In addition to Richards comments

As a guide I use to run two spread spectrum links. One on 2.4GHz over
about 10km and another on 5.6Ghz over 8km.

The 5.6GHz link was one of those Cisco patch panel things with 30dBm
EIRP. The RF power was about 8dBm. We had 56MB/sec about 90% of the
time. (Including the BER) (keeping in mind that this is aggregate)

The 2.4GHz link was initially setup wrongly. There is a parameter one
has to set that defines the max distance of the link I think to reduce
packet retries and collisions. When it wasnt set the rate was a real bad
and flakey 1MB/sec but when fixed 11Mb/sec was good about 80% of the
time (incl BER)

What eventually killed the 2.4GHz link was mainly other users on the
same freq. The radio design didnt seem to allow it to hop away from
interfering signals. A cold power boot often resolved the issue as it
chose another clearer freq. We eventually dropped it to 2MB/sec with
about 50% reliability. We didnt really have any major multipath problems
that were noted in the design phase. We did however have a building go
up in the path and for a while were firing between two concrete floors!
(We moved one end later)

We used a 2 metre gridpack horiz polarization at each end (to avoid some
user interference). One end had a 16m run of LMR400, the other about a
12m run. I dont remember the calcs/margin we did off hand, sorry. We
didnt however exceed the 30dBm EIRP legal limit. (The company had a very
good standing with the ACA/ACMA so we were kind of pedantic about doing
it right)

Both links were kind of high point to high point accross Sydney. ie
There was a large series of valleys between each site.

Hope you find this helpful

Cheers Bob

Tekmanx wrote:

Also, I heard 802.11g sucks outdoors. This true? And would you guys say
my 400mw radio is overkill for 4-10mile shot?