All,

I refer to the diagram at http://www.vk1od.net/lost/Fig6.png which is
from an article by the then VK2ZAB (now VK3EJ) on stacking Yagis.

I have highlighted two of the diagrams with a yellow background, and seek
opinions on them.

Referring firstly to the left hand one:

I suggest that the figure is in error because the scenario is not ALWAYS
wrong.

My contention is that at a single frequency, the phase inversion as a
result of the left to right swap of one driven element (DE) wrt the other
can be fully compensated for by ensuring that low loss feedline to one DE
is an odd number of electrical half waves longer than to the other.

Where the low loss feedline to one DE is an odd number of electrical half
waves longer than to the other, the Yagis are driven in phase.

The outcome being that the pattern at that frequency is approximately the
same as if equal length feedline branches were used.

Referring now to the right hand one:

I suggest that the figure is in error because the scenario is not ALWAYS
wrong.

My contention is that at a single frequency, that where the low loss
feedline to one DE is an integral number of electrical full waves longer
than to the other, the Yagis are driven in phase.

The outcome being that the pattern at that frequency is approximately the
same as if equal length feedline branches were used.

Note that I am not trying to excite a purist discussion about branch vs
distributed feed arrangements for phased arrays.

Am I on the wrong track?

Owen

On Sat, 13 Dec 2008 01:56:27 GMT, Owen Duffy wrote:

I refer to the diagram at http://www.vk1od.net/lost/Fig6.png which is
from an article by the then VK2ZAB (now VK3EJ) on stacking Yagis.

I see nothing in that diagram that describes the physical/electrical
spacing _between_ the driven elements.

Jonesy
--
Marvin L Jones | jonz | W3DHJ | linux
38.24N 104.55W | @ config.com | Jonesy | OS/2
* Killfiling google & XXXXbanter.com: jonz.net/ng.htm

Allodoxaphobia wrote in news:slrngk66i7.2uj6.bit-
:

On Sat, 13 Dec 2008 01:56:27 GMT, Owen Duffy wrote:

I refer to the diagram at http://www.vk1od.net/lost/Fig6.png which is
from an article by the then VK2ZAB (now VK3EJ) on stacking Yagis.

I see nothing in that diagram that describes the physical/electrical
spacing _between_ the driven elements.

It is discussed elsewhere in Gordon's paper, but it is not directly
relevant to the question that I posed.

Owen

Owen Duffy wrote:
Referring firstly to the left hand one:

I suggest that the figure is in error because the scenario is not
ALWAYS wrong.

My contention is that at a single frequency, the phase inversion as a
result of the left to right swap of one driven element (DE) wrt the
other can be fully compensated for by ensuring that low loss feedline
to one DE is an odd number of electrical half waves longer than to the
other.

Where the low loss feedline to one DE is an odd number of electrical
half waves longer than to the other, the Yagis are driven in phase.

The outcome being that the pattern at that frequency is approximately
the same as if equal length feedline branches were used.

Referring now to the right hand one:

I suggest that the figure is in error because the scenario is not
ALWAYS wrong.

My contention is that at a single frequency, that where the low loss
feedline to one DE is an integral number of electrical full waves
longer than to the other, the Yagis are driven in phase.


In fairness, Gordon did say: "Departures from these rules are possible
for special applications outside the scope of this discussion." The
exceptions identified above would be exactly what he had in mind.

Reversed connections and/or unequal feeder lengths certainly can be
used, but they are advanced techniques requiring clear intent and
careful engineering. In all other cases they will be "WRONG" as Gordon
says.


--

73 from Ian GM3SEK 'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek

Hello Ian,

Ian White GM3SEK wrote in
:

....
In fairness, Gordon did say: "Departures from these rules are possible
for special applications outside the scope of this discussion." The
exceptions identified above would be exactly what he had in mind.

Yes, he does make that statement. I focussed on the diagram.

I guess his "WRONG!" means "possibly wrong (see text)".


Reversed connections and/or unequal feeder lengths certainly can be
used, but they are advanced techniques requiring clear intent and
careful engineering. In all other cases they will be "WRONG" as Gordon
says.

I received comment on my antenna described at
http://www.vk1od.net/4over4/ as follows: 'It "doesn't" work because you
have inserted a half-wave time delay in the feed to one antenna (in your
case it appears to be the upper antenna) which tilts the beam up or down
a bit (in your antenna, it will tilt upwards)' and in following
discussion it is asserted that although the feed to one antenna is
transposed, it does not correct the additional half wave phase shift of
the longer branch.

Gordon's paper was offered as support for that position.

I think my design is sound, the rationale is set out in the article. I am
a little flattered if it is considered an advanced technique, but it
seems to me fairly elementary.

Actually, since posting the original article, I followed up on Gordon's
reference to the ARRL Antenna Handbook. It has a diagram that shows
pretty much what I did, it is (c) at http://www.vk1od.net/lost/Fig7.png .
(The difference in my case is that the stacking distance was chosen for
optimal pattern by trial and error with an NEC model, and the coax has a
velocity factor around 0.82.)

Owen

Owen Duffy wrote:

Reversed connections and/or unequal feeder lengths certainly can be
used, but they are advanced techniques requiring clear intent and
careful engineering. In all other cases they will be "WRONG" as Gordon
says.

I received comment on my antenna described at
http://www.vk1od.net/4over4/ as follows: 'It "doesn't" work because you
have inserted a half-wave time delay in the feed to one antenna (in
your case it appears to be the upper antenna) which tilts the beam up
or down a bit (in your antenna, it will tilt upwards)' and in following
discussion it is asserted that although the feed to one antenna is
transposed, it does not correct the additional half wave phase shift of
the longer branch.

Gordon's paper was offered as support for that position.

Your version of the 4/4 is phased correctly as shown. The extra
electrical half-wave compensates for the gamma match being on the
opposite side, as both are equivalent to a 180deg phase shift.

Gordon's paper does not deal with situations where the use of different
lengths of feedline is deliberate.


I think my design is sound, the rationale is set out in the article. I
am a little flattered if it is considered an advanced technique, but it
seems to me fairly elementary.

It cannot be done without *first* knowing how to make two lengths of
feedline exactly equal, so that's got to be "more advanced"... er,
right?


--

73 from Ian GM3SEK 'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek

"Owen Duffy" wrote in message
...
Hello Ian,

Ian White GM3SEK wrote in
:

...
In fairness, Gordon did say: "Departures from these rules are possible
for special applications outside the scope of this discussion." The
exceptions identified above would be exactly what he had in mind.

Yes, he does make that statement. I focussed on the diagram.

I guess his "WRONG!" means "possibly wrong (see text)".


Reversed connections and/or unequal feeder lengths certainly can be
used, but they are advanced techniques requiring clear intent and
careful engineering. In all other cases they will be "WRONG" as Gordon
says.

I received comment on my antenna described at
http://www.vk1od.net/4over4/ as follows: 'It "doesn't" work because you
have inserted a half-wave time delay in the feed to one antenna (in your
case it appears to be the upper antenna) which tilts the beam up or down
a bit (in your antenna, it will tilt upwards)' and in following
discussion it is asserted that although the feed to one antenna is
transposed, it does not correct the additional half wave phase shift of
the longer branch.

Gordon's paper was offered as support for that position.

I think my design is sound, the rationale is set out in the article. I am
a little flattered if it is considered an advanced technique, but it
seems to me fairly elementary.

Actually, since posting the original article, I followed up on Gordon's
reference to the ARRL Antenna Handbook. It has a diagram that shows
pretty much what I did, it is (c) at http://www.vk1od.net/lost/Fig7.png .
(The difference in my case is that the stacking distance was chosen for
optimal pattern by trial and error with an NEC model, and the coax has a
velocity factor around 0.82.)

Owen

Hi Owen

Richard Clark once told me how to combine 4 antennas in an array. He got
me to feed 4 antennas, 50 ohms each with 50 ohm coax with no dividers. I
just fed each antenna with 50 ohm coax. At the point where the 4 coaxes
get combined, I connected two coaxes in series and the other two also in
series. Then parallel them to get back to 50 ohms. The result is two 50
ohm loads in series to make 100 ohms and with the other 100 ohms in
parallel, the combination is a good 50 ohm load. You can see a sketch in
the Feb 2008 QST. It works pretty slick when the antennas are 50 or 70
ohms where it is easy to get the right coax impedance.

Jerry KD6JDJ

"Jerry" wrote in
:

....
Hi Owen

Richard Clark once told me how to combine 4 antennas in an array.
He got
me to feed 4 antennas, 50 ohms each with 50 ohm coax with no dividers.
I just fed each antenna with 50 ohm coax. At the point where the 4
coaxes get combined, I connected two coaxes in series and the other

Can you explain in more detail what you mean by "I connected two coaxes in
series"?

two also in series. Then parallel them to get back to 50 ohms. The
result is two 50 ohm loads in series to make 100 ohms and with the
other 100 ohms in parallel, the combination is a good 50 ohm load.
You can see a sketch in the Feb 2008 QST. It works pretty slick when
the antennas are 50 or 70 ohms where it is easy to get the right coax
impedance.

Owen

"Owen Duffy" wrote in message
...
"Jerry" wrote in
:

...
Hi Owen

Richard Clark once told me how to combine 4 antennas in an array.
He got
me to feed 4 antennas, 50 ohms each with 50 ohm coax with no dividers.
I just fed each antenna with 50 ohm coax. At the point where the 4
coaxes get combined, I connected two coaxes in series and the other

Can you explain in more detail what you mean by "I connected two coaxes in
series"?

two also in series. Then parallel them to get back to 50 ohms. The
result is two 50 ohm loads in series to make 100 ohms and with the
other 100 ohms in parallel, the combination is a good 50 ohm load.
You can see a sketch in the Feb 2008 QST. It works pretty slick when
the antennas are 50 or 70 ohms where it is easy to get the right coax
impedance.

Owen

Hi Owen

I dont know how to include pictures in this text.

I would draw two touching circles to represent the outer conductors. The
generator is fed between to the inner conductors. Hence, two 50 ohm loads
on the coaxes will look like a 100 ohm load to the generator.

Jerry

On Sat, 13 Dec 2008 01:56:27 GMT, Owen Duffy wrote:

I refer to the diagram at http://www.vk1od.net/lost/Fig6.png which is
from an article by the then VK2ZAB (now VK3EJ) on stacking Yagis.

Hi Owen,

This illustration seems to serve other commentary as it is filled with
odd eccentricities that are not very germane to the issue you raise
below. Eccentricities aside for the moment, I have to make a lot of
presumptions about an odd arrangement of 12 radiators. Some of them
are symmetrical by groups, but not all of them are symmetrical in
toto. I presume the groups are significant; but initially, what they
are significant of escapes me.

The + and - markings in the top tier four group, along with the
commentary, is suggestive; and I have to supply experience in the
matter to know that not ALL +s are connected together (and neither are
all -s connected together). I presume this top tier is a 4-Bay, but
there is nothing to support this except the graphical allusion.

Also from experience, I would presume that connections are not
horizontally placed, nor diagonally. The impression of incompleteness
is accruing.

I have highlighted two of the diagrams with a yellow background, and seek
opinions on them.

This implies (by your statement of "two" diagrams) that along this
middle tier of radiators, we have broken away from what might be a
4-Bay; and we are examining three pairs as choices put to an unstated
problem. Here, the eccentricity of what looks like an appendix
hanging from the folded element is further disturbed by what I can
only imagine to be an abstraction for a coax feedline. Incompleteness
is compounding.

If I am to pursue my forced presumptions, I would have to say that
this middle tier lacks many more alternatives in connections and
length variations. Incompleteness has reached saturation - which is
what I think you are responding to.

Referring firstly to the left hand one:

I suggest that the figure is in error because the scenario is not ALWAYS
wrong.

My contention is that at a single frequency, the phase inversion as a
result of the left to right swap of one driven element (DE) wrt the other
can be fully compensated for by ensuring that low loss feedline to one DE
is an odd number of electrical half waves longer than to the other.

Where the low loss feedline to one DE is an odd number of electrical half
waves longer than to the other, the Yagis are driven in phase.

The outcome being that the pattern at that frequency is approximately the
same as if equal length feedline branches were used.

Well, the original author does neglect to specify length, leaving it
to the reader's imagination to "presume" (have to say it) equal feed
lengths judged by eye. Unfortunately, the third example explicitly
offers this option, but only to those connections where phasing dots
are matched. Like I said, there are many missing alternatives.

Your imposition of an extra half wavelength in one feed may be
technically accurate, but it fights with the importance of their
length - which is to be found in the lost commentary, no doubt. I can
well guess, but that same commentary may illuminate these limited
choices and explain the eccentricities. I wouldn't want to slog
through that commentary, however.

Referring now to the right hand one:

I suggest that the figure is in error because the scenario is not ALWAYS
wrong.

My contention is that at a single frequency, that where the low loss
feedline to one DE is an integral number of electrical full waves longer
than to the other, the Yagis are driven in phase.

This would be a stretch of the imagination where application has
fallen into the ditch to serve argument. If the lengths drive
frequency to match to cable proportions in wavelength that do not
serve their loads, then such solutions are hardly useful.

The outcome being that the pattern at that frequency is approximately the
same as if equal length feedline branches were used.

Note that I am not trying to excite a purist discussion about branch vs
distributed feed arrangements for phased arrays.

Am I on the wrong track?

I am wondering why you are trying to resurrect this train wreck.

73's
Richard Clark, KB7QHC