"james.thompson" <jwthompson2@gmail.com> hath wroth:
>I am planning on deploying a test 802.11g point-to-point system to
>hopefully replace an old coax network.
You're always better off with wire, coax, or fiber. The problem is
not rain fade. It's interference from other users on the 2.4GHz band.
Interference is a problem you don't normally consider on a wired
network, but which is of paramount importance on a wireless link.
Excessive interference is also why many commercial users have switched
from uncontrolled unlicensed operation, to licensed microwave links.
>I am wondering what the impact
>can be from rain, fog and snow.
|
http://www.telex.com/Wireless/faq.ns...%20Propagation
Lots of good references and reading.
>Where I am located gets all three and
>in abundance and variety.
By the lack of specifics, I presume you only want additional reading
material and not specific calculations. If you want help with
calculations, numbers are usually helpful.
>A full deployment on site would consist
>eventually of multiple point-to-multipoint and a handful of
>point-to-point links with the distances ranging from a few hundred feet
>to almost a mile.
Once you get beyond a few hundred feet, you'll probably need to go
with directional or sector antennas. This will make the link topology
rather complex. You can probably survive with an omnidirectional
antenna at the Point to Multipoint hubs if interference is minimal.
The problem with omni antennas is that they pickup interference from
all directions, while directional panels and sector antennas minimize
the areas from which interference might arrive. If the "full"
deployment involves a large number of outside radios, you are likely
to create much of your own interference problem.
>The longest run would have to contend with other
>issues too such as NLOS due to a tree that we can't quite get high
>enough to shoot over.
Trees are not fatal but difficult to deal with. A very rough guess is
that one medium size tree that you can see through is good for about
0.5dB of attenuation per meter at 2.4GHz. It can be radically more or
less depending on season and type of tree. The real problem with
trees is that you can usually find a hole in the foliage to shoot
through, but that the signal will not be stable because it infringes
on the Fresnel Zone. At 2.4GHz, you need a much larger hole than what
you can see through. For example, at 1 mile, at midpoint, you need
about a 19ft radius zone that's clear.
|
http://www.terabeam.com/support/calc...esnel-zone.php
>I have read a few posts on the subject that suggest rain and snow don't
>have quite as much impact as fog but I am wondering if anyone can point
>me in the direction of quantified impact?
In my limited experience, rain, fog, and such are minor problems at
2.4GHz as compared to interference, lack of fade margin (SOM), and
sloppy installations. For example, deploying a wide are network in an
area infested with a municipal wireless network. The fade margin
problem is that many links are deployed with limited calculations. I
try to aim for an minimum of 20dB fade margin. This is rather
difficult to obtain with low power hardware. Try your longest link on
this procedure from the FAQ:
|
http://wireless.wikia.com/wiki/Wi-Fi#Link_Calculations
and see how close you get. For the time being, ignore fog, trees, and
such. When you arrive at a first approximation for fade margin, just
subtract out your worst case estimates for these in dB. If you end up
with a lower number than 20dB, you may have problems.
Note that you can relate fade margin to link reliability:
Reliability Fade Margin
90% 8 dB
99% 18 dB
99.9% 28 dB
99.99% 38 dB
99.999% 48 dB
Sloppy installation is the most common problem. Water in the coax
connectors and cable can really ruin a wireless link. I have
reinstalled a few wireless links for waterproofing (and alignment
errors) that suggest that the problem is all too common.
>I'm intending to make sure
>that the links have more than enough "power"
Possibly wrong. Increasing transmit power at one end of a link
creates an "alligator" which is an animal with a big mouth and small
ears. It transmits much farther than it can hear unless you also
increase the power at the other end of the link. If this Point to
Multipoint system is properly designed, it will have equal xmit power
at all points of the system. That's not always possible or
economical. Anything less creates a wide area jammer. As I
previously noted, you may be your own primary source of
self-interference.
>but I was wondering how
>much would be an appropriate margin for the various types of weather
>events that can impact 2.4GHz signals.
|
http://www.telex.com/Wireless/faq.ns...opagation_rain
I use 0.1dB/mile for fog. No big deal.
>Also, what impact does light
>foliage cause?
|
http://www.utexas.edu/research/mopro...pter02-toc.htm
The results for L band should be sufficiently close. Otherwise, use
the hints for frequency scaling.
>Can it be "blasted through"
No. Or rather, yes it will work for a while, but it will not be
stable unless there is sufficient fade margin.
>or is some other solution
>more appropriate?
Cables, wires, or fiber. Licensed microwave. I think FSO (free space
optics) are out due to fog.
--
Jeff Liebermann
jeffl@comix.santa-cruz.ca.us
150 Felker St #D
http://www.LearnByDestroying.com
Santa Cruz CA 95060
http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
Snow/Fog/Rain Guidelines? 
Linear Mode
