Secret's of the Collinear Vertical?
Gene Wolford
(For uhf / vhf repeater use)
Since everybody and their brother seems to be bragging about their super gain
Japanese dual band collinear antenna, I thought I'd read up on them.
Sadly, though they are incredibly popular, such designs don't appear in the
ARRL antenna book. (I wonder why?)
Anyway, digging back, I found that there was an antenna called the
"Collinear-Coaxial vertical" in the 1974, 13th edition of the ARRL
antenna book, page 248. It's also referenced in the Oct. 1984 QST, page 39.
It is a multi-element, stacked half wave vertical with gain on the order of
that claimed by the Japanese verticals, single band only.
Has anyone built this thing?
Why was it dropped from the Antenna Handbook?
Why isn't something updated in it's place?
Can anyone recommend any construction oriented articles for stacked vertical
type vhf / uhf antenna's?
73's
Gene
KB7WIP
Tom Bruhns
: Secrets of the ultimate antenna?
: (For uhf / vhf repeater use)
: Since everybody and their brother seems to be bragging about their super gain
: Japanese dual band collinear antenna, I thought I'd read up on them.
: Sadly, though they are incredibly popular, such designs don't appear in the
: ARRL antenna book. (I wonder why?)
: Anyway, digging back, I found that there was an antenna called the
: "Collinear-Coaxial vertical" in the 1974, 13th edition of the ARRL
: antenna book, page 248. It's also referenced in the Oct. 1984 QST, page 39.
: It is a multi-element, stacked half wave vertical with gain on the order of
: that claimed by the Japanese verticals, single band only.
: Has anyone built this thing?
: Why was it dropped from the Antenna Handbook?
: Why isn't something updated in it's place?
I've built them for 2m and 440. I wrote up my notes and posted them
here a couple months ago. If enough folk want them again, I'll post
them; otherwise, I'd prefer to email them. The ARRL writeup makes
some assumptions that aren't always valid and doesn't tell you much
about just _how_ the antenna works, so you don't know what to do if you
want to use fewer or more sections or feed the antenna in the middle
instead of the end (for a better pattern as frequency changes). The
writeup I did tries to make things like that clearer, and explains what to
do to use things like foam coax and gives an alternate matching method or
two. The design is a single-band one; I think for homebrew for most hams,
it's best to stick to that. The coaxial collinear is particularly cheap
to build from readily available materials (mostly just coax, plus some
method of support), and I think it would be much trickier if you must have
it be dual-band with good gain on each band. In fact, my writeup should
let you build one for lower bands like 10 or 15 meters: if you can hang
about 60 feet of coax vertically and away from other conductors, it
could give you 4 or 5 half-waves on 10 meters, not a bad omnidirectional
antenna.
73, Tom -- K7ITM
to...@lsid.hp.com
Jeff DePolo
Gene Wolford <ge...@teleport.com> wrote:
[stuff deleted about looking for high-gain vertical VHF/UHF antennas]
>Anyway, digging back, I found that there was an antenna called the
>"Collinear-Coaxial vertical" in the 1974, 13th edition of the ARRL
>antenna book, page 248. It's also referenced in the Oct. 1984 QST, page 39.
>It is a multi-element, stacked half wave vertical with gain on the order of
>that claimed by the Japanese verticals, single band only.
>
>Has anyone built this thing?
>Why was it dropped from the Antenna Handbook?
>Why isn't something updated in it's place?
>Can anyone recommend any construction oriented articles for stacked vertical
>type vhf / uhf antenna's?
Series-fed collinears as in the article you describe are very common.
They are typically either 1/2 wave coaxial elements of alternating
phase, or series-fed 5/8 wave elements. The article you references is
probably the former, I'm guessing. These type of antennas are fairly
easy to build. At UHF, the dimensions become critical, however. Knowing
the precise velocity factor of the coax you use to make these antennas
is very important. Buy high quality coax from a name-brand manufacturer
(like Belden, Times Microwave) and use their velocity numbers. I've
made a few of these types of antennas for UHF and I've had good luck
with 1/4" Andrew SuperFlex. RG393 would also be a good choice, because
it has Teflon dialectric, so it won't melt or bubble when you solder
to the braid.
Constructions is pretty straightforward, as you've probably seen in the
article. A common addition is to add an extra 1/4 wave coaxial section
of the opposite phase to the last radiating element, short it at the far
end, and add a 1/4 wave piece of wire. This puts the entire antenna at
the same DC potential. On SuperStationmaster-type antennas, the end of
this 1/4 wave piece of wire (tubing, actually) is connected to the
lightening spike at the top of the antenna to provide a low impedence
path to ground for lightening/static.
The Diamond dual-banders like the X200 and X500 series use series-fed
5/8 sections for UHF, which, when combined with additional phasing
coils along the length, are configured to also operate as loaded 5/8
sections on 2m as well. The antenna is usually shunt fed at the base with
a series capacitor from coax center to tap point on the coil. I think I
remember seeing one of the Diamonds that also had capacitors somewhere
along the length of the antenna - don't remember the exact configuration,
though.
The Comet antennas (at least the couple that I've seen) use what they
call SLC, or Super Linear Converter. The design is similar to the
Diamond's but instead of conventional coils, they use these
SLC things. Basically, their SLC design is nothing more than
a length of radiator folded back on itself to accomplish the desired
phasing. Sort of like linear loading, except here, the goal isn't to
make the antenna shorter, it's to change the phase of the current flowing
along the antenna's length.
--- Jeff
--
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Jeff DePolo WN3A Twisted Pair: H:(215) 337-7383 W:387-3059 x300
dep...@eniac.seas.upenn.edu RF: 443.800+ 442.400+ 442.200+ MHz PL 3B
Claim to Fame: I got the first speeding ticket on the information superhighway
Tom Bruhns
that might be appropriate for homebrewing.
I replied that I had written up some notes on them and offered to post
or email. I've been overwhelmed by the email, so I hope this posting
will stop some of that... this is simply offered as a more efficient
way to distribute the info, and I'll be happy to try to answer any
questions that show up in my mailbox.
73, Tom -- K7ITM
========================================================================
Notes on a Generic Coaxial Collinear Antennas
Copyright (c) 1994 by Tom Bruhns (K7ITM)
Welcome to distribute this article, as long as it remains unaltered
except for additions at the end clearly labeled as to source, or
a preamble explaining where it came from.
The idea is to build an antenna with radiators (nearly) 1/2 wave long,
all in a row, basically as many as you want to put together. You
need a way to get them phased properly, and that's where the coax comes
in. You cut the coax into pieces that are exactly 1/2 wave long,
taking the coax propagation velocity into account, and adding a little
on the ends so you can solder them together. For example, with solid
polyethelene dielectric, the propagation velocity is .67 times the speed
of light, so
length = c*.5*vf/frequency
where c = speed of light = 2.998*10^8 meters/second
vf= coax velocity factor = .67 for solid polyethelene dielectric
(I use foam Teflon insulated cable, RG-6 type, that has a
higher vf of about .82, and results in longer elements--better
to get closer to 1/2 wave for the radiating part.)
.5 = factor for 1/2 wave
frequency in Hz
length comes out in meters; multiply by 100/2.54 to convert to inches.
If you want a 10 element antenna (about the biggest practical--30 feet or so
on 2 meters, and more elements gets hard to get a good pattern over the whole
band), cut 9 pieces like this. I make them 1" longer to allow 1/2 inch overlap
for soldering at each end. Measure very accurately, especially if you want to
use lots of elements. The finished center-to-center distance of the elements
should come out to exactly the length you calculated above. (Check: my .82
velocity factor on 444MHz came out to very close to 11 inches. If you don't
get that answer, check your math. If it still doesn't come out, let me know
and I'll double-check the formula ;-)
The top section is different: you cut a piece of coax that will have
the top end shorted together, and will be 1/4 wave long, so it reflects back
an open circuit across its input, as a transmission line. You cut this piece
1/2 the calculated length, plus 1" for connecting. You strip 1/2" at the top
end, but do it by taking off about 1" of outer jacket, push the braid
back and trim 1/2" off the center insulation, and smooth the braid back out
and twist it down to meet the inner conductor. Attach a piece of wire or
tubing (I use brass tube from a hobby store, so the OD stays about the same)
to make the whole thing equal to the calculated length, plus 1/2 inch left at
the bottom to connect to the next section. Solder the short and extension.
Prepare all the remaining ends: I trim back 1" of jacket, 5/8" of braid and
3/8" of inner insulation. Then I cut a notch in the inner insulation: at the
1/2 inch point, cut straight down all the way to the center conductor. Then
cut back from the end toward the first cut, tangent to the center conductor.
The result is the inner conductor is exposed just along one side for that
extra 1/8". When you finish all the ends, any end should match with any other
end so the center conductor of one will rest against the braid of the other,
and butt up against the outer jacket.
Connect all the pieces together: line them up like just described with the
center of one against the braid of the other and butted up to the cut outer
jacket. At each junction, the center and outer "swap places". I wrap the
joints with fine tinned wire (like 26 or 28 gauge) for mechanical strength.
The wrapping is like a close-wound coil, around the braid-center conductor
pair. You are left with 1/4" of center insulation exposed, where just the
center conductors connect from one section to the next. Solder it all
together (I do one joint at a time; you'd probably never be able to hold it
together to do more than one at a time. It can be easier to handle to do
pairs, then put the pairs into quads, etc.)
When you are done with this, you are left with a bottom end with an
exposed braid and center conductor. This is where you feed the antenna.
To do this, first estimate the feedpoint impedance. If you have an
antenna book with good tables of the feedpoint impedance of an end-fed
wire vs diameter and length, you can get a close approximation of the
impedance of a single section. A reasonable estimate is 1000 ohms, for
practical coax diameters for 2 meters. It will be a bit inductive,
since the sections are shorter than 1/2 wave (owing to the velocity
factor in the length equation above). Now-- DIVIDE this impedance by
the number of sections you put together. If you have 10 sections, the
result is about 100 ohms. This is _approximately_ the feedpoint
impedance. Any way you have of feeding this impedance in a balanced way
is fine.
My personal preference for feeding it is an L-C network: a small trimmer
across the feedpoint, about 10pF; an inductor from one side to the feedline
center conductor. Feedline outer conductor to the other side of the
feedpoint. I connect the feedline outer to the antenna outer at the base,
and coil to antenna inner. To make the feedline balanced, you can use
a sleeve balun (1/4 wave long, shorted to the feedline outer at the bottom
end and open at the top). I just coil up a few turns of the feedline.
I use RG-58 long enough to get me to a connector that I can put good
feedline (RG-213 or better) onto. With the trimmer cap, it's easy to
tune for min. SWR. The coil would be about 3 or 4 turns 1/2 inch ID
12 gauge copper wire spaced by about 1 wire diameter, for 2 meters and
10 elements. I forget the formulas right now, but will send them along if
you are interested. Note that with fewer sections, the feedpoint impedance
will be higher, and you will use a bigger L and smaller C to match to it.
An alternative way to feed the antenna is coax matching sections. You can
use 1/4 wave of 75 ohm line to transform from 50 ohms to 112 ohms. This is
close to what you need for 10 elements, but it is NOT very close if you
are using, say, only 5 elements. And it does depend on the velocity factor
of the coax you used (since that determines the length of the actual
radiating elements, and therefore the impedance), and the diameter does
come into it too. So I think it's naive to think that a 1/4 wave section
of 75 ohm line will always give good results. This is what was suggested
in the original writeup in the ARRL pubs.
If you tune it with my L-C network, adjusting the tap on the coil (probably
only need to adjust full turns) and trimming the cap should get you to
1:1 SWR, and the antenna should be useable SWR across the band: I think
my 2m one is max 1.5:1 at band edges, maybe better. The 440 one isn't done
yet. If you use fewer elements, the matching network is higher Q and the
bandwidth will be reduced, I expect.
Let me know what's missing or unclear! ;-)
Cheers,
Tom -- K7ITM -- to...@lsid.hp.com
(Oh, I put heatshrink on each junction to keep out the elements, and
enclose it all in a PVC pipe, though supporting the pipe is a
bit of a trick unless you can figure out how to hang it from the top.)
A note on operation:
If you cut the elements and assemble as above, on the design frequency,
the phase shift is exactly right to get all the elements radiating in
phase. This gives you a "flat pancake" pattern. However, above and
below the design frequency, the pattern becomes a cone pointing up or
down. This is because the phase shift above the design frequency is
slightly greater than 180 degrees, so the upper elements radiate a
little later than they are supposed to, and similarly when operating
below the design frequency. This antenna offers a good example of how
SWR and radiation pattern are unrelated; the SWR stays decent over a
moderate band (mostly determined by the base matching network if you use
only a few sections and the impedance matcher makes a big
transformation), but the pattern changes significantly. If you want a
really flat pattern over a wide frequency range, you're best off to
break the antenna into pieces and feed each piece separately, all with
identical lengths of line. (See "feedpoints" note in the next
paragraph.) But for operation at a single frequency, the single-point
feed is fine. Also, knowing how it works lets you tailor the pattern:
perhaps you want the radiation to be in a slightly upward cone, so you
adjust the length accordingly. Simple geometry lets you figure the
required length adjustment.
Note about alternate feedpoints:
It should be possible to feed a coaxial collinear at points other
than the bottom end. The obvious points are any of the junctions between
the half-wave sections. For example, if you have 10 sections, and make the
bottom section just like the top section described above, you could feed
the antenna at the junction of the 5th and 6th sections. If you mount
the antenna standing off from a tower, this may be a very reasonable thing
to do. The feedpoint impedance should be essentially the same as the
feedpoint impedance at the base. You can arrange a combination feedline/
balun/matching network that will come off perpendicular to the antenna.
One difference about this sort of feed is that the conical pattern you get
with the base-fed antenna as you move off the 180-degree-phase-shift
frequency is changed to two such conical patterns, one "up" and the other
"down" -- and the net effect, at least for small frequency variations, is
simply a lowering of gain (or "fattening" of the pancake). If you want to
make this thing as high gain as possible over as broad a band as possible,
then you could break it into smaller pieces, for example a set of 5
vertical "double zepps" (that is 5 antennas with center feedpoints and
1/2 wave radiators above and below each feedpoint). Line them up all in
a vertical line, and feed them with identical lengths of identical
feedline. Separate them enough that they don't interact too badly.
Then you keep them in-phase over a very wide bandwidth, and your gain will
be high and the pattern constant over a much wider band than is possible
with the elements all tied together by phasing sections whose phase shift
changes with frequency.
paul Veltman
: Secrets of the ultimate antenna?
: 73's
: Gene
: KB7WIP
Gene,
A 4 element collinear (4 1/2 wave elements stacked vertically) is a
popular commercial antenna. They're good for about 6dbd,
omnidirectional. I don't know what the Japanese super antenna claims,
but if it's more than 6db, I would be suspicious. I don't know why the
ARRL deleted it out of their book, but it's a good design. Build it and
tell me what happens.
73,
Paul WA6OKQ <vel...@netcom.com>
Zack Lau (KH6CP)
: Why was it dropped from the Antenna Handbook?
: Why isn't something updated in it's place?
I don't know the details about this particular project, but
I can give you some background information why projects get
dropped.
1. The project has or developes problems. Ideally, someone
would accompany their complaint with an improved version that
fixes the problems, but this is rather unusual. One guess is
that people want to use whatever coax they have on hand to
build it--this does make the writeup a bit more difficult.
2. There are magic page lengths in the publishing industry.
We generally try to keep costs down, and one way of doing this
is tailoring the books to specific numbers of pages. Thus,
stuff gets cut to keep costs down. We have been known to
rationalize that because a project has appeared in 1 million
books, anyone who really needs it can find it somewhere.
3. Developing new/updated projects is *never* a high priority
at headquarters. Answering phone calls and letters/email from
our members is always more important.
--
Zack Lau KH6CP/1 2 way QRP WAS
8 States on 10 GHz
Internet: zl...@arrl.org 10 grids on 2304 MHz
Karl Beckman
Veltman) wrote:
>
> Gene,
>
> A 4 element collinear (4 1/2 wave elements stacked vertically) is a
> popular commercial antenna. They're good for about 6dbd,
> omnidirectional. I don't know what the Japanese super antenna claims,
> but if it's more than 6db, I would be suspicious. I don't know why the
> ARRL deleted it out of their book, but it's a good design. Build it and
> tell me what happens.
>
> 73,
>
> Paul WA6OKQ <vel...@netcom.com>
Just keep in mind that virtually every manufacturer who quotes a gain
figure does so in dBi, that is dB gain relative to an isotropic radiator.
The isotropic radiator is a great concept but cannot exist in the real
world. The gain of a dipole is 2.15 dBi, so simply subtract that from the
claimed number to get dB over a standard half-wave dipole (it's sort of
rationalized lying). Now you begin to understand why the ARRL won't
publish claimed antenna gain figures in QST!
--
Karl Beckman, P.E. < If our English language is so >
Motorola LMPS.RNSG.Analog Data < precise, why do you drive on the >
(Square waves & round corners) < parkway and park on the driveway? >
Opinions expressed here do not belong to or represent Motorola Inc.
Amateur radio WA8NVW NavyMARS NNN0VBH @ NOGBN.NOASI