On the 25th of August 2007 several tests were carried out on SOWN hardware.
We tested the ZeplerNode hardware with a NL-2511CD card and two antennas - the new RF Technics and the old SUSU Omni. Both antennae achieved a 250m range from the door to the demo room in vista to the SUSU concourse.
Testing 802.11a on 5.8Ghz was not so successful when using only one antenna however when two omni antennae were used (one on the main connector and one in the aux connector) a rate of 54Mb/s was achieved over a distance of about 50m with 0 packet lost and a ~1.7ms response time. The laptop at the other end did not have a dedicated 5.8Ghz antenna. From this it is believed that many of the problems associated with 802.11a are caused by multi-path interference which is avoided by using diversity (2 antennae).
Info on diveristy: http://pyramid.metrix.net/trac/wiki/FrequentlyAskedQuestions
Also the signal appears to suffer a large amount with small movements of the laptop during the tests.
Futher testing of this took place and the results are still inconclusive. We will try once more and come to a conclusion.
Because 802.11a performance may be improved with the aid of diveristy this means we need to have 2 antennas for 11a, and one for 11b. Previously when there have only been single antennas for 11a and b on separate poles. With three antennas, attaching 3 poles to the building may be excessive. Some kind of trident arrangement may be appropriate.
11a 11b 11a | | | | | | +--------+-------+ | | | Node ------ROOF----------