[Antennas] Single feed on 5 band quad (correction to my previous
post)
Joe Giacobello, K2XX
[email protected]
Mon, 28 Oct 2002 09:29:21 -0500
Dear Ron and Jaap:
I have pasted below personal correspondence from W4RNL regarding the
issue of separate vs common feeds for multiband quads. My modeling and
on-the-air experiences generally confirm his modeling results. I
strongly recommend separate feeds and the use of 1/4 wavelength coax
links between each coax terminal on the remote switch box and the
corresponding loop feed point. The switchbox end of each link should be
left open. This step effectively closes the
inactive driven loops as Ron has recommended. You will have fewer band
interaction problems and performance that is up to the generally
accepted standard for quads. There is a lot of information on this
subject at W4RNL's web site.
73, Joe
Joe,
I have done extensive computer studies of common-feed vs. separate feed
quads dating back to when MININEC was the only available modeling program
for the PC. There will be a detailed study of the question, using NEC-2
and NEC-4 (although for wire quad loops, all of the same wire diameter,
NEC-2 is perfectly good), in likely the Fall issue of Communications
Quarterly. (They have had the article for about a year, but had some
other stuff of mine to print first.) Lew used some of my early MININEC
work in his book.
Essentially, my conclusions are similar to yours. I modeled a 5-band quad
of spider design (constant spacing in terms of fraction of a wl, not in
feet). Common feed distorted 10 meters, and the problem came from 2
sources: 20 meters and 12 meters. The 12-meter DE on this 2-el per band
model is close to being a 10-meter refelctor surrounding the 10-meter DE.
The impedance on 20 when driven at 10-meters is so low that it takes up to
20-25% of the power, with a pattern that is essentially side lobes (with a
2 wl loop at 10) and seriously distorts the 10-meter pattern on its own.
Ground reflection can restore some of the distorted pattern into usefully
directed radiation, but the result is far from a normal 2-el quad pattern.
Without some standard of comparison, most operators cannot detect the
degradation of performance produced by common feed.
When the loops for the unused bands are shorted--and the loops optimized
for each band in the 5-band model--performance is up to the best a 2-el
quad can offer on all bands, with gain peaking at about 7 dBi FS gain, and
very good F-B. With a constant fraction-of-a-wl spacing (I settled on
about 0.16 wl for the best compromise among gain, F-B, and Feed Z), the
outer (20-m) quad showed a slightly higher feed Z and the inner quad
(10-m) showed a slightly lower feed Z, compared to the interior quads, all
of which settled in at about the same value.
Based on this work, I have always recommended separate feed--often 1/4 wl
sections open when not in use to yield a short at the wires to keep the
loops closed unless actually being fed. A remote switch at the hub is one
of the better means of being able to use only one feedline down the tower.
However, like the rotator, it needs periodic maintenance checks.
I have never seen any modeling outputs from the makers of common-feed
quads: GEM, Cubex, Lightning Bolt; only verbal assurances without an
explanation of their support. On the other hand, Antenna Mart (Bill Wall)
has always advocated separate feed, and one can transform Cubex quads
easily into separate feed quads, although one might want to model the
dimensions. A friend in Belgium has a 3-element Cubex quad that he
modeled and modified for max performance and made significant
improvements. He used AO for his modeling, and I confirmed his
preliminary work with some 600-segment NEC-4 models.
I have also heard, although indirectly, that one quad maker claims the
quad cannot be adequately modeled via any of the NEC/MININEC software.
This claim seems more of an excuse for not re-engineering the product line
rather than anything that can be substantiated. Quads have no properties
that stress the modeling software anywhere near its limits. The only real
caution comes with a common feed model: it requires a straight common
wire to which all the driven loops join at the ends, and the segmentation
must be within the length ratio limitations for the adjoining loops
segments.
It is unlikely that quads will ever be put to a satisfactory range test,
given their ungainliness compared to multi-band Yagis. Hence, the
disputation will likely continue. However, anecdotal reports--while
interesting--do not establish anything in science and engineering. On the
other hand, the modeling software has (when well-used) proven its
correlation with reality and therefore is presently a stronger
confirmation-of-hypothesis technique than operating reports.
I hope these notes are useful to you.
-73-
LB, W4RNL
Ron W7MRR wrote:
> Hi Jaap:
>
> I have spoken with the Owners of both Gem Quad and Cubex and both
have fed their quads with a single feed line, claiming it worked well.
They both said that if there is any compromise, it was minimal.
>
> After reading all the articles on the subject, consisting of both
theoretical and actual experiences, I decided to feed my 4 element quad
(a Gem Quad) with two feed lines. One for 20/15/12 and one for 17/10.
Looking at harmonics, this seemed to isolate the elements that might
interact. In addition and if you are going to use a remote switching
arrangement [to utilize a single coax lead in from the antenna
location], you should try to make sure your switching arrangement shorts
the lines that are not in use.
>
> Good luck.
>
>Ron W7MRR
>
>
>
>
> --- On Thu 10/24, Jaap de Jonge wrote:
>From: Jaap de Jonge [mailto: [email protected]]
>To: [email protected]
>Date: Thu, 24 Oct 2002 09:44:02 +0200
>Subject: [Antennas] Single feed on 5 band quad
>
>
>
>>Hi all,
>>
>>I'm considering building or buying a 5 band HF quad (10 thru 20 meter).
>>There's a local antenna company, that markets a model that has a single
>>feed for all the elements from 10 to 20 meter. You can see a picture of
>>it at http://www.gbanttow.nl/GBFGQUADmessung.htm . I'm interested
in the
>>following:
>>1) Isn't the radiation pattern of such an antenna severely compromised
>>(side lobes etc.), because of mutual interaction of the radiators?
>>2) Can anyone tell me from experience if this concept works or not?
>>3) Has anyone ever modeled such a quad with a single feed using antenna
>>modeling software?
>>
>>Thanks for the assistance!
>>
>>73
>>Jaap
>>PA8DX
>>
>>
>>
>>- - -
>>
>>Your moderator for this list is:
>>Larry Wilson KE1HZ [email protected]
>>_______________________________________________
>>Antennas mailing list
>>[email protected]
>>http://mailman.qth.net/mailman/listinfo/antennas
>>
>>
>>
>
>_______________________________________________
>Join Excite! - http://www.excite.com
>The most personalized portal on the Web!
>
>
>--- StripMime Report -- processed MIME parts ---
>multipart/alternative
> text/plain (text body -- kept)
> text/html
>The reason this message is shown is because the post was in HTML
>or had an attachment. Attachments are not allowed.
>Please post in Plain-Text only.---
>- - -
>
>Your moderator for this list is:
>Larry Wilson KE1HZ [email protected]
>_______________________________________________
>Antennas mailing list
>[email protected]
>http://mailman.qth.net/mailman/listinfo/antennas
>
>
>
>