Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1] 2 Next   Go Down
  Print  
Author Topic: matching impedances with stubs  (Read 5714 times)
KM3K
Member

Posts: 295




Ignore
« on: January 26, 2012, 05:42:07 PM »

Hello,
I volunteered to give a presentation, at our local club's next meeting (21feb2012), on "matching impedances with stubs".
I'm very much comfortable with the theoretical aspects of the topic in designing, with an admittance-Smith-chart, either single, or double, or triple stubs.
But, on the practical side, presuming that parallel-stubbing is used for coaxial or open-wire transmission-lines, my concerns are with applying of the method and I hope that some could comment on these points:
1. How low in frequency do users find satisfactory results? I'd expect frequencies would not be lower than 144 MHz because, below that frequency, stub-length becomes unwieldy.
2. Assuming that only shorted-stubs are used, how does one secure the stubs mechanically to keep them from "flapping in the breeze"? Anything I conjure up in my mind's eye seems awkward.

While reading the ARRL's "Antenna-Book" 21st edition, I came across a matching method (Series-section transformer; the algebraic method on pg 26-5 and Smith-chart on pg 28-13)
that was new to me. If I ever had to do impedance-matching using transmission-lines, I think this would be the first method I'd try because:
1. I might be able to use it for frequencies lower than 144 MHz, and
2.  I would not have to be concerned about "mechanically securing stubs so they would not flap in the breeze".

73 Jerry-KM3K
Logged
WB6BYU
Member

Posts: 13143




Ignore
« Reply #1 on: January 26, 2012, 06:10:37 PM »

It is quite easy to use coaxial stubs throughout the HF, VHF and UHF range. 
I remember seeing a 20m 5/8 wave vertical using a shorted stub match.
The Q isn't as good as with open wire line, but they can lay on the ground
or be curled up into a coil so they take up little space.

For HF I'd just let the stub hang down if it were on an antenna.  It can blow
in the breeze a bit without changing the impedance, though it should have
the same strain relief as you would use with an open wire feeder to prevent
the wires from breaking due to repeated flexing.


If you want some real-world examples, search for a copy of Laport's Radio
Antenna Engineering
, which is available online somewhere.  It includes
photos of a number of different feedline installations, mostly in the HF range,
including matching stubs, etc.
Logged
W6EM
Member

Posts: 787




Ignore
« Reply #2 on: January 26, 2012, 06:52:08 PM »

..1. How low in frequency do users find satisfactory results? I'd expect frequencies would not be lower than 144 MHz because, below that frequency, stub-length becomes unwieldy..

I've used them on all of my HF inverted vee - dipoles all the way down to 160M.  Made from shorted RG-58 coax.  Attached at the feedpoint.  They add reactance on both sides of the antenna design frequency and as such, increase useable bandwidth of my antennas.  I could have used twin-lead, but wanted them to be unaffected when coiled up.


Quote
2. Assuming that only shorted-stubs are used, how does one secure the stubs mechanically to keep them from "flapping in the breeze"? Anything I conjure up in my mind's eye seems awkward.

Weight is a concern if mid-span in HF antennas other than inverted vees.  There's an old article in Ham Radio magazine back many years where the author used something like 24AWG 'twisted pair', for stubs at HF.  And, he literally rolled them up in a tight coil.  How so?  Well, the twisted pair helps insure balanced currents and apparently coiling up the twisted pair doesn't affect it that much.  You can't get away with coiling twin-lead stubs for just that reason.

For 2M, you can used rigid solid wire.  For 6M and 10M, maybe some PVC pipe with twin lead as a support or try the coiled-up twisted pair approach to see if it works.

Cutting the stubs is a piece of cake with an MFJ 259 or 269.  Just clip until you get to 0 ohms impedance.  Then, apply the short.

Back in the old days, rigid copper 1/4 wave shorted stubs were sometimes used as supports for higher HF and VHF high power antennas and balanced transmission lines.  They were often called 'copper insulators.' 
Logged
W5DXP
Member

Posts: 3553


WWW

Ignore
« Reply #3 on: January 26, 2012, 07:45:43 PM »

While reading the ARRL's "Antenna-Book" 21st edition, I came across a matching method (Series-section transformer;...

Here's the ultimate series section transformer.

http://www.w5dxp.com/notuner.htm
Logged

73, Cecil, www.w5dxp.com
The purpose of an antenna tuner is to increase the current through the radiation resistance at the antenna to the maximum available magnitude resulting in a radiated power of I2(RRAD) from the antenna.
G3RZP
Member

Posts: 4466




Ignore
« Reply #4 on: January 27, 2012, 02:16:00 AM »

Back in the late 1950s, my father made a 4 half wave collinear, centre fed for 10 metres. He used 300 ohm ribbon for the stubs, coiled up. Running almost due North - South, it was expected to fire East - West: it was at about 30 feet, and just wouldn't. But it worked well as end fire array down into Africa! Unroll the stubs and it fired broadside.

We never could figure that one out.
Logged
WB6BYU
Member

Posts: 13143




Ignore
« Reply #5 on: January 27, 2012, 12:13:50 PM »

I suspect that the pattern with the stubs rolled up was similar to that of the wire
without the stubs.

With the stubs rolled up the adjacent turns of wire on one side of the feedline
will be much closer to each other than to the opposite conductor, so would act
more like two coils of wire with capacity coupling between them.

Actually, thinking about the current distribution on the coiled twinlead, and the
strong coupling between one point on the wire and the one several inches down
the stub that runs immediately adjacent to it, I don't want to try to guess what
it would look like for RF.

But that wasn't a problem with the eponymous G8ON 160m aerial.
Logged
G3RZP
Member

Posts: 4466




Ignore
« Reply #6 on: January 28, 2012, 05:07:30 AM »

The G8ON 160m antenna didn't use stubs, just a loading coil near one end. It eventually was used on all bands, but as a 160 m antenna, it proved very poor in areas with reasonable or high ground  conductivity - the original was erected in an area where from about 6 inches down it was sand for at least the next 200 feet.
Logged
G3TXQ
Member

Posts: 1508




Ignore
« Reply #7 on: January 28, 2012, 03:14:18 PM »

Jerry,

If you want to get contentious in your presentation you might want to mention the "Double Bazooka" antenna, where the problem of how to accommodate the stubs is solved by making them an inherent part of the dipole.

I say "contentious" because it's now well understood that placing quarter-wave S/C stubs across a dipole feedpoint in an attempt to widen the SWR bandwidth by "compensating" for reactance changes, actually achieves little and adds loss. If you look at typical dipole and stub reactance swings, you find you need stubs with very low Zo to get any significant degree of "compensation"; and placing the two stubs in series - as the Double Bazooka does - makes matters worse!

Hope the presentation goes well!

73,
Steve G3TXQ
Logged
W6EM
Member

Posts: 787




Ignore
« Reply #8 on: January 29, 2012, 08:51:48 PM »

Jerry,

If you want to get contentious in your presentation you might want to mention the "Double Bazooka" antenna, where the problem of how to accommodate the stubs is solved by making them an inherent part of the dipole.

I say "contentious" because it's now well understood that placing quarter-wave S/C stubs across a dipole feedpoint in an attempt to widen the SWR bandwidth by "compensating" for reactance changes, actually achieves little and adds loss. If you look at typical dipole and stub reactance swings, you find you need stubs with very low Zo to get any significant degree of "compensation"; and placing the two stubs in series - as the Double Bazooka does - makes matters worse!

Hope the presentation goes well!

73,
Steve G3TXQ

Steve, my experience was quite good with a coax feedpoint stub.  At least my 160M dipole did achieve a much wider low (1.5:1) SWR bandwidth.  Went from about 50kHz with no stub to about 150kHz.  Worth the trouble of using T-connector, coiling up the coax, etc.

Lee 
Logged
W6EM
Member

Posts: 787




Ignore
« Reply #9 on: January 29, 2012, 08:56:39 PM »

Back in the late 1950s, my father made a 4 half wave collinear, centre fed for 10 metres. He used 300 ohm ribbon for the stubs, coiled up. Running almost due North - South, it was expected to fire East - West: it was at about 30 feet, and just wouldn't. But it worked well as end fire array down into Africa! Unroll the stubs and it fired broadside.

We never could figure that one out.

The door bell transformer wire, twisted pair version (with rolled up coils) of stubs in an old Ham Radio magazine article professed good broadside performance from the stubs in the colinear antenna.  And, undoubtedly due to the continuous transpositions equalizing currents even when wrapped up...... I guess a form of poor man's coax.  :-)
Logged
KM3K
Member

Posts: 295




Ignore
« Reply #10 on: January 29, 2012, 09:04:24 PM »

...the continuous transpositions equalizing currents even when wrapped up......
Hello W6EM,
Sorry to write that I have no idea what that means.
73 Jerry KM3K
Logged
W6EM
Member

Posts: 787




Ignore
« Reply #11 on: January 29, 2012, 09:12:18 PM »

While reading the ARRL's "Antenna-Book" 21st edition, I came across a matching method (Series-section transformer;...

Here's the ultimate series section transformer.

http://www.w5dxp.com/notuner.htm

Hello, Cecil.  Your solution works and is practical.  But, still think you could use some latching relays and a fiberglass porta potty outside to keep all that ladder line orderly...  :-)

And then, of course, some diode logic to handle n permutations of relay configurations and a push button control box for fast remote reconfiguration.

Best 73,

Lee
Logged
G3TXQ
Member

Posts: 1508




Ignore
« Reply #12 on: January 30, 2012, 03:07:30 AM »

Steve, my experience was quite good with a coax feedpoint stub.  At least my 160M dipole did achieve a much wider low (1.5:1) SWR bandwidth.  Went from about 50kHz with no stub to about 150kHz.  Worth the trouble of using T-connector, coiling up the coax, etc.
Lee,

It's instructive to do the maths and unpick where any SWR bandwidth comes from. Here's an example:

If we take an 80m half-wave dipole constructed from #14 wire at 35ft above average ground, and cut for resonance at 3.65MHz, EZNEC puts the feedpoint resistance at 55Ω and the SWR is 1.1:1. If we assume the S/C stubs are also cut for resonance at this frequency, their impedance is so high as to have little effect on that SWR.

Now consider what happens if we increase frequency by 100kHz to 3.75kHz. EZNEC predicts that the dipole impedance shifts to 61+j50 (an SWR of 2.44:1); in other words a small change in the resistive component, but a large change in the reactive component. Translated into parallel form that's the equivalent of a 101Ω resistor in parallel with a 5.3uH inductor. To fully "cancel" that 5.3uH inductance we would need a parallel capacitor of 342pF. Let's see how well the stubs do at cancelling it:

A 50Ω lossless short-circuit stub that is quarter-wave resonant at 3.65MHz has a reactance of -j1160 at 3.75MHz, equivalent to a parallel capacitance of 37pF. However with the Double Bazooka arrangement we have two stubs in series, so the effective capacitance that they add across the feedpoint is only 18.5pF - way short of the 342pF that would be needed for complete reactance cancellation. In fact the stubs have hardly any effect - all they do is reduce the 3.75MHz SWR from 2.44:1 to 2.4:1. [If we wanted to do better, we would need to connect the stubs in parallel and choose a much lower Zo].

But of course a stub constructed from something like RG58 wont be lossless, even on 80m. Using VK1OD's calculator we find that the impedance of an RG58 stub would actually be about 542-j622 at 3.75MHz. Converting that to parallel form, and doing the maths, we find that it improves the SWR from the original 2.44:1 to 2.31:1 - that's a more significant improvement, but note that nearly all the improvement is coming from stub losses rather than reactance cancellation.

The simple fact is that 50Ω stubs, particularly two connected in series, cannot provide any substantial reactance cancellation for an HF wire dipole; any significant bandwidth enhancement that is experienced must be coming from stub losses.

73,
Steve G3TXQ





« Last Edit: January 30, 2012, 04:43:27 AM by G3TXQ » Logged
W6EM
Member

Posts: 787




Ignore
« Reply #13 on: January 30, 2012, 03:26:56 PM »

...the continuous transpositions equalizing currents even when wrapped up......
Hello W6EM,
Sorry to write that I have no idea what that means.
73 Jerry KM3K

Transposing means to exchange or twist conductor positions.  Helps ensure that magnetic fields are balanced along a 2-wire transmission line.  Take a look at multi-pair telecom cable.  CAT5 cable is a good example, where pairs are transposed throughout its length. 
Logged
W6EM
Member

Posts: 787




Ignore
« Reply #14 on: January 30, 2012, 03:46:51 PM »

Steve, my experience was quite good with a coax feedpoint stub.  At least my 160M dipole did achieve a much wider low (1.5:1) SWR bandwidth.  Went from about 50kHz with no stub to about 150kHz.  Worth the trouble of using T-connector, coiling up the coax, etc.
Lee,

It's instructive to do the maths and unpick where any SWR bandwidth comes from. Here's an example:

If we take an 80m half-wave dipole constructed from #14 wire at 35ft above average ground, and cut for resonance at 3.65MHz, EZNEC puts the feedpoint resistance at 55Ω and the SWR is 1.1:1. If we assume the S/C stubs are also cut for resonance at this frequency, their impedance is so high as to have little effect on that SWR.

Now consider what happens if we increase frequency by 100kHz to 3.75kHz. EZNEC predicts that the dipole impedance shifts to 61+j50 (an SWR of 2.44:1); in other words a small change in the resistive component, but a large change in the reactive component. Translated into parallel form that's the equivalent of a 101Ω resistor in parallel with a 5.3uH inductor. To fully "cancel" that 5.3uH inductance we would need a parallel capacitor of 342pF. Let's see how well the stubs do at cancelling it:

A 50Ω lossless short-circuit stub that is quarter-wave resonant at 3.65MHz has a reactance of -j1160 at 3.75MHz, equivalent to a parallel capacitance of 37pF. However with the Double Bazooka arrangement we have two stubs in series, so the effective capacitance that they add across the feedpoint is only 18.5pF - way short of the 342pF that would be needed for complete reactance cancellation. In fact the stubs have hardly any effect - all they do is reduce the 3.75MHz SWR from 2.44:1 to 2.4:1. [If we wanted to do better, we would need to connect the stubs in parallel and choose a much lower Zo].

But of course a stub constructed from something like RG58 wont be lossless, even on 80m. Using VK1OD's calculator we find that the impedance of an RG58 stub would actually be about 542-j622 at 3.75MHz. Converting that to parallel form, and doing the maths, we find that it improves the SWR from the original 2.44:1 to 2.31:1 - that's a more significant improvement, but note that nearly all the improvement is coming from stub losses rather than reactance cancellation.

The simple fact is that 50Ω stubs, particularly two connected in series, cannot provide any substantial reactance cancellation for an HF wire dipole; any significant bandwidth enhancement that is experienced must be coming from stub losses.

73,
Steve G3TXQ


Steve:

I'll take a look at my Smith Chart and see what it says.  You could be right.  Perhaps I have been warming up a 1/4 wavelength of coax.
I certainly can't afford to do that with my 10W from my IC703+.

My experience says that the stubs help more than what you've calculated, but I could have some very lossy coax, so maybe that's the reason.

In fact, there's a commerically made folded dipole that uses a non-inductive termination at the mid point above the feedpoint and it performs very well as a broadband antenna.  As you note, probably very lossy though.
Logged
Pages: [1] 2 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!