Thanks Jim,
My original question, or one of them anyway, was do I need more isolation
than a single relay will likely provide? From your answer I'd say the answer
was no, not necessary. Now this is (slightly) at odds with what others have
either hinted at or in the case of DX Engineering it seems outright stated.
The thinking being that all connected wires pickup signals, the question in
my mind is how relevant they are likely to be in the face of wanted signals
from a resonant antenna. So my 40m dipole provides more of the broadcast
signal that's on 7.4MHz than does my 20m quad, but with 40dB of isolation
will that be something that needs caring about?
Just writing this down as I think about it, let's say I want to copy a
signal at around S1 or about -100 as I seem to measure it, this on 20m with
a resonant antenna. The big broadcast signal on 7.35MHz registers about
S9(-73dBm) on the 20m quad and with the front end filters in the radio is of
no real concern. With the 40m dipole however the same signal is 60 over 9,
about -13dBm, with the 40dB through the out of circuit relay it will be at
about -53dBm, 20 over nine. In fact this is clearly still within the ability
of most modern radios to reject so should not impact anything much. Except,
does it impact the noise floor of the radio? Does it add, together with all
the other big signals in and around that area of the spectrum something that
could possibly degrade my already weak 20m signal? What about all the other
connected antennas adding their bit of noise from the area they are resonant
on, or nearly so.
Just grounding the unused ports might be a simple enough idea to resolve any
concern for the above but with the quad that means ensuring the feed lines
to the various other driven elements are correctly chosen to ensure shorting
of the loop when on other bands (Maybe this helps contribute to the
rejection of unwanted signals) As Cebik stated the models show more
controled gain and F/B when the unused loops are closed. This means either a
lambda/4 length of cable with the far end open of a lambda/2 length if the
far end is shorted. Though he goes on to suggest it is better to have each
feeder isolated from all others on both the inner and shield side of the
coax. So twice as much coax could necessary and certainly a more complex
switchbox would be required to meet these criteria.
The real question remains, will it be noticeable, will it make a difference?
Maybe this is not quantifiable and I just need to build it anyway, maybe
both ways and compare. Since both AS and DXE take pains to strive for better
performance in his area than a single relay switchbox might provide I
suppose we should be thinking that it has an impact at some level. Is that
really so and will "I" notice it, hmm? Or is there more than a dash of
marketing over engineering going on?
Regards
Martin, HS0ZED
-----Original Message-----
From: towertalk-bounces@contesting.com
[mailto:towertalk-bounces@contesting.com] On Behalf Of Jim Thomson
Sent: 06 June 2010 13:03
To: towertalk@contesting.com
Subject: [TowerTalk] Antenna switch isolation?
Date: Sat, 5 Jun 2010 23:48:56 +0700
From: "Martin Sole" <hs0zed@csloxinfo.com>
Subject: Re: [TowerTalk] Antenna switch isolation?
Back to my switch project it seems clear that layout is critical as may be
the choice of relay to minimise coupling. Whilst I think I have a stray
coupling problem there also seems to be quite a few variables at play.
Martin, HS0ZED
p.s. This switch box is for my quad project, feeding the loops ONE at a
time.
## OK... since this appears to be a 1 in 3-5 out box... for a quad... then
why do u require all the isolation?? My Array solutions 1 in 6 out box
uses
6x DPDT relays. They use 3/4" wide Cu strap..exactly 5/16" above the
chassis.
This forms a 50 ohm strip line. Input is at far left... and outputs 1-6
are to the right of this. COM of each relay is connected to the center
conductor of each coax connector [7-16 din]. ALL NC contacts are
connected to the grnd buss.. which is just
more 3/4" wide cu strap..running the entire width of the box. ALL the
NO contacts
are connected to the hot 2 kw buss. [ more cu stap]. Far left hand side of
this hot buss is
connected to the input connector.
## in normal operation... any relay that has it's coil DE-energized.. [ not
operated]
has it's ant center conductor grnded ! Any relay operated has it's ant
center conductor
fed to the input hot buss. There really is no port to port isolation per
se.... since all unused
ports are dc grnded. They have also paralleled all contacts to increase
power hanndling by 4 x .
## I also have the AS 6 pack box... uses 12 x relays. They have the
diagram of it on the AS site.
These box's use 12 x 3PDT relays. The 3rd set of contacts is used to
control the ..'master-slave'
function.. so which ever of the 2 x ops gets the ant 1st.... locks out the
2nd op. There is loads
of isolation with the remaining contacts. Each of the 6 x ant ports
has a PAIR of 3PDT relays.
## If you just require a remote switch box to save on coax... and toggle
ONE xcvr between a bunch
of DE's on your quad... then almost anything will suffice. Isolation in
that case is a non issue.
later........ Jim VE7RF
_______________________________________________
_______________________________________________
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
|