.

Jon Tebbs wrote:

> is there any correlation between small diameter secondaries showing
> a propensity for racing arcs? I'm thinking along the lines of the
> _sharper_ radius of curvature.

Hi Jon. Welcome to tssp! That's a real good point you raise there.

Jon said:

> Thinking further out loud, the smaller wire diameter typically used
> is probably more of a contributing factor as the voltage per unit of
> length increases...

and Marc said:

> if a coil is wrapped with a larger wire...less turns...higher turn
> to turn potential...a too large topload would breach the threshold

which I think they are both right, it depends whether the volts/turn
remains the same (Jon assumes) when you change wire size, or whether
the total secondary voltage remains the same (Marc assumes). Chances
are neither would. I think the question we need answering (as far as
the modeling is concerned) is:  At what V/turn does a coil of a given
wire size and radius break down?  This is a 'localised' question, in
that it doesn't involve total volts or number of turns or voltage
profiles - it just asks a specific question about an arbitrary pair
of adjacent turns on a wound surface. Provide an answer to that
question, and we can then say

 a) at what bang energy a particular system will break down along the
 secondary - if at all - assuming it is properly built and tuned.

 b) what the maximum tolerable toroid is for that coil.

 c) for a given toroid and bang energy, what the highest safe k factor
 is.

(all for a given BPS).

That would be really something.  Surely the answer to this breakdown
threshold is sitting in a textbook on HV engineering somewhere? The
planet is covered in HV transformers, OK, most are not RF, but we can
surely find a starting figure. Do we simply use twice the insulation
breakdown value?  I did this for my coil, but I've no idea whether the
answer represents the actual breakdown limit for the thing - it would
be nice and straightforward if it did.

If we could collect statistics from coilers who can demonstrate racing
arcs - model each of their systems to estimate the likely V gradient,
it would give us a rough idea of the typical maximum stress that a
secondary can stand, and we might be able to see if there is any
correlation of this with coil radius, wire size, and insulation
break-down voltages.

There's some well-gathered info on the web site of Stefan Kluge,

 http://www.stefan-kluge.de/

see especially the four charts at,

 http://hot-streamer.com/stk/tc/sub1.htm#eff

It's hard to know what to make of these, but the reliance on spark
length doesn't help - we really need top volts instead.  There are
obvious patterns beginning to form, but they have quite a large
deviation. Do they tell us anything about the limits and capabilities
of the coils themselves, or are the patterns simply a reflection of
current trends in coil building, I wonder?

I think we need to take a closer look at Stefan's work - it's the best
performance survey I've seen so far, I think.

Bart wrote:

> ...a particular run where the coil ran just fine until I hit a
> particalar power level. At that moment, all hell broke loose and the
> coil profusely produced racing arcs...
> ...problem was tuning due to using a couple different toroids and
> not spending enough time in the tuning department

That will account for many cases, but we need to seek out the well
built properly tuned coils that are driven to the limit. Lets find
those systems, look at their limits, model them, and see what the
secondary gradients peak at.

Marc asked:

> in the case of too high a coupling, is it just the field that
> intersects the secondary from the primary? or could it be that the
> field from the upper side of the primary "and" the field from the
> lower side of the primary are interacting

No, only the part of the primary field which actually passes through
the secondary coil cylinder itself is effective in inducing volts on
the secondary. If a primary field line does not pass inside at least
one secondary turn, it will not be seen by the secondary, and will only
contribute to the primary's self-inductance, not its mutual.

> what happens when the lower field is brought into play on the
> secondary? if the parting line between fields is brought above the
> bottom turn of the secondary (or close there) wouldn't a form of
> bipolar coil be introduced?

Yes, as far as the induction is concerned, but the secondary base is
still clamped to earth, not free as in the bipolar.

Interestingly, I expect that k will reduce as you raise the primary
further up the secondary (above the current maxima), because the
secondary current is less - a moot point in view of the inevitable
breakdown, but I'll eventually get around to modeling the case just to
see.

Marc again:

> if primary field shape does cause a "localized" strong point...
> ...i would think this could be enough to create unequal gradients

Yes, indeed the lumpiness of the primary induction appears to be the
stimulation for the mode 3 (that's 3/4 wave resonance) excitation. If
you could contrive to have a smooth induction from the primary, in such
a way that the induction from the primary had the same profile along
the secondary that the secondary free 1/4 wave resonance has by itself,
then I think we'd see no higher modes at all.  The dominant modes, ie
1 and 2, (which are both 1/4 wave resonances but one has the opposite
phase primary current to the other)  should ideally have the same
secondary V/I profiles, ie the difference between modes 1 and 2 should
be just the primary current sign, and no other.  For practical
primaries this is unlikely to be possible - the induction will be
lumped towards the secondary base.  You've seen the mode shapes in,
say

 http://www.abelian.demon.co.uk/tssp/tmod.html,

As the primary induction is formed into a more and more compact lump,
we need to introduce higher and higher modes in order to match its
shape along the secondary.

> primary field shapes might have to be given more "visualization"
> and thought?

Definately. I'm beginning to suspect that there is one, and just one,
optimum shaped primary for a given secondary/topload pair, and that
this magic shape is probably not one that is feasible to use from the
point of view of breakdown.

Marc wrote:

> i personally try larger toploads until i only get 1-2 streamers
> breaking out at once, with a rough toroid like tape coated duct,
> this can be larger then a smooth toroid.

Those are the ones we should be measuring the breakout topvolts on, and
comparing them with ET6.12 predictions, but that will be another
thread.

Well, those are my thoughts on the subject, and I welcome any
corrections.  Meanwhile, back to some coding...

Cheers All,
--
Paul Nicholson,
Manchester, UK.
--


Maintainer Paul Nicholson, paul@abelian.demon.co.uk.