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

Re: Racing Spark Prediction

Original poster: "Dr. Resonance" <resonance@xxxxxxxxxx>

It really doesn't matter exactly what the measured coupling is.

Just run the coil up and down until you find the "sweet spot" at lower power levels (35-40%) and you have the longest possible sparks --- hopefully the goal here.

I've done this very procedure with several different size coilforms and the results are always within 5%. After you achieve the performance that is max for your particular setup, then measure the coupling if you desire. The goal is optimum performance but not at the expense of excessive standing waves on the sec coil --- the cause of "racing sparks".

Dr. Resonance

> My hunch is that overtones have something to do with it

Dmitry wrote:
> then how about "folded resonator" concept?

Bob wrote:
> Can you explain why a folded resonator may cure racing sparks?

Dmitry wrote:
> why are you asking me about ways to stop racing sparks?

Hey you guys, go easy!  These sound like fair questions if we
suspect HF signals are causing racing arcs. Folding - meaning it
appears,  we wind some of the secondary in reverse direction to
the rest - is worth putting on the table.  If a resonant mode
is identified as responsible, we must relocate it or dampen it.
Changing the reactance distribution by folding might be one
possibility, but until we know more about which HF modes are
responsible, if any, I don't see how we can pursue that idea.
Let's identify the cause before considering the cure.

Dmitry wrote:
> can your simulator model this situation?

Yes, with just a small alteration to allow us to select which
portion(s) of the secondary are reverse wound.

> this software is downloadable, GNU-able, can i compile it
> for myself?

Yes.  But at the moment there are some incomplete changes. The
main defect in tssp software lies in the process of solving for
the quadratic eigenvalues and for the eigenvectors of the large,
dense, non-hermitian 'solenoid matrix' which is defined in


The process used at the moment is slow and brute force, and
requires some hand-holding of the algorithm: selection of starting
parameters so that the thing converges and also does not miss any
modes, especially closely located ones during weak coupling and
highly damped modes during conditions of heavy loading.  I'm slowly
learning enough math to tackle the problem properly.  That's why
pn1401 above is unfinished - this key step is too embarrassingly
crude at the moment for me to document and publish!  Better ways
need to be found.  I'll write more about this if anyone is

Dmitry wrote:
> do you really suppose that his [D.C's] coils were such horrible
> quality or were coupled well below 0.1?

No, neither.  I assume the coils are good and the coupling is
modest but not super low as D.C.'s setup procedure would suggest.
I suspect some combination of factors is upsetting the sweep.

D.C. wrote:
> No racing sparks here in over 30 years!!

So he must be doing something right, we just need to understand
exactly what and how.  Simple reduction of k doesn't give the
answer - the HF modes are not reduced sharply enough. There must
be some other factor, such as reduced coupling between primary
overtones and the secondary, etc.

Guys, we are running out of things to say on this topic. Most of
these points have been stated before. We are stalled, needing
some measurements of coils near to producing racing arcs so we
can actually see what signals are present and at what level.
I feel that lurking in this topic is something new and useful
waiting to be discovered, hidden not too deep - and we have the
technology too!  We 'theorists' can only speculate ad nauseum
until this ball returns from the experimenter's court.
Paul Nicholson