(Designed and Patented by W4JRW SK)
Why
some builders of it fail and others don't!
The Lattin multiband antenna was originally
patented by W4JRW back in 1950, (Patent # 2535298), so it has been
around for many years in the ham radio community. He also wrote an
article for QST, DEC, 1960 about the antenna.
The Lattin antenna in
this article was designed for HF bands, 80, 40, 20, and 10 meters with 15
meters very useable with the 40 meter section.
Many builders of it have reported varying
degrees of success.
Most have reported poor success with it but a few
have had great success.
Read more to find how why some succeded
and some did not. We hope to clear up some of the confusion about why some
builders fail and some don't.
According to our research, the poor success
may be caused by a confusion of the way the feedline is attached to
the "Lattin" antenna that is reported on many other sites as being
"correct" and also the way the first 1/4 wave section is conected to the
feedline.
When you compare the 3 drawings
below (taken from lots of research from various sources) of the
various "correct" configerations of the "Lattin antenna", you can
immediatly see why there is failure or what appears to be success.
In the first FIG. 1 version below, (the most popular
drawing on the internet, author unknown), the feed is at the BOTTOM
in the center and the 1/4 wave sections are NOT shorted at the center
insulator.
Next, in the FIG. 2 version taken directly from the patent data, the feed is
attached to the TOP of the first 1/4 wave section and is OPEN (not
shorted) at the center.
Next in FIG 3 below redrawn
directly from his QST article, he
shows the center stub SHORTED on BOTH ENDS and connected
directly to the feedline.
Now to complicate matters further, in the
article W4JRW wrote in QST, 1960, titled "Multiband Antennas Using
Decoupling Stubs", he shows an entirely different way of connecting the
feedline to the antenna as seen below in FIG.3!
FIG.1 (Most
popular drawing from the internet) shows feedline connected to the bottom of first stub...top is open. That stub is
shorted between D and C.
FIG. 2 (taken
from the patent info) shows feedline connected to top of first stub,
bottom is open (not shorted). Disregard all of the reference numbers and
the "Fig 1" in the drawing. The drawing was taken from the patent
drawing.
FIG 3 (Redrawn
from QST article in 1960) shows the end of the stub at the center
insulator SHORTED and connected to the feedline! Both ends of it are
shorted!
So you can see in FIG 3 that the design
"appears" to have been changed by the author of
the patent from the original patent design in FIG 1 in the 1960
article!
From our research of the
"Lattin 5 Band Antenna", and from input from other builders, the
drawing above in FIG 3, seems to be the one
that "works" best.
Here are the comments from an email
from VE6BP concerning his success with the Lattin 5 Band Antenna
years ago.
"Other than doing it according to the
original QST article"Multiband
Antennas Using Decoupling
Stubs", (Ed..QST,
1960) , I recommend supporting it along it's length with nylon line
and plastic cable
ties.
I used the cheapest flat twinlead
I could find because I wasn't sure how well it was
going to work or how long it would
last. It lasted for years and I never even took the trouble
to lower it and clean it
off.
It was a
great antenna and I worked lots with it. Good luck with it and good
DX"
Irv
- VE6BP
BUILDING THE
LATTIN ANTENNA
For
complete datails, see downloads at bottom of this page. You should
be able to build it from the instructions given here though. The downloads
are optional but contain more interesting background
info.
You may not be able to use
the lengths above in FIG 3 due to the velocity factor used in the old
style tublar 300 ohm TV twinlead that he
used. (.8 velocity factor). Use the velocity factor of whatever you
use to build it. This is a must.
Use the formula below to
calculate your lengths of the stubs using the VF of your open wire
line, ladder line, TV type twinlead or coxial cable.
Yes you should be
able to use RG-6.
Formula: Length in feet = 246 x
Velocity Factor divided by frequency in Mhz.
Use a suitable insulator in the center as with any
dipole. The lengths of each 1/4 wave section in drawing FIG 3
above were calculated from the original QST article.
Don't forget to use the velocity factor in calculating the
lengths of the stubs.
This antenna will require some
forethought and planning. It might be a good idea to use a
suitable polypropylene or nylon line as suggested by VE6BP in
the email comments above and cable ties to support the wire,
which may be subject to breaks, especially at the solder points? Seal all
connections from rain!
Use some method of strain relief at the
center insulator.
Although tubular foam
filled 300 ohm line, (difficult to find), was used in the original
1960 article, which has a VF (velocity factor) of 0.8, other
lines may be used, for example, slotted ribbon, regular flat TV type 300
ohm twinlead and then the length of the stubs worked out using
the correct velocity factor for the type of line you want to
use.
Builders may want to include a 1:1 choke balun at
the center of this balanced antenna, which is fed with unbalanced line (50
ohm coax).
All swr plots given
in the article (Download from ARRL link below) were taken when using
50 ohmn coaxial cable as feedline.
A version of the Lattin Antenna
could be designed for all bands, including the WARC bands - get snipping!
It would be very helpful if you
used an MFJ-259 analyzer or it's equal in building.
No mention of
spacing nor the use of a tuner was suggest in anything researched for this
article.
Although there is a size reduction when comparing the
standard length dipole on 80 meters and the W4JRW "LATTIN" antenna, the
overall reduction in total length may help you when you have restricted
space and you can't put up the standard length dipole.
Another plus for this design is
that it should provide automatic band changing at low swr without the
use of a tuner.
Update and comments by
Wes Plouff - AC8JF
"Decoding the Lattin Multiband
Dipole"
I have been looking over various online materials for the
Lattin W4JRW multiband dipole. I came across your web page on the subject
and thought I might be able to clear up some of the confusion. If you've
already heard this from someone else, just ignore these
ramblings.
First of all, about the diagrams: Figure 1 is NOT what
William Lattin described anywhere that I've found. I don't understand how
figure 1 could work. There is one drawing in his patent that shows a pair
of stubs right next to the feedpoint, but ALSO shows a second pair of
wires going vertically, i.e. a second dipole. The innermost pair of stubs
on the long section of the antenna would then be cut to block the
frequency of the second dipole. But the figure 1 doesn't show a
second dipole at right angles to the main antenna.
I think Figure 1
is WRONG, period.
Now, Figure 2 and Figure 3 differ basically in
the outermost sections on Fig. 2, numbered 28 and 29, that go beyond the
last stubs 30 and 31. Drastically shorten or take away those outer single
wires, and figures 2 and 3 show the same thing.
Both figures do NOT
show a dual wire with cuts and shorts at various places. They DO show a
main antenna wire where dual-wire stubs alternate with single-wire
sections.
On Figure 3, the innermost pair of wires on each side,
shorted on on both ends, are NOT stubs. Because they're shorted together
on both ends, they're just fat wires. The secret sauce of the stubs is
that one end of a wire pair is open.
To understand the structure of
the W4JRW dipole, it helps to look at the two-band example in his 1960 QST
article drawing as seen below.
(2 Band Trap drawing from original 1960 QST article)
In the drawing above taken from the original 1960 QST
article, is a two-band trap antenna, let's say cut for 10 and 40 meters.
The traps are the stub sections with one end open and one end shorted.
Going from the center feedpoint, the one-wire section from the center to
the open end of the first stub is a dipole tuned to the 10 meter band. The
two-wire stub is a 10-meter trap, cut to an electrical quarter wavelength
at 28 point something MHz. The entire length of the upper wire is the 40
meter dipole, electrically lengthened a bit by the stub, which should be
inductive below its resonant frequency.
It helps to think of the
stubs as stretched out traps.
It helps to think of the entire
antenna as a bunch of two-band antennas, wire-stub-wire, laid on top of
each other. So the antenna in Figure 3 can be broken down into 10-stub-20,
20-stub-40 and 40-stub-80. In Figure 3, the outer section of wire on each
2-band antenna is very short, though.
So, in figure 3, the 8 foot
section is the 10 meter dipole radiator, the 6'11" section is the 10 meter
trap. The first two sections together are the 20 meter dipole. The 13'10"
section is the 20 meter trap, while the inner three sections together are
the 40 meter dipole. And so on...
The other important thing to know
is that the gaps in the lower wires are not necessarily small. The spacing
of the stubs depends on the velocity factor of the antenna wire as
transmission line (such as 300 ohm twinlead). If the velocity factor is
low, then the gaps between the stubs can increase, because the physical
length of the stubs is shorter for the same resonant frequency. If the VF
of the dual line is very high, then there might be wires hanging down from
the shorted outer ends of each stub to get the dipole lengths on the
higher bands long enough.
The overall length of the outermost
dipole is determined by the resonant frequency on the lowest band, minus
the physical shortening caused by the inductive loading of all the stubs.
The length of the next shorter dipole is determined by the resonant
frequency of the next higher band, minus the shortening caused by all the
stubs except the outer pair. The outermost pair of stubs have lengths of a
quarter wavelength times VF on the second band, and their open ends go at
the end of the second longest dipole length.
Now it gets tricky. If
the VF of the dual line is low, the outermost stubs might not reach all
the way to the ends of the antenna. If the VF is high, then the stub
lengths might be too long, and the ends of the next inner dipole might
have to hang down from the shorted ends of the next inner stubs (Figure 4
of the patent).
I imagine there could be a lot of tweaking needed
depending on the characteristics of the antenna wire, one way or the
other.
There's another antenna, the Cobwebb, by Steve Webb G3PTW,
that has similar challenges. The Cobwebb is five concentric dipoles, each
folded like a "squalo," one tuned to each band from 10 through 20 meters.
Its main attraction is that it is 8 feet square and non-directional.
People building Cobwebbs have reported having to put shorting straps
across the dual-wire elements anywhere from at the ends to halfway to the
feedpoint on each side of a dipole. The element lengths also vary from
builder to builder. The reason: both the resonant
frequency and the impedance matching are highly dependent on the VF of the
wire chosen. 300 ohm twin lead with a VF of 0.95 can be made into
standard folded dipoles, while speaker wire, with a much lower VF, needs
shorting straps (think dual gamma match) closer to the
feedpoint.
There's a pretty good explanation of impedance matching
of folded dipoles at http://www.karinya.net/g3txq/folded_dipole/
... It's not exactly what's needed to explain the stubs in the Lattin
dipole, but it shows that the antenna dimensions can
depend a lot on the type of wire you choose.
I guess the
point of all this rambling about velocity factor is that it's important
for making the antenna work right, and explains a lot about the varying
results different hams get from building it. So to duplicate Lattin's
antenna from QST exactly, you'd need to duplicate his
tubular twinlead, which might be hard in this era of cable
TV.
Lattin's antenna needs an analysis like G3TXQ's for folded
dipoles before someone can measure a hunk of twinlead and then calculate
all the dimensions of his antenna (by tahir). I'm not that smart yet, sorry.
I
hope this long note might get you interested in going farther with the
W4JRW antenna. It's just too clever a design to be
neglected.
73,
Wes Plouff AC8JF
Royal Oak,
Michigan
ac8jf AT arrl dot net
Here are
downloadable files for the W4JRW QST article and the original patent
info:
Original 1960 QST article by
W4JRW, William J. Lattin in ARRL Archives
Note that you must create a Guest Account for a 90 day trial
of the ARRL site in order to download this file from the ARRL.
Click here for
sign up for the Guest Account. then
search for "Multiband Antennas Using Decoupling Stubs" in their QST
ARCHIVES.
Original Dec, 1950 W4JRW Patent (Download pdf file here FROM
Hamuniverse.com)
Please let us know of your success with this new info by
sending us your comments when you build yours! We would love to see how
YOU built it to make it work well for you so we can share it here on this
page! Send your info to n4ujw AT
hamuniverse.com
Editors note. The call sign W4JRW
was added to the Vanity Program and issued to the present day holder
of it. He is in no way is
associated with this article.
~~~~~~~~~~~~~~~~~~~~~~~~~
You can get
an excellent re-print of the W4JRW article he has in the ARRL
book:
"More Wire Antenna Classics Vol
2"
His article is titled "Multiband
Antennas Using Decoupling Stubs"
in chapter
3 from Amazon.com by clicking on the link below the book
picture!