> The length of the jumper cables between the cans has a profound
> effect upon the insertion loss at the pass frequency, and relatively
> little effect upon the isolation at the notch frequency.
>
> 73, Eric Lemmon WB6FLY
Eric,
I'm curious why you say this, as it contradicts what I would believe to be
the case in theory.
At the pass frequency, the Z should be very close to 50+j0, so having
the "wrong" cable length won't affect the insertion loss between each of the
pairs (or three, for a 6-pack) of cavities on either side of the duplexer.
In contrast, for proper reject notch performance, there has to be correct
phasing between cavities. The notches are effectively shorts at the notch
frequency, and if they are not repeated at 1/2 wavelength intervals between
cavities, they will tend to not align when cavities are cascaded due to
transmission line transformer effects (ever try tuning a duplexer and the
notches act like they're "chasing each other" and you can't get them to fall
into place?).
Likewise, between the antenna T and the first cavity on each side, having
the wrong cable length will cause the notch "short" to not be properly
transformed to an open at the tee (via the odd 1/4 wave section between the
tee and the first can of the opposing side of the duplexer). This will make
the insertion loss appear to go up when looking from Tx input to antenna or
antenna to Rx output, and also throw off the pass Z since the opposing side
of the duplexer is no longer "invisible" as it should be.
Regarding the question posed by others for finding cable lengths and tuning
methods:
For experimentation purposes, a line stretcher is the easiest way to find
optimum cable lengths when re-cabling a duplexer. However, you can usually
just scale the lengths of the original harness to the new operating
frequencies using simple ratios of the old and new frequencies; an error of
1/4" or so isn't going to make a noticible difference on VHF, and may even
be tolerable on UHF. The other option is that if you have a duplexer that
was, say, originally on 160 MHz and you want to move it down to 2m is to add
elbow adapters to extend the length of the cables. Even if you don't leave
them in on a permanent basis, it gives you a good approximation of how much
the cables need to be lengthened by.
As far as tuning duplexers, many manufacturers' tuning instructions give a
simple how-to using just a spectrum analyzer and tracking generator. While
this might get you close, the passband performance is almost guaranteed to
not be properly optimized. At the factory, duplexers and filter cavities
are tuned on a network analyzer so both reflection (return loss) and
transmission (passband insertion loss and reject notch depth) can be
measured and optimized concurrently. When looking at insertion loss only,
pass/reject duplexers appear to have a relatively broad pass response, but
in reality, if you look at return loss, the pass is really quite sharp.
If you don't have a network analyzer, a return loss bridge is a great, and
relatively inexpensive, piece of equipment to have to give you the ability
to measure return loss using a typical spectrum analyzer/tracking generator
in a service monitor. You do need good termination loads to go with it -
most run-of-the-mill high-power dummy loads don't have enough return loss to
yield accurate results. Spend a few bucks and get precision terminations to
screw directly onto the port(s) without patch cables, and measure them with
the return loss bridge to make sure they're good (30 dB should be the bare
minimum). There are lots of good loads floating around on the surplus
market - I see new Narda 370's show up at hamfests and Ebay regularly in the
$10 range and they usually do better than 35 dB up through UHF. I don't lug
my network analyzer out to sites, but I do keep the return loss bridge in
the truck in case I have to do a quick-n-dirty retune on-site when I can't
afford the downtime of transporting and tuning back at the shop.
--- Jeff
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/Repeater-Builder/
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
