Thanks for the info on 802.11a. I haven't had an opportunity to study it
much.

If the delay-spread limit for 802.11a is only 800 ns, then it's obvious
802.11a was designed only for indoor use (or perhaps outdoor LOS links
with directional antennas only).

Outdoor systems such as pt-mpt systems and cellular encounter multipath
delays of 3 or 4 ms, up to 8-10 ms max [please see papers by Cox,
Devarsivatham, and many, many others].  No wonder you don't see very
good performance with 802.11a outdoors (aside from the power
limitations).  

Many other OFDM systems were explicitly designed for outdoor channels
(one sterling example is the European HDTV system for which OFDM was
chosen specifically because it's great at handling multipath).


In a line-of-sight (LOS) path you have a very strong direct
first-arrival signal, followed by much weaker delayed multipath energy
(multipath happens because signals get reflected off objects or get
refracted around objects, reflection and refraction are lossy
phenomenon).  If a system only requires a signal-to-interference ratio
(SIR) of 10 dB or so (as is typical), only the multipath that's above
-10 dB compared to the strong direct first-arrival signal will cause any
concern; the rest doesn't affect the modem.  Hence what we call the
"multipath delay spread" (i.e. the time over which energy that's of
concern to modem is spread) is usually very short.

In a non-LOS path you do _not_ have a very strong direct first-arrival
signal.  Instead, all you have are weaker reflected or refracted
signals, and you have many more delayed multipath signals that are
within 10 dB of the strongest signal (or whatever threshold is important
for the modem you're using), hence the multipath delay spread for
non-LOS paths is usually much longer than for LOS paths.  

In other words, if you have a strong signal you usually get to ignore
most of the multipath, but if all you have is weak multipath, you have
to pay more attention to the multipath!

Think about it.


Greg





On Thu, 2003-07-24 at 23:16, Jim Thompson wrote:
> 
> Greg DesBrisay writes:
> 
> > Criticisms from others about 802.11a on this list don't seem to be
> > related to OFDM, they seem to be related to other issues of receiver and
> > transmitter design.
> 
> 802.11g or 802.11a deals well with delay spread until you smack up
> against the GI (800ns), then it falls apart.
> 
> In a situation where there is no LOS signal, the delay spread will be
> smaller, on average.  (Think about it.)
> 
> This is not a 'criticism' of 802.11a or 802.11g.
> 
> Jim
> 
> -- 
> "Speed, it seems to me, provides the one genuinely modern pleasure."
>                       -- Aldous Huxley (1894 - 1963)


--
general wireless list, a bawug thing <http://www.bawug.org/>
[un]subscribe: http://lists.bawug.org/mailman/listinfo/wireless

Reply via email to