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

Re: the cure for racing sparks



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

D.C. Cox wrote:

> A second note is that you may be "overcoupled" which causes the
> "double-hump" on the freq, ie, splitting the res. freq. into two
> different freqs which beat against each other. 

This is another myth, which seems to have been quite popular
during the early 90's.  The idea that mode splitting occurs when
the coupling reaches a certain level is false, and the presence
of the double hump in the frequency response doesn't imply
overcoupling.

The quarter wave resonance of the secondary is 'split' into two
separate resonances by the introduction of a tuned primary, and
this occurs no matter how small the coupling.  The two frequencies
are Fres/sqrt(1+k) and Fres/sqrt(1-k) for high Q coils where Fres
is the resonant frequency of the uncoupled coils.  At normal TC
coupling coefficients, the mode frequencies are offset around 10%
each side of Fres, and as k is reduced to zero the two mode
frequencies converge smoothly to Fres.  

The two modes thus produced share about equally 95% or more of the
firing energy, creating a beating envelope as the two decaying sine
waves alternately cancel and reinforce every 1/2/(F_upper-F_lower)
seconds.  For example, see waveforms from Marco's Thor system [*]
at k circa 0.17 in

 http://www.abelian.demon.co.uk/tssp/md110701/

Far from being a problem, the 'beat against each other' effect is
an essential feature of the energy transfer mechanism of a TC.

It is often demonstrated that reducing the coupling coefficient,
even just slightly is enough to cure a case of racing arcs. Why
should this be?  

Well the following paragraph is one hypothesis:

In a period of half a beat, the firing energy is transferred
completely (less losses) to the secondary.  Increasing the 
coupling coefficient reduces this period. In the example above,
the transfer takes 3 cycles, during which streamer formation is
taking place.  In coils that are breaking out, we expect that the
topvolts is to some extent clamped by the draining of charge into
streamers and this clamping can only occur if streamer formation is
fast enough to take up all or most of the charge arriving from the
secondary. Now if the k factor is increased from 0.17 to 0.24, the
transfer time reduces to only 2 cycles, and streamers therefore
have only 2/3rds the time in which to form.  If streamer formation
isn't quick enough to accept all the displaced secondary charge,
then voltages will rise significantly above the lower-k values,
even for just a small increase in k.  Reducing the k allows 
streamers to develop further over more cycles, thus taking more
charge and clamping the top voltage more effectively.  There is
evidence to suggest that a topload with a full development of
streamers can have 20% or more effective capacitance.  Thus if
the streamers aren't given enough cycles to develop, then V may
have to rise 20% higher to accomodate the same charge.

Conversely, as the coupling is made lower, more energy will be
wasted in losses.  Therefore for any system there should be an
optimum k factor for best spark length. Too high = not enough time
for streamer development = many short streamers, with risk of
racing arcs.  Too low = few streamers, longish perhaps, but feeble.

Another thing to look out for is that it may be necessary to open
out the primary tap to retune as k is increased and the topload is
breaking out.  The streamer loading increases the effective
topload C and lowers the secondary Fres.  If the primary remains
tuned to the original Fres, then there is incomplete energy
transfer and perhaps 10% or more of the firing energy is wasted by
being left behind in the primary.  When the primary is tuned a
suitable amount on the low side of the secondary Fres, the
streamer loading can bring the two into alignment and the complete
energy transfer takes place. To see if this is happening in any
system, tune for best performance, then measure the primary and
secondary resonant frequencies (or the mode frequencies if you
prefer not to remove the secondary).  Another good reason to allow
a few extra turns when designing a primary.
 
As regards the sphere, any topload will tend to break out from the
point on its surface which has the highest field strength.  In the
case of an isolated sphere, this could be anywhere.  But put it
atop a secondary coil and the highest field strength could well
be just where you don't want it - where the secondary meets it!
A sphere just doesn't do a good job of directing the breakout to a
desirable place.

One advantage of a toroid is that the highest field strength occurs
at the outer rim, and the area where the top of the secondary meets
it is a fairly quiet backwater as far as the field strength is
concerned.  Another advantage is that, for a given breakout
threshold voltage, a toroid will hold more charge than an
equivalent sphere, so that when it does break out, it will send
out a longer streamer.  Toroids are rather nice for this job, you
have two degrees of freedom in the design - thickness and diameter.
Roughly speaking, choosing a thickness sets the breakout voltage
(by setting the minimum radius of curvature), and choosing a width
determines the charge available for streamer formation (by setting
the capacitance).

Steven Ward wrote:

> It takes a few seconds, then WHAMMM!!! a
> huge 28" arc will jump out into the air. 

Your existing radius of curvature (6") seems quite large compared
to that of toroids typically used with this size and power of coil
(say 4-6" thickness = 2"-3" ROC?).  Therefore your secondary is
doing a pretty good job of containing the voltage applied to it in
the presence of an unhelpful topload.  So I'd recommend switching
to a toroid.  This way you can reduce the peak voltage reached by
the system, and at the same time maintain or increase the available
charge.  You might well find that for best operation, the primary
ends up being tuned 5-10% below the secondary Fres.

> i use a sphere (12" diameter) because its the coolest damn
> thing ive ever used for a topload!

Hmm, sit the sphere on top of the toroid?

[*]
 http://personal.inet.fi/atk/dncmrc/thor.htm
--
Paul Nicholson,
Manchester, UK.
--