Archived from groups: alt.internet.wireless (
More info?)
On Sun, 26 Dec 2004 12:45:20 GMT, viradio
<viradio.1hvggy@WiFi-Forum_dot_com> wrote:
>> Another piece of trivia is that the cable lengths inside the box on
>> many routers are different between the two antennas. Find a picture
>> of the inside of your router, and select the antenna with the
>shortest
>> coax, or preferably, no coax cable. That's worth about 1dB which
>> might make a difference on a long distance link.
>This is also for antenna "co-phasing" or matching the two antennas to the
>load.
>Do not change the length!!!!
Wrong. Co-phasing is a term used (almost) exclusively in CB radios.
http://www.signalengineering.com/ultimate/co_phasing.html
It is also used in astronomy to describe the alignment requirement of
segmented telescope mirrors.
Co-phasing is applicable for dual CB antennas, and other such
abominations, because both antennas are operated simultaneously. This
results in some "interesting" antenna patterns and cable length
requirements to get a proper impedance match. The coax cable "phasing
harness" allegedly forms a power splitter, where the cable lengths are
critical.
However, the wireless access point has one additional piece of
circuitry that is lacking in a typical CB radio and antenna system.
It has a diversity switch located between the antennas and the radio.
The switch scans between the two antennas, looking for the best
receive signal. The algorithm varies by chipset, but is basically a
stored antenna selection based upon the last successfully received
packet.
When one antenna is selected and active, the other antenna is isolated
from the radio and shorted to ground by the PIN diode switch diversity
switch. However, the antennas are in sufficient proximity to create a
somewhat directional pattern, and I agree that this pattern is highly
dependent upon the length of internal coax cables. However, it is NOT
critical by any stretch of the imagination, as demonstrated by the
ability to orient the two antennas in any random manner and still have
a functional access point. One can even remove or short one of the
antennas, and see little effect on performance.
Anyway, since the coax cables are NOT part of some phasing, matching,
power splitting, or beam forming network, and are impedance matched at
both ends, the actual lengths are NOT critical.
Incidentally, I never suggested changing the length of the internal
coax cable. My comments were in reference to the loss between the
radio assembly and the antenna being slightly less on the port without
the added loss of the coax cable. You are probably correct about not
changing the internal coax cable length. I see some things in the
circuit that look like the diversity switch is running at a much
higher impedance than 50 ohms (to reduce switch loss), and that the
coax cable (and a few components near the antenna connector without
the coax) are part of a matching network. That will make the coax
cable length critical to insure 50 ohms at the connector.
--
Jeff Liebermann jeffl@comix.santa-cruz.ca.us
150 Felker St #D
http://www.LearnByDestroying.com
Santa Cruz CA 95060 AE6KS 831-336-2558