Here's a guess at what happened.
When diversity is turned on in Cisco AP 340s, the APs listen to one antenna until they don't hear a signal. Only then do the APs listen to the other antenna. This is not the optimum way to impliment antenna diversity, but it works pretty well in practice. If diversity was turned on, but the second antenna port wasn't connected to an antenna or a 50-ohm termination, then it could easily pick up low-level spurious emissions from another nearby AP. Sounds like the APs were picking up low-level spurious emissions from the other AP and getting stuck listening to the antenna ports that weren't connected to antennas! (I'm also theorizing that since antennas are tuned circuits they won't radiate or receive as much of the low-level off-frequency spurious emissions, so the problem goes away if proper antennas are connected.) Just a theory. Greg DesBrisay On Wed, 2003-06-25 at 18:30, Jon Katz wrote: > > > Just to save anybody else from going through this one... > > We had a mysterious problem with two Cisco 340 AP's on a mountaintop. They > are connected back-to-back with a crossover ethernet cable, and their > antennas point in different directions; they're also on different > channels. They stopped working one day, and in troubleshooting I noticed > that clients would associate if one or the other AP was turned on, but once > they were both turned on, the connection would drop, and no clients would > associate. The problem was resolved (eventually) by turning off antenna > diversity; apparently the leakage between the unconnected antenna ports was > causing the AP's to lock to those ports, disconnecting the external > antennas. AP's may not associate with each other, but at least the 340's > definitely listen to each other. > _____________________________________________________________________ > > Jon Katz > Center for Religion, Ethics, and Social Policy > Anabel Taylor Hall, Cornell University, Ithaca NY 14853 USA > Office: 607-255-6202 Fax: 607-255-9985 > Resident Technical Consultant, Rural Alternatives Center > El Limon de Ocoa, Dominican Republic > Email: [EMAIL PROTECTED] > http://www.sas.cornell.edu/cresp/ecopartners > _____________________________________________________________________ > > > -- > general wireless list, a bawug thing <http://www.bawug.org/> > [un]subscribe: http://lists.bawug.org/mailman/listinfo/wireless -- general wireless list, a bawug thing <http://www.bawug.org/> [un]subscribe: http://lists.bawug.org/mailman/listinfo/wireless
