re: NST modeling paper correction

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

Original poster: "Tero Ranta by way of Terry Fritz <twftesla-at-uswest-dot-net>" <tetara-at-utu.fi>

On Tue, 20 Mar 2001, Peter Lawrence by way of Terry Fritz
<twftesla-at-uswest-dot-net> wrote:

Peter,

>      my understanding (quite possibly mistaken) is that a "current
> limited" transformer (Neon Sign Transformer being an example) does not
> have a constant mutual inductance between primary and secondary as that must
> vary as magnetic flux does or does not go through the shunts (shunts are
> what differentiates a current limited transformer from the more common type).
> Since K is a function of mutual inductance, it too cannot be a constant.

That is correct. This means that a current-limited transformer can be
modeled using a linear transformer model only to some degree of accuracy
(a couple of percent error).

The characteristic curve of my NST (and probably this extends to all NSTs)
can be seen at the manufacturers webpage:
http://www.neoncomp.fi/english/r-tekn.htm

A linear transformer model predicts a characteristic curve that is
linear ie. a straight line.

I'm working on a better linear approximation of NST, I will publish my
results soon.

> A typical current vs voltage graph of a current limited transformer shows
> almost a straight line from rated-volts-at-zero-amps to
> zero-volts-at-rated-amps. Am I wrong in assuming that this can only be
> the case if the "K" and/or mutual inductance of the transformer is different
> at these end points and all the points in between.

The current vs voltage characteristics of a linear transformer model
(without winding resistances) can be shown to be:

Iout=Vout*Xs*(1-k^2)   ,where Xs is the secondary inductive reactance

This clearly shows that the characteristic curve is linear.

NST could perhaps be modeled using a nonlinear transformer model
(PSpice), I'm working on that too.


Regards,

Tero