Help With Antenna Diversity

[Michael Shuler]

I'm trying to understand how the tx/rx diversity works if anyone would
be kind enough to enlighten me about their experiences I would really
appreciate it.

Here is what I have tried so far:
1 -> Installed 2 Maxrad 10db omni's on standoff arms at about the top of
a 100ft tower spaced 15ft apart horizontally, ran 2 separate Andrew's
1/2" Andrews Heliax cables to each antenna.

2 -> Attached one antenna cable to first port on Zcomax 200mw card.

3 -> Sampled signal at several locations around the tower (about 2500ft
from the tower; not all are line-of-sight to both antennas and some
times neither antennas have line-of-sight; spaced locations in an even
circle around the tower; took 8 samples)

4 -> I then ran 'prism2_param wlan0 antsel_tx 1' and 'prism2_param wlan0
antsel_rx 1'.

5 -> Reran tests from step 3

Results after diversity seemed to be exactly the same...or negligible
differences.  I was expecting something amazing to happen and was quite
disappointed.  Am I doing something wrong or is antenna diversity a way
for antenna manufacturers to sell more antennas?  

Is there a way to see tx/rx stats based off by the antenna instead of
the card as a whole?  That way I would know for sure that both antennas
were doing something or if something was wrong with one of them.  


Here is what I am thinking of trying but want to check if I am wasting
my time:
Use two Zcomax 200mw cards and 4 90 degree sector panels (two to each
card, one to each antenna port).  One card would be set to channel 1 the
other to 11 and tx/rx diversity would be turned on.  This 'should' give
me 360 degrees of coverage and all of the benefits of sector panels
(smaller window of noise, higher gain, etc).  The only disadvantage
should be only 2 of 4 panels could listen/transmit at a time...this also
could lead to high collisions since a station wanting to transmit on an
inactive panel would think it was clear to transmit, when in reality the
radio card may already be transmitting or receiving on its other panel
and in turn would cause that client to retransmit...which I guess
wouldn't be the end of the world since the other panel would barely
pickup the nose from that client.  

Thanks,
Mike