Re: Fo shift issues with spark C loading? was, Tesla Coil Blunderbusses

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

Original poster: "by way of Terry Fritz <twftesla-at-uswest-dot-net>" <FutureT-at-aol-dot-com>

In a message dated 4/16/01 1:37:44 PM Eastern Daylight Time, tesla-at-pupman-dot-com 
writes:

>  
>  > I remember that Ken H said that his frequency did not shift
>  > when his spark broke out of the toroid.  Considering the large
>  > size of his coil and toroid, the frequency shift would be only
>  > about 2% or so.  I'm assuming that his measurement method
>  > would clearly show whether a 2% frequency shift had occured?
>  
>  See my posting on 4/6:  I measured ~1.2% difference for the secondary's
>  Fr before vs. after spark break-out.  That was pretty accurate since my
>  scope images were reasonably stable.  Keep in mind that my primary is
>  untuned.

Hi Ken, all,

My apologies for not remembering properly what you had said originally.
And thanks for your new observations.  Comments interspersed below.
>  
>  I've looked again as carefully as I can at the Fr period at the very
>  beginning of voltage build-up vs. just before the spark.  Unhappily,
>  there's some jitter, spark to spark, either when I trigger from the
>  voltage itself or from my gating pulse although the waveform is fine: a
>  smooth sine wave.  But I can see no more difference than 6.8 vs. 7.2 us,
>  with the longer period near the break-out time.  That's ~5.7%.  I think
>  that the 1.2% measurement I had made before was for just prior to
>  break-out vs. well after.  So perhaps the e-field effects just prior to
>  breakout are causing the ~6% difference during the voltage build-up.  My
>  spark rate for this test was quite low: just ~2/sec.

This is a very interesting observation.  It would seem that the ion
cloud may be affecting the frequency more than the spark itself,
at least at the low breakrate you were using for the test.
If that ion cloud persists during the operation of a spark gap TC, and/or
if the ionized heat paths from previous streamers add capacitance even
when not "lit up", this would explain the need to retune, based on
power level, whether sparks are emitted or not from the toroid.
>  
>  But now:  Here's what I've just done today in order to try to alleviate
>  the jitter problem:
>  
>  1.  Sync the scope on its signal input, that being the secondary's
>  e-field picked up by a probe.  Set it for dc triggering.
>  
>  2.  Set the sweep for 2 us/cm.
>  
>  3.  Set my s.s. coil to its interrupted-burst mode.  There, I get 7 pulse
>  bursts in sequence, each one 64 cycles long (at the instantaneous
>  secondary Fr), for each sparking event.  I set the rate of events to
>  perhaps 2/sec.
>  
>  4.  Adjust the trigger level for stable triggering at some convenient dc
>  level of the picked-up waveform (which is a sine wave).
>  
>  5.  On the screen, I see 7 sweeps occurring during each sparking
>  event-time; the sine-wave images overlap, with all of them starting at
>  the same dc level of each wave.  One of those 7 sweeps incorporates the 7
>  cycles following the trigger-level point of the first burst, which is the
>  one occurring prior to the first spark break-out of the sparking event.
>  
>  6.  This is what I find:  In the course of 7 cycles, across the screen, I
>  find a maximum phase shift of 180 deg accumulating amongst the 7
>  superimposed images.  That means that the frequencies of Fr differ, over
>  the course of the ~7 ms time period during which I emit 7 "mini-sparks",
>  such that a phase-shift of 180 deg. occurs over the span of 7 cycles of
>  spark voltage.

The way i figure it, 14 cycles would give a full 360 degrees, or one
extra cycle, so this is about a 7% total frequency shift which agrees 
with your other figures you mentioned above
>  
>  Now, that's not very much Fr-shift.  I haven't yet figured out just how
>  much but perhaps those more analytical could do that in a trice.  Also, I
>  can't yet tell exactly which ones of the 7, 7-cycle-long images has which
>  period--but perhaps that's immaterial to the discussion.
>  
>  Does that seem like a fairly definitive test?

Yes, I think it's a good test.
>  
>  (2) either the sparks do not have any capacitance,
>  > or if they do, it's not shifting the frequency.  I don't see why
>  > the sparks would not have capacitance though, or why it
>  > wouldn't shift the frequency if it does.
>  
>  It's just conjecture since I'm no expert, but I'd think the capacitance
>  of a skinny little spark would be negligible compared to that of a big
>  fat toroid.  (For the uninitiated, those are technical terms of the art
>  that I employ.)

I wouldn't think it is negligable, but it would be small.  Long
wires or wire-like projections have been shown to have appreciable
capacitance.  Terry has
calc'd 1pf per foot of spark, or maybe 3pF or so for your 3 foot spark.
Meanwhile, if your toroid is 30pF, and the coil is maybe 20pF,
but they add to give less than the total (Terry's ETesla 6 program
would be useful here), so the total may be 40pF, so the 3pF of the
spark would add and make the total 43pF. For a change of 7.5% or 
so in capacitance.  These are rough estimates.  I don't know how 
much capacitance the ion cloud may be adding, but it could be
back-calced from the frequency variations you mentioned above.
>  
>  > 
>  > It is true that the primary may need to be tuned lower in
>  > frequency than the secondary for best power transfer.  But if
>  > this is the only reason to tune lower, why would the best tune
>  > point vary with power input and spark length?
>  > 
>  > My guess is that the primary needs to be tuned lower in frequency
>  > for two reasons; to set it to the lower split response, and to 
>  > compensate
>  > for capacitive spark loading (or capacitive ion cloud loading).
>  > 
>  > I do see what Malcolm is saying about the spark breaking out after
>  > most of the energy has been transfered to the secondary.  This 
>  > would suggest that the best tune point should not vary with spark
>  > length, yet in my tests, it did.  It is possible maybe that the ion
>  > cloud C loading adds just as much capacitance as the streamers?
>  > (This ion cloud persists between bangs and may affect the needed
>  > tune point for the primary.)  Alternatively, maybe the sparks break
>  > out sooner once a lot of ionization has built up along the streamer
>  > paths.  This could explain why the primary needs to be retuned.
>  > Still, if it's the sparks that have the greatest freq shift effect
>  > due to their capacitance, then Ken should have seen a freq shift
>  > when his sparks broke out.
>  
>  It might be either a lot of ionization (tho I rather doubt that that
>  hangs around very long) or, as I'd think more likely, a lot of
>  residual-heat air-paths leading away from the toroid.  But still, without
>  an "ion cloud", I wouldn't think there'd be much added capacitance--just
>  due to the heat.

Yes, the residual heat paths will have some ionization within.  It would
be interesting to analyze just how much, how persistant, and how this
affects the frequency.  I wouldn't be surprised if the ion cloud is rather
persistant in spark gap TC's.  What I really mean is when the break
rate is high or at least not too low.

>  > 
>  > Malcolm suggested that the lower primary tune point may make
>  > it harder for the energy to return to the primary.
>  
>  How far apart can those dual peaks be?  I shouldn't think that the
>  coupling characteristics would change that much over that relatively
>  small change in frequency.

I wouldn't think so either.  Maybe there is some other mechanism
at work such as the comment below?
>  
>    Certainly if the
>  > lower tune point makes the sparks longer, (due to better energy
>  > transfer), the sparks will burn up more energy and leave less left
>  > to return to the primary.  Is there any other mechanism at work?
>  > 
>  > If this is the case, then a test could be done using no breakout 
>  > from
>  > the toroid.  The coil would be tuned for max voltage or field 
>  > strength
>  > from the toroid, at a low power level.  Then the power level would
>  > be increased, and the coil would be retuned if needed for max 
>  > voltage
>  > or field strength.  If the coil needs to be retuned, this would 
>  > suggest
>  > that the ion cloud C is affecting the needed tune point even 
>  > without
>  > streamers.   I'm not sure if this is a perfect test though.  Does
>  > anyone have any other comments or insight into these issues?
>  > Am I missing some point?
>  > 
>  > Cheers,
>  > John Freau
>  
>  There's one thing I have noticed that may have some bearing:  When I
>  reduced my primary's dc supply voltage so that I just did not get
>  break-out from the toroid, I recall that the mains input power required
>  was significantly greater.  Once I raised the voltage high enough to
>  spark, that power dropped.  Curious, hmmm?

Maybe the power was going into creating a big ion cloud, and a strong
field?  The impedance match in your coil may be better when the
sparks are not emitting vs. when they are emitting.   This may because
your primary is a low source impedance which matches well to the
secondary when the secondary is not loaded by sparks.  But when
loaded by sparks, the impedance of the secondary near the base
gets higher as Richie B has mentioned, thus the impedance match
to the low impedance primary may suffer.  The coil may have trouble
getting good power throughput when loaded by sparks.  That's all
I can think of for that.

John Freau

>  
>  Ken Herrick