Thanks for the data, Andy.  The longer path to the tower is the worst case.
The shorter link puts much more power onto the passive repeater (about 40dB)
so that path may have a very good margin.  I saw some active boosters of
about 2 watts (about $300) that could be used at the site more distant from
the tower (ie, the 3.88 mile leg) for an extra 23 or so dBm onto the passive
repeater.  If the actual path loss numbers are as marginal as I think, that
extra boost at the client's site should be enough to make it all quite
reliable (ie, with suitable margins).

Joel.. for a small fee like lunch and a great deal on a Cisco card I'll help
you run the numbers and do a field test if things look OK.


> -----Original Message-----
> From: tcwug-list-admin at tcwug.org [mailto:tcwug-list-admin at tcwug.org]On
> Behalf Of Andy Warner
>
> Chuck Cole wrote:
> >
> > Estimating path loss will indicate how feasible this may
> be.  Some handbooks
> > have graphs for this.  The transmitter power is radiated
> into the solid
> > angle of the antenna pattern and is received by the area of the next
> > antenna.  Assume that the passive link has zero dB loss.
> Then the next
> > segment has an area loss also.  The loss for the longer
> segment alone looks
> > like more than 60dB, or around 80dB overall.  I don't know
> what the receiver
> > requires as signal level, but it might be -120dBm.  With all these
>
> Assuming that Joel is thinking of using Aironet stuff, Cisco
> claims the following:
>
> 	1 Mbps: -94 dBm
> 	2 Mbps: -91dBm
> 	5.5 Mbps: -89 dBm
> 	11 Mbps: -85 dBm
>
> (source:
> http://www.cisco.com/warp/public/cc/pd/witc/ao350ap/prodlit/a3
> 50a_ds.htm)
>
> So, for a 20dB output radio, you've got a total link budget
> of 105dB to maintain 11 Megs. Add a fudge factor of at least 10dB
> and you've got 95dB of budget to deal with.
>
> Might work, might not - only one way to really tell :) Certainly
> no way for us armchair critics to tell with greater certainty from
> the data provided.

A real calculation with antenna spec data would produce a good path loss
prediction.  I'm sure that data can be had but it hasn't been posted here.
That calculation is worth doing, but it's only an estimate since "lumps in
the beam" can upset real accuracy (ie, real antennas aren't perfect, etc).


 ---
Chuck