Re: OLTC update

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

Original poster: "Paul Nicholson by way of Terry Fritz <twftesla-at-qwest-dot-net>" <paul-at-abelian.demon.co.uk>

I guess the main difficulty with the OLTC is that a large primary
cap is required in order to get a decent firing energy.  Thus an
impractically small primary inductance is called for.  

Barry wrote:

> Would it be easier to build a secondary that couples to a
> higher harmonic of the primary?

Hmm.  Putting the energy into a primary (or any coil) by 
connecting a charged cap across its ends tends to excite the coil
at its lowest resonant mode, which is what we'd call its Fres.
Only a little of the firing energy goes into higher modes
of the primary. 

To excite the primary at a higher mode (not a harmonic), one way
would be to simultaneously discharge several caps, strategically
placed along the winding, alternated with ground connections. 
The initial firing energy would be shared across all the caps.  
This would seem to be feasible given the use of semiconductor
switches.

Apart from the difficulty with tuning, so far so good. 

The problem then is to couple effectively to the secondary. Now
with the primary at a higher mode, current in parts of the primary
will tend to be pulling against current in other parts of the 
primary as far as the inductive coupling to the secondary 
quarter wave is concerned.  And as the mode number gets higher,
the overall cancellation of the field from the primary becomes
more complete.

Therefore, it would be necessary, in order to obtain a cooperative
field from each half-wave current segment of the primary, to
reverse the winding sense between each firing cap.  But doing so
would destroy the transmission line properties which were giving
you the particular mode, and the system degenerates almost into 
a set of parallel primary windings, each with its own personal cap.

So to persue Barry's rather interesting idea, I think it would be
necessary to make sure that the secondary was closely coupled to
only one half-wave's worth of the primary, and that the rest of
the primary contributes relatively less to the coupling.  Acmi
can be used to model the coupling to a secondary from a primary
which has caps placed along it's length, so long as these caps are
large compared to the self-C of the primary itself, which it's
safe to assume. 

Note that this might not be possible with traditional spark gaps
because of the need for simultaneous switching. 

Apart from tuning, I think the other main problem with this
approach would be how to ensure that large amounts of energy don't
converge onto a single cap of the primary, which it's perfectly
free to do so once the thing is released into resonance - given
unpredictable reflections from the top of the secondary.  We
wouldn't want any of the caps to suddenly find itself carrying
a whole lot more energy than it started out with.

But then we do have scope for controlling the timing of the 
switching, so maybe things could be controlled automatically
somehow.  If you get into that sort of territory, you can 
conjecture building up a traveling pulse waveform along a primary,
by firing caps consecutively, so that the energy is built up
into a single broadband pulse.  You'd set things for a steady
impedance transformation along the way, and the primary would
couple to or merge into a secondary, which would continue the
impedance transformation (by some cunning choice of coil profile)
to give a single huge voltage pulse at the top.  Picture a deep
atlantic wave approaching a beach, rolling up in the shallows,
trading width for height.  The deep sea wave sees a gradual
impedance change as the beach applies some extra boundary
conditions.  Surfers know the beaches with the best profiles -
wonder what shape of coil would turn a long duration, low-voltage,
high-current, broadband pulse applied at one end, into a short,
high voltage, low current pulse at the other.  Note that this is
quite unlike a normal TC.  We're now firing up multiple resonant
modes, and timing them so that they momentarily converge to a
single giant voltage pulse at one end (preferably the far end)
of the coil.  But before anyone gets too excited, there are some
problems with this approach and it's not likely to give you any
more topvolts per joule than a regular TC, if my sums are correct.

But, I think I like Barry's idea, because it opens up some 
interesting possibilities.  Whether it would bo any better than,
say, a normal primary with lots of separate caps and switches
paralleled across it would have to be considered.  All in all,
I'm following this OLTC stuff with great interest.
--
Paul Nicholson
--