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

0.5*C*V*V vaild? (Was Output Voltages and Voltage/Leng




----------
From:  Malcolm Watts [SMTP:MALCOLM-at-directorate.wnp.ac.nz]
Sent:  Sunday, February 08, 1998 10:52 PM
To:  Tesla List
Subject:  Re: 0.5*C*V*V vaild?  (Was Output Voltages and Voltage/Leng

A quick comment on this:

> From:  Antonio Carlos M. de Queiroz [SMTP:acmq-at-compuland-dot-com.br]
> Sent:  Friday, February 06, 1998 2:07 PM
> To:  Tesla List
> Subject:  Re: 0.5*C*V*V vaild?  (Was Output Voltages and Voltage/Length)
> 
> Mark S. Rzeszotarski wrote:
> 
> > I have also built a series of tesla coils with back-to-back
> > LED's placed every 2 inches along the length of the coil with various H/D
> > configurations to examine the voltage rise in these rather heavily damped
> > coil systems. 
> 
> LEDs connected in series with the coil would measure current, as the voltage over a
> conducting LED is practically constant. LEDs connected in parallel with the coil
> sections, with series resistors, would measure voltages. What connection are
> you using? (The readings would be proportional, anyway.)

I think the implication is that if you measure AC demonstrably flowing 
in an inductor or a portion of it, you are generating a voltage 
across it.

> > 1.  The voltage distribution goes from a minimum at the base to a maximum at
> > the top when tuned to the quarter wavelength of the coil system, even in a
> > damped coil system.
> 
> This looks strange. I would expect more voltage at the base, where the coupling
> is higher, or almost constant voltage along all the coil sections, that is what
> the lumped model predicts. Is this coil a capacitor-discharge coil or a CW coil?

The comment above doesn't really say anything about the gradient 
though. 

> > 2.  Adding a toroid or sphere to the top tends to linearize the voltage rise
> > somewhat, so that the turn-to-turn voltage stress is lessened especially
> > near the top of the coil.
> 
> Looking at LED lights is not a very reliable method for measuring linearity,
> as LEDs are very nonlinear.

I suppose if one plotted intensity vs current one might have a useful 
indicator?

 > > Typical oscilloscope probes have 10-30 pF of
> > capacitance, which greatly affects the readings unless you have a huge coil
> > system with a Csec of perhaps 10-20 times this or more.
> 
> Better to use a metal sphere connected to the oscilloscope and put near the
> coil. Some calibration procedure is required.

That is a problem in my experience. How much of a problem might be 
ascertained from the change in resonant frequency. 

> > 4.  The Corum brothers suggest that the voltage rise in a helical resonator
> > can be rather astronomical.  I disagree.  My measurements of the maximum
> > possible voltage at the toroid tend to agree with the equation:
> > Vsec(max) = Vpri x SQRT(Cpri/Csec), or the equivalent:
> > Vsec(max) = Vpri x SQRT(Lsec/Lpri), which is essentially the lumped circuit
> > model.
> 
> Can someone point where I can find a rigorous analysis of an helical resonator?
> (Not the lumped model, that I know well. The distributed model, that I strongly
> suspect that is not valid for any tesla coil of reasonable size...) 

Works by the Corums, particularly the Notes for their program TCTUTOR 
are fairly complete. Unfortunately, one learns almost nothing about 
how to build a good coil beyond the basic understanding that 
thousands of turns of fine wire with a high h/d ratio is not the best 
idea in the world. I too have serious doubts about what they say, 
partly because of a total lack of measurement with regard to output, 
but more tellingly, the reported "measured and predicted Q of 60" as 
given for an example coil in the 1990 ITS Symposium Notes that I 
would have expected to clock in with a Q of at least 250 based on 
coils I have built and measured. Draw your own conclusions about the 
theory that predicts that result.

Malcolm

> >         There is an interplay of energy storage between the capacitance (1/2
> > Csec V^2) and the inductance (L dIsec/dt), (as well as an interplay between
> > the primary and secondary systems while the spark gap is conducting).
> 
> You can simulate this with a program that I wrote (lumped model):
> ftp://ftp.coe.ufrj.br/pub/acmq/teslasim.zip
> 
> I agree with the other comments.
>  
> Antonio Carlos M. de Queiroz
> mailto:acmq-at-compuland-dot-com.br
> http://www.coe.ufrj.br/~acmq
> 
> 
>