Re: Help with difficulty finding overtones or nodes?

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

Original poster: "Paul Nicholson" <paul-at-abelian.demon.co.uk> 

Phil LaBudde wrote:

 > No topload: ~400kHz resonance. With 3 different toploads
 > resonance drops to 278, 247, and 170kHz. If the overtones
 > should then be within the 600kHz range of my oscillator,
 > I still couldn't pick them up ...

 > ...I've got three so far: 4x19", 6x24", 8x36".

Thanks for the extra info.  I couldn't get my model of your
coil to go down to 400kHz unloaded, so maybe you've got a
tail of wire coming off the top, or something.  Lowest I
could get was about 460kHz.  Anyway I'll continue with it...

With a little calculation we get (kHz)

             Unloaded    4x19   6x24   8x36
1/4 wave     461         281    245   195
3/4 wave    1139         938    923   918
5/4 wave    1666        1516   1512  1517 [*]

(I couldn't get 170kHz either for the 8x36. Something wrong
  there - it shouldn't be so far out.  Was anything attached
  to the 8x36 topload?).

 >>  Set your generator to do square waves and try to excite
 >>  the coil with harmonics.
 >
 > Hmmm... I think I'd need to obtain another generator.
 > Vintage sine-wave only vac tube oscillator here.

Not necessary - you can probably clip the output waveform
by various means, eg by heavy shunt loading of the output
terminals, or by using a resistor and a diode or two.
Anything that distorts the sine wave will do.

 >> You would need to look for
 >> coil responses through the scope probe,
 >> not the oscillator voltage.
 >
 >  Shouldn't operating at an overtone still cause a change in the
 > resonator's impedance?

Yes, but only at the frequency of the overtone, so if you're just
looking at the total voltage you may not notice the slight dip
as the coil draws current at one of drive the harmonics.  Also,
you may not be sure which harmonic is exciting the coil... if you
use the scope probe near the top of the coil to pick up the
resonance, the overtones will be sharp peaks of pure sinusoid
and you can check against the scope timebase to see which harmonic
you've got.

 > From discussion here, I understand that everything is truly
 > "distributed", and "lumped" is a convenient model. You could
 > probably argue that almost any useful model for any system is
 > "lumped" in some way, but that's why we have models! On the
 > other hand, if it doesn't behave "distributed" enough to warrant
 > worrying about it, the "lumped" model is just dandy.  "Lumped"
 > certainly seems good enough for the average coiler to get up
 > and running.

Absolutely.  Even our so-called 'distributed' models are really
just elaborate lumped models that approximate the distributed
reality by using lots of small L and C elements.

 > But I'm curious about exploiting the "distributed" effects,
 > if they are of any significance.

Me too, but I can't think of any real benefit to the practical
coiler from all our theoretical musings on distributed models.
There's nothing coming out of the models that tells coilers to
do anything different to what they're already doing.  But it's
nice stuff to ponder though, if that's your cup of tea.

 > I don't see how one has choice over the mode of operation,
 > other than building the secondary differently.

Well there physically isn't a choice to be made at all.

 > Speaking of efficiency, the Corums say to measure VSWR as a
 > guide, not spark length. But even if I concentrated on
 > measuring SWR's, what relation (if any) would they hace
 > to spark lengths?

Not much, directly.  The VSWR in question is just the ratio
top volts divided by base volts.  This turns out to be equal to
the Q factor of the resonator, so you can talk in terms of VSWR
or Q, take your pick, but VSWR only makes sense when you're driving
the coil through its base terminal.  So I'll use Q to be more
general.  There are two Q factors:- a high Q when the coil is
not breaking out, and a lower Q when loaded by its streamer
discharges (or whatever the load is).   The efficiency of energy
transfer into the load is just the usual

  100% * (Q_unloaded - Q_loaded)/Q_unloaded

Other things being equal, the greater spark length would be
obtained with higher efficiency, so you want the highest unloaded
and the lowest loaded Q factors possible.  That's about as far
as it goes for Q or VSWR.  There are many inter-related factors
affecting spark length and we don't really understand it very
much . . . yet.  Q is really just a small factor.

 > But the question is always, "What is gained by going through
 > the trouble of the more complex/accurate model and is it
 > worth the effort?"

Good question.  For me, the theoretical modelling is pretty much
and end in itself.  There's a certain satisfaction to be had in
correctly predicting some of the more subtle coil behaviours by
feeding laws of physics into a computer. I like it when you line
up measured and predicted waveforms and see that even the little
squiggles match up.  Weird, eh?

 > If I understood correctly when I first read that, when a
 > topload is added, it obviously drops the fundamental resonance,
 > but the overtones are still left way up there.

Ok, you pick this stuff up pretty quick,

 > does that mean that even with my largest topload dropping
 > the fundamental to 170kHz, the overtones are going to remain
 > out of the 600kHz range of my current oscillator?

Yes, as you can see from the frequencies predicted for your coil
with/without topload.  Adding top loaded can only truncate the
top quarter wave of each resonance, so the 3/4 wave can only really
drop by at most 50%, the 5/4 wave by 20%, and so on.

 > I have a feeling that an efficient secondary is going to be
 > pretty "lumped".

Yes, indirectly.  Efficient operation means good coupling to the
load, which means a low characteristic impedance for the resonator,
which requires a large topload, and this as you see pulls down the
fundamental much lower than the lowest of the overtones.  So the
result is very lumped looking operation because the overtones are
way up there out of the way.

 > Thanks for the help!

You're very welcome.  I'm relieved to see someone avoid getting
stuck on a wild goose chase after a non-existent alternative mode
of operation!

Good luck with the measurements - I'm sure if you can roughen up
that generator waveform you'll get enough harmonic content to see
the overtones - at least you know where to look now! I hope you'll
report back and tell us if we're right or not!

[*] You want to know why 5/4 wave for the 8x36 is a little
higher than the others?  Well to try to get your 170 kHz I
lowered the top load way down over the coil top to increase its
capacitance to the ground plane.  When so low, the shielding
effect over the top quarter wave of the 5/4 wave resonance
actually relieves it of some of the distributed shunt capacitance,
raising the frequency a little.

--
Paul Nicholson
--