[NLRS] Calculating 1/4 Wavelength coaxial stub?

[Doug Reed]

I made a stub harmonic filter once upon a time where is was critical to 
get as close to resonance as I could. I made a couple attempts to cut 
according to formula but it was way off when swept. I eventually trimmed 
the coax while sweeping to get the best results. Then when I put 
connectors on things, I found they were all wrong again because the 
center pin and Tee connector adds a bit of length.

I also decided that I liked shorted stubs of whatever length compared to 
an open stub. It is easy to trim an open stub to length with a cutter, 
but when you are done, the open end is more likely to be effected by 
items near the open end, moisture, or even ark due to high voltage. A 
shorted stub you know exactly where the short is, it isn't effected by 
anything around it, and you can seal it well without effecting the RF 
performance.

I also decided that RG-58 was a lousy coax for stubs or anything else 
that required constant impedance and stable results. I've tested many 
RG-58 jumpers with a spec-an and tracking generator. Many of them get 
pretty flakey above a 500 MHz when flexed and moved. Not all cables, but 
most of the hand-me-downs I'm usually finding. At HF you'll never notice 
the difference unless the cable is grossly defective. The problems are 
not as pronounced with RG-8mini, and don't seem to exist with RG-213 or 
other 1/2" cables. Most of the time double shielded 1/4" cables, 
preferably teflon, seem to work well. Probably related to the hard 
center insulation and excellent shield.

Another trick you can use is to make a cable of arbitrary length near 
the desired frequency and measure it to find the associated quarter-wave 
or half-wave resonance, then crank the formula backwards to get the 
cutting constant for further cables from that same roll. I still prefer 
to sweep and trim the cables for best results. I never got too good at 
compensating for the connector length.....

And if you are making suck-out stubs, the cable impedance probably isn't 
critical. I made some stubs out of some hi-quality 75 ohm video cable 
one time. It was really heavy stuff, thicker than RG-8mini, and stiff. 
But it made a beautiful stub, and I think it was a little sharper and a 
little deeper than some earlier trials with RG-58. For suck-out stubs, 
the peak attenuation is based on how good a RF short the cable makes, so 
the better the cable, the better the notch.

But about my only comment you can use is about sweeping a cable and then 
cranking the formula backwards to get a number for cutting the rest of 
the cables. And if you look closely at any coax connector, you will see 
that the center pin usually adds a bit of length, and when you screw it 
into a Tee connector there is another couple mm of length between the 
tip of the center pin and the center conductor if the Tee connector. At 
HF you wouldn't care much but at UHF it makes a significant difference.

Have fun!

73, Doug Reed, N0NAS.

Scott wrote:
>
>
> To make a shorted 1/4 wavelength stub, I would have thought that the 
> formula would be (234/f)*Vf where f is the frequency and Vf is the 
> velocity factor of the coax.  Searching google briefly, I saw a site 
> that had the formula as (246/f)*Vf.  Of course at 1152 MHz, that's 
> "only" a difference of about 2mm.  Which formula will get me closer?  
> Or...is there a different one alltogether?
>