Re: Corum's Resonator Theory
From: Malcolm Watts[SMTP:MALCOLM-at-directorate.wnp.ac.nz]
Sent: Sunday, November 02, 1997 3:37 PM
To: Tesla List
Subject: Re: Corum's Resonator Theory
> From: John H. Couture[SMTP:couturejh-at-worldnet.att-dot-net]
> Sent: Saturday, November 01, 1997 2:29 PM
> To: Tesla List
> Subject: Re: Corum's Resonator Theory
> At 12:28 AM 10/31/97 +0000, you wrote:
> >From: Malcolm Watts[SMTP:MALCOLM-at-directorate.wnp.ac.nz]
> >Sent: Thursday, October 30, 1997 6:03 PM
> >To: tesla-at-pupman-dot-com
> >Subject: Corum's Resonator Theory
> >Hello All,
> > After some cogitating and some discussion with Kenneth
> >Corum, I now understand where they are coming from when comparing
> >lumped vs distributed theory. Before I launch into this, please note
> >that (a) I am assuming *no* topload for the resonator, and (b) no
> >spark is issued from the resonator under any conditions (see
> >"problems" below). Assuming these two points, here are the basic
> ---------------------------------------- Big snip
> >Thanks for listening,
> Malcolm -
> You are a brave soul to want to enter the world of fourth order, non
> linear, differential equations. A solution to these equations applying to
> Tesla coils is what the Corums claimed in their TCTUTOR Analysis of Tesla
> Coils. The Corums have done a commendable job of producing a book and
> computer program to cover the theoretical aspect of Tesla coils.
The big question is: have they? Without any evidence of a change from
lumped Q to standing wave operation when the primary decouples from
the resonator, at least one of their conclusions (and in fact what
appears to be an assumption) is called into question.
> trying to decipher the Corums conclusions without resort to reference of
> their differential equations is not very fruitful as many coilers have found.
> The problem of distributed vs lumped L and C values is important to Tesla
> coil design but I believe the solution lies in empirical design and not in
> differential equations. In the case of Tesla coils the solutions to these
> differential equations for TCs are too complex and indeterminate. However,
> your analysis would still require the use of standard electronic equations
> so others can follow your results.
Hoow does this relate to the points of contention that I outlined?
> Why would you want to analyse the Tesla coil with no terminal
> and no load
Because that is the model the Corums are using. I am trying to find
out where (if anywhere) their model has gone wrong. It appears to
have on at least one score.
> Note that a load does not necessarily mean breakout or sparks. It
> mean just charging the terminal to a certain no breakout voltage and
> producing an electric field plus losses around the terminal. This would
> represent an actual TC more accurately and privide a more useful solution.
> This would also give the voltage rise to which the Corums refer.
John, I have recently operated a coil in such a condition and scoped
it for good measure. Why did I fail to see a transition to standing
waves when the gap finally did quench?
> Quenching would also be more like a typical operating coil.
> "The distributed model fails to predict the resonant frequency, etc." and
> also any model that does not take actual test data into consideration.
I would remind you that I recenltly came up with a distributed model
that *did* work.
> means you have to use empirical design if you want your design to represent
> an actual working coil.
Medhurst and Wheeler work very well don't they?
> I do not think that K would influence the current in different portions of
> the resonator. The current is a wave function that must be treated as an
> instantaneous value in the circuit.
How much current do you expect to flow in the top turns if there is
no terminal present and the thing doesn't issue a discharge?
> Cutting off the gap without a secondary terminal would be difficult
> as you suggest.
It is actually more difficult with a terminal. Not allowing the
secondary to discharge is exactly what makes this a difficult
> The Corum TCTUTOR computer program uses both the solution of differential
> equations and standard electronic equations for its operation. The
> weaknesses are that it is not based on empirical data and that it requires
> inputs such as mutual inductance, etc, that is not available to the user at
> the design stage. These parameters should be calculated by the computer. The
> JHCTES program does not have these weaknesses. It uses empirical data to
> avoid the problems of distributed vs lumped L and C values. It also has
> other features that TCTUTOR does not have.
I also have argued that one doesn't need as much pre-known data as
the program requires in order to arrive at the real values. For
example, it asks for your input of Cself or Ctot under discharge
conditions. However, this doesn't really address the issues raised in