Dennis Ferguson wrote:
> g <wh@t.me.worry> wrote:
>> Dennis Ferguson wrote:


> Yes, this is correct, the frequency-dependent component of free space
> path loss is really just a fiction. It is, however, a useful fiction
> since it lets you talk about path losses and antenna gains (over isotropic)
> as if they were independent issues instead of just a single issue related
> to how antennas work.

Definitely it is useful, though any practical application of the model
involves incorporating the characteristics of a physically real antenna.
There's no problem at all with the model as long as one keeps in mind
what it is modeling.

> This is true, though it is probably worth pointing out that the limiting
> direction for mobile phones is not the tower->handset direction, but
> rather the handset->tower direction. The power output from the handset
> is limited, and while the power output from the tower is not the reciprocal
> performance of the antenna/path/antenna system means that increasing the
> power output from the tower much beyond that of handset doesn't help.

This is correct for symmetric uses such as two-way voice, performance is
ultimately limited by handset batteries and antenna size.

It's interesting that for at least one data usage, EVDO, this is not
true because the *entire* carrier power (20+ watts) goes to servicing
one user at a time. EVDO is not CDMA in the sense of sharing the carrier
among multiple users; it's TDMA. But for higher rate downloads the
handset need only get an ACK returned while the base has to deliver
larger information rates. This tends to even things out considerably.

As devices are required to go to symmetric broadband and higher
rates(3G+, 4G) the handset definitely is the limiting factor. In these
cases the only way to maintain performance is to improve the paths,
either by reducing the impairments (foliage, buildings etc) or greatly
reducing the path length (cell size)... which is more or less how we got
started on this thread.

g