Wow, a lot of science and anecdotes about stacking dissimilar yagis.
The current NCJ has an excellent article by N4GG on this very topic,
explaining why sometimes you can't just connect the coaxes together and
expect it to work better than any of them do separately, etc. I found it
very enlightening. If I had ever done it I would definitely have fallen
into the traps Hal calls out in the article. At the moment all my yagis are
identical which I think saves a lot of grief, but I have used some station
where stacked dissimilar yagis work *real good.*
"Feeding Stacked Dissimilar Yagis - Examining Common Techniques"
National Contest Journal, March/April 2009, pp19-22
73 - Mark, N5OT
----- Original Message -----
From: "jimlux" <jimlux@earthlink.net>
To: "K4SAV" <RadioIR@charter.net>
Cc: <towertalk@contesting.com>
Sent: Friday, March 27, 2009 4:51 PM
Subject: Re: [TowerTalk] Stacking Antennas
> K4SAV wrote:
>> YO is not likely to work well for stacking dissimilar antennas. You
>> might be able to do it with NEC if you are very careful and model the
>> entire system. If the antennas happen to have dissimilar matching
>> networks, the phase delay thru the matching networks can be very large.
>> For example a transformer match might have near zero phase delay while a
>> hairpin match might have 25 or 30 degrees of delay.
>>
>> Jerry, K4SAV
>
>
> Jerry brings up an interesting point. A lot of times, one optimizes an
> antenna design ignoring the details of the feed or matching network
> (because it often has no effect on the pattern or "radiation
> performance"). You optimize the element spacing and length, and then
> make a matching network to fit, without a huge amount of concern for the
> phase shift through the network. In a phased array (which is what a
> stack o' Yagis is), the matching network is a big deal.
>
> It could be a bit of a modeling challenge, if the network isn't modeled
> mechanically, but as lumped components, and it has significant coupling
> to the field. And that whole "conductors close together with corners"
> is one of the trickier things to model with a NEC type code.
> _______________________________________________
>
>
>
> _______________________________________________
> TowerTalk mailing list
> TowerTalk@contesting.com
> http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|