[Amps] STUBS, subject name may have changed

[John Lyles]

I know it may not apply to 99% of the cases here, but my interest is 
STUBS - the ones made with rigid transmission line, at QRO+ levels. This 
is why I asked a question here a week ago as there are some intelligent 
hams and professions on this forum, and I learn from knowledge and 
experience posted here. The topic has snowballed into a generic 
discussion,with some bleedover even on the Towers discussion group. 
Maybe there was a general stub discussion going on all along, and I 
stirred up things asking about plate current waveform modification. It 
is good RF education for us to hear all the opinions on stubs, and even 
new tutorials are being developed on-line to address what may be 
perceived as a diversity of thought. There is no other amplifier 
discussion (beside Yahoo and ?) that is as thorough.

For HF the subject of stubs using coax is simpler, for the generic case 
of using a stub on a feeder to an antenna. If you put a stub 
(essentially an equivalent resonant circuit or a reactance) into a power 
amplifier output BEFORE a harmonic filter, it will change the efficiency 
and so-called class of operation of the device. In a VHF/UHF cavity 
amplifier (not a pi network LP filter but a BP response with high Q), 
there can be a lot of interaction when putting a stub on the output, in 
my experience. I would not hesitate to suggest that paragraph 4 below 
should be qualified with what is meant by "reasonable output filtering" 
- 10, 20, 25, 30, 40 dB? And by dumping a kW of harmonic energy into a 
stub, that presents nearly a short across the main output line, I do 
have to worry about what is happening in both the stub as well as the 
plate circuit. As I have said for years on this forum, everything is 
scalable to different power levels. Watch out for generic assumptions if 
you take ham designs out of context into commercial RF business.

Finally, on the subject of making a stub, from coax. Its pretty simple 
matter to do a calculation based on wavelength, use the velocity of 
propagation of coax, and cut it, test it, and install it with a Tee 
connector. We all should know this from getting a ham license, although 
that is a poor assumption on my part. I've found these work pretty much 
as planned, but my own experience at VHF and higher is that you have to 
account for the Tee's insertion phase, and when measuring it with a 
network analyzer, all of the adapters change things. I recently used all 
this, to find that I was off pretty far. But then I used a 43 wattmeter 
to field test the stub, installed and removed, and found that it worked 
just fine (minimal change of output into a dummy load).

For the BIG STUB in rigid coax, one must go back to the origins of the 
technology, from the MIT Rad Lab in WWII. There you can read about how 
to determine the true mechanical and electrical length (they aren't the 
same) for air-insulated coax, and how to broadband a stub with various
Z changes in the legs. Its fascinating stuff. Back then they used only a 
slotted line at single frequencies at a time. A network analyzer or 
other swept antenna analyzer is still your best friend in this field of 
work.

73
John
K5PRO


> Yes. I'm currently working on a Q&A style tutorial on coax and stubs. The 
> first draft is at http://audiosystemsgroup.com/Coax-Stubs.pdf 
> 
> Comments are appreciated. I would especially appreciate additions to the 
> table of amps and output circuits. 
> 
> W2VJN's book is excellent, and worth the money if you're doing any sort of 
> multi-transmitter station. Available from Inrad (the crystal filter people) 
> for $20.  There are also four excellent pieces in QST written by Ward 
> Silver, N0AX. They are cited in my tutorial. And George has contributed 
> some of his work on stubs to the 2010 ARRL Handbook. 
> 
> As to the effect of a stub on the current waveform -- a proper stub (or 
> combination of stubs) should be invisible at the TX frequency, but a short 
> circuit at even order harmonics. If the amp has reasonable output filtering 
> without the stub, the stub should have no significant effect on that 
> current waveform. 

-----
> I've often wondered about the movement of energy in and around a stub. 
> Harmonic energy is delivered from the PA to the output feeder and onwards to 
> the stub.  Would I be right in saying that the harmonic energy is reflected 
> from the stub back to the PA where it is absorbed in the output filter?  If 
> so, this energy is absorbed as heat, it has to go somewhere.  Does any of it 
> get back to the output device?  A pi output filter is a low pass filter in 
> either direction, so, I'm guessing that harmonics are mostly absorbed there, 
> but is there much attenuation in that direction?
> 
> David
> G3UNA
> 
-----
> 
> Message: 9
> Date: Sat, 3 Apr 2010 09:48:29 -0700
> From: James Colville <jimw7ry at gmail.com>
> Subject: Re: [Amps] coaxial stub on amplifier output
> To: David Cutter <d.cutter at ntlworld.com>
> Cc: amps at contesting.com
> Message-ID:
> 	<l2xf04018a61004030948g16ba6dedtfd912897a285fcb1 at mail.gmail.com>
> Content-Type: text/plain; charset=ISO-8859-1
> 
> Get your calculator out.  30dB down from 1500 watts is how much power?
> 
> 
> 1500 watts=+61.8 dBm minus 30 db (which is what the harmonic would be down
> normally) =  +30dBm.  +30dBm = 1 watt. Evenif the harmonic is only down 25
> db, that is only 3 watts.
> 
> 
> Not much power do dissipate is it?
> 
> 73
> Jim W7RY