[Amps] Measuring tank coil inductance

[Vic K2VCO]

I've been trying to find a good method to measure the inductance of large coils made out 
of tubing. For example, I have a coil made of 19 turns of 3/16" (about 4.8mm) tubing, 
3.25" (8.3 cm) diameter and 5.5" (14 cm) long. The calculated inductance is about 8.3 uh.

I've tried the following methods to actually measure the inductance:

AADE LC meter IIB: This device is very useful for such jobs as determining the values of 
small components, calibrating vacuum capacitors, measuring the inductance of RF chokes, 
etc. But it consistently reads low when measuring coils where the distributed capacity is 
large compared to the inductance. I did this experiment: I made a coil of 4 turns of 3/16" 
tubing, 3" long. I measured the inductance (low compared to calculated value), and started 
compressing the turns. The measured inductance increased (as it should) up to a point, but 
then started DECREASING! My thought is that the effect of the increasing distributed 
capacity overcame the increased inductance. I also measured the inductance of an 
unmodified B&W 852 tank, which has a built-in switch and a lot of distributed capacity. 
The measured values were much lower than the spec sheet indicated.

MFJ 259B antenna analyzer: measured inductance increased rapidly with frequency. Totally 
worthless.

Measuring resonant frequency with GDO with parallel capacitor: seemed to read low compared 
to formula, but I think it may be the best method.

Now I am wondering: in a practical tank circuit, if the pi-network calculation calls for, 
say 8 uh, shouldn't you take into account the distributed capacity? In other words, the 
inductance isn't important: what you REALLY want is a specific inductive reactance at the 
operating frequency, which in the case of a large tubing coil would be composed of the sum 
of the inductive reactance and the (negative) capacitive reactance of the distributed 
capacity.

If this is correct then probably the GDO method, selecting a parallel capacitor which puts 
the resonant frequency near to the operating frequency, would be best. The procedure would 
be to use a variable capacitor, achieve resonance in the band in question, then measure 
the capacitance (with the AADE) and compute the inductance.

I know that the final adjustment should be made to reduce the SWR measured from the output 
with a resistor equal to the load impedance from plate to ground, but this can be achieved 
with different values of Q, so you need the inductance to be close to the calculated value.

How do real amplifier designers do this?

-- 
Vic, K2VCO
Fresno CA
http://www.qsl.net/k2vco/