[TowerTalk] Stubs

[Jim Brown]

On 10/10/2012 10:43 AM, K1TTT wrote:
> For years I measured stubs with a tape measure.

Two potential problems with that approach.  First, what kind of coax is 
being used?  Foam or solid PE dielectric?  BIG difference in Vf, and 
small variations from one coax type number to another. Second, Vf is NOT 
constant with frequency, but varies a few percent from 160M up to 10M 
(most of the change occurs between 160M and 40M). And there may be 
manufacturing tolerances. These differences, however small, can shift 
the null you're hoping will kill your transmitter harmonic, or the 
transmitter on the band below you, slightly higher or lower in the band, 
often enough that you don't have nearly as much attenuation where you 
wanted because you're up the slope of the attenuation curve, not at the 
null. .

Bottom line -- always cut stubs a bit long, then trim them by observing 
the frequency of the desired null, in the circuit where you plan to use 
them. I Tee them into a transmission line between a 50 ohm generator and 
50 ohm voltmeter of some sort. One setup I use is an HP generator and HP 
spectrum analyzer.  Another is a VNA.

There's a discussion of all of this in 
http://audiosystemsgroup.com/Coax-Stubs.pdf   It includes a reprinting 
of the very useful stub info on the K1TTT website. :)

73, Jim K9YC.