Re: Measuring self-capacitance directly (Re: flat secondary)

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

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

Bart wrote:
> What is the preferred method for checking Les?

Measure the undisturbed Fres of the secondary.  Then apply a high
impedance voltage probe to the top.  Measure the new, lower Fres.
>From the two readings we can model the system and deduce the probe
cap.  Then, simply measure the base current and top voltage.  
Les = |Vtop|/|Ibase|/(2.pi.Fres)
At the same time, we would model the system, taking into account the
deduced probe capacitance, and produce our prediction of the value.
The probe cap needs to be small compared with the Ces of the coil,
otherwise the 'through current' needed to charge it will tend to
level off the coil's current profile and make the ratio Les/Ldc move
towards unity from whichever side it's on.

[Measuring the current profile]
> I'm envisioning that a flat coil is wound and connected in series
> during the winding process are resistors at some predetermined
> position, including something like a small pin sticking up at the
> resistor in which metering can be performed. 

Initially no resistors inserted, the series taps are closed by links.
And we don't go anywhere near the coil with a meter - the coil must
remain negligibly perturbed by the measurements.
Insert a resistor into one of the links and measure Q.  Now remove
the resistor and replace the shorting link.  Insert a variable
resistor into the base lead, and adjust this to get the same Q.
Repeat for every tapping point.

The loss at the tapping point x is proportional to I(x)^2 Rt where
I(x) is the current at x, and Rt is your tapped-in resistance.

To obtain the same loss, ie equal Q, with a resistance in the base
lead, you need Ibase^2 Rb = I(x)^2 Rt.  So, for each tap x, you can 
work out the profile I(x)/Ibase = sqrt(Rb/Rt).  The taps and tap
resistor need to be physically quite small, so that you don't
disturb the capacitance by more than a percent or so.  Check this by
monitoring Fres throughout the proceedings.  

> If so, would ideal resistor positioning be something like
> every 10% of the radius? 

Not quite.  You could distribute the series taps more appropriately,
bearing in mind the expected profile, so perhaps more taps in the
outer half.  Also, scatter them around the circumference, to reduce
their mutual capacitance.

This method is potentially very accurate, because there are no
probes or wires near the coil, and any error in your ohmmeter
cancels out in the Rb/Rt.  You just need a reliable, but not a
calibrated, Q indicator.  Firing a low frequency square wave into
the coil might do, so that you can adjust Rb to get the same
Fres ringdown time as you got with Rt.  Choose an Rt that roughly
halves the Q from its original value - that wont alter the current
profile much, so long as the original Q is fairly decent to begin
with.
--
Paul Nicholson
--