The Access Point

For over $900 there is not much to an access point. It feels like an empty plastic box.

The Antenna

Same with the antenna. It looks and feels like a cheap lighting fixture.

The Antenna Location

After a bit of scouting around, the best location was determined to be on the top of the north outside wall.

Beta test with short cable

After disappointing first results where the signal barely passed the cemetery, testing the access point with a much shorter cable led to better results.

Further research shows that the original amount of cable would cause a 75% loss in signal strength. This was meant to be compensated for by an amplifier on the line. The shorter cable test run had the amplifier on it.

Coverage area from Fenwick

With the shorter cable, coverage extended out to Clyde Street. If we are to use this shorter cable instead of using the longer cable which extends out to the telecom room, we will have to either buy or make an all-weather enclosure for the access point with power and thermostat/heating element.

Coverage area from the East side

Coverage area from above

How things stand now: over the next couple of weeks we will test the access point again, this time with functioning network connection, and test reception on the ground. At this time the antenna is still on Fenwick but the access point is not.

The coverage area and power of transmission were less than we had hoped for. Still, the populated residential portion of the area falls within the radio footprint. The question will be if our power levels are enough to penetrate green leaves and roofs without exterior antennas.

There is no more powerful commercial gear available to us. One possibility is that we can make an old computer into our own version of an access point that would still be below Industry Canada power limits but have more power than the commercial gear. We are looking into this now.

Part of the problem is the large number of home routers using the 2.4 GHz frequency - more than a dozen within the top two floors of Fenwick Place alone.

The contractor advised using the less trafficked 802.11a frequency for backhaul networking purposes between nodes or from the top of Fenwick to nodes. Unfortunately this has an even shorter range than the 802.11b/g results illustrated above so it would likely have to be a point to point link with directional antennas at both ends to one or two special "seeder" nodes no further away than Clyde Street if we were to go this route.

The contractor also suggested the way to set up a mesh network "properly" would be to have an access point and antenna (as in first two pictures, c. $3000) every two blocks, mounted on a power pole or at pole height. Economically this is simply out of our price range.

Experiments with 802.11n draft specification equipment have been very successful with a clear increase in range and power over 802.11b/g. Unfortunately, since the specification for 802.11n is still draft and will be until summer 2007, 802.11n equipment purchased now may be incompatible with the final specification and/or draft equipment from other manufacturers.