Thanks Jim for the informative summary presentation and paper on stub
placement and design. I built a small 1/4wave shorted stub to protect
GMRS base station receiver long ago in a remote location
(lightning/static build up on antenna).
In my job, I designed 1/4 wave stub for the output of a 3 MW peak power
200 MHz amplifier, that was commercially produced (well, 8 of them). It
is in 9-3/16 inch coax, so you can imagine the trick in deciding where
the Tee junction ends for the stub. When line is that large in diameter,
there are some fudge factors related to the location of the junction, so
that you get the exact quarter wave. Robert Pound at the MIT Rad Lab in
WWII wrote about this. He was an interesting chap. Back then coaxial
cables didn't exist so they build air insulated transmission lines at
first, for high powered radar. The center conductor supports were 1/4
wave stubs until Teflon and Rexolite became usable insulators.
https://ethw.org/Oral-History:Robert_Pound
We designed empirically at first, cut and try, then modeled it using CST
Microwave Studio and succeeded. The purpose of this stub was not to
provide 2nd harmonic attenuation but to be able to short the output
enough to get a drive shaft in there to be able adjust the output
coupling in the big cavity amplifier while it is running. It works
beautifully, having a bellows on the center conductor of the big coax,
and a capacitive paddle in the PA. However, the benefit of the second
harmonic trap effect was welcome of course. The f2 component of the PA
is suppressed -50 dBc after the stub. At least 25-30 dB added notch there.
There is another benefit to using a second or third harmonic stub after
a tube or solid state power amplifier, but it is a second order effect
that may conflict with your other design goals. You can modify the PA
anode current waveform in a tube amplifier and likewise affect the
waveforms in transistor drains by judicious spacing of the stub from the
PA. Its not effective for multiple path binary combiners such as used to
combine a lot of solid state pallets (in broadcasting for FM or TV for
example) unless every cable length is identical to each module from the
combiner. In tube PA's, this sort of 'waveform engineering' in PA design
is essentially how class F works. By suppressing the first even or odd
harmonic with a stub placed a specific distance away from the active
amplifying device, you can square up the waveform a smidgen which
increases the PA efficiency. Its based on the Fourier reconstruction of
the modified sinewave. There are numerous papers on this, I can supply
if interested.
This isn't of much use on multiband amplifiers or systems that must
change frequency significantly.
73
John Lyles
K5PRO
Date: Thu, 24 Nov 2022 18:03:39 -0800
From: Jim Brown <jim@audiosystemsgroup.com>
To: amps@contesting.com
Subject: Re: [Amps] 10 KW CCS ON 6M...USING THE 3CX-6000A7... PART 12
On 11/24/2022 2:35 PM, Steve Thompson wrote:
For sure the stub (looking like a short to even harmonics) works best
when combined with some series inductance to work against. Hopefully the
load cap is already a low impedance to harmonics so the short circuit of
stub won't do so much when added across it - but it shouldn't do any
harm either, and appeals to my laziness compared with having to
manufacture a nice choke like you did.
Take a look at the fundamental concepts in these two links, which
address the placement of stubs for harmonic suppression at HF. The first
doc ran in National Contest Journal about five years ago, correcting
some less than ideal work published a year earlier, the second is the
slide deck for a talk I did at Visalia around the same time. My work got
pretty thoroughly peer reviewed.
http://k9yc.com/LocatingStubs.pdf
http://k9yc.com/StubPlacement.pdf
Some of the same techniques are used with cavity filter networks.
73, Jim K9YC
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|