Before the Nextel 800 MHz rebanding, a major regional transit agency had its main passenger transfer yard near the city center. The bus drivers began to complain about poor reception in this yard.
What They Tried First
Their maintenance provider leapt into action, checking bus radios and antennas, finding some buses had 450 MHz shark fin antennas instead of 800 MHz shark fins. They found numerous Maxtrac and GTX radios a few kHz off frequency as the crystals had aged. They didn’t find anything seriously wrong, not the actual cause of the periodic signal loss problem. Then I was called in.
What I Tried First
I asked if there was any periodicity to the problem. They said every third day or so, the problem got much worse. This was a big clue, as the 800MHz trunked system cycled between each of the three control channels about once a day. This was a feature of the controller so that no one repeater would suffer excessive wear and tear from constant control channel transmitting.
Then I was rather surprised, and the bus company was also rather surprised, to find that they were operating on a tower that was almost in another county, 40 km away. Their main operations occurred in a low elevation downtown area. The flat terrain besides that helped yield downtown signal strengths in the -100 dBm range. They just got used to the static I suppose, never realizing they were constantly on the fringes of the repeater system they paid to use.
There was a Nextel system less than 300 meters from the downtown transit terminal. The Nextel panel antenna were downtilted. As noted in a prior case study, Nextel invoked the rule allowing them to put transmitters at locations not on their license, as long as they didn’t let signal outside their licensed footprint. This meant that less than 80 dB of path loss existed to attenuate the Nextel order 50dBm EIRP signal, so the bus radios were getting blasted with -30 dBm signals while trying to hear -100 dBm signals on nearby channels.
Nextel signal strength = 50 - 20 log10(4 pi d f/c) = -30 dBm
Where d is distance in meters, f is frequency in Hz, c is speed of light in meters/second.
Chirp the Professor
I was in touch with one of Nextel’s PhD engineers, a former professor who didn’t mind trading travel for dinero. Over Nextel Direct Connect (what else) we came up with a plan to verify no excess spurious emissions were coming from Nextel. We hypothesized that the -30 dBm signals were overloading the receivers. Since this receiver overloading effect is highly non-linear and threshold-like, we supposed that turning the Nextel antennas, reducing Nextel transmit power and removing some of the antenna downtilt would help. The intermodulation / spurious / adjacent channel rejection specification of the GTX are all 65 dB, so informally we wanted to get the Nextel signal level at the bus terminal in the -50 to -60dBm range.
The next week I and the Professor were at ground level with a good spectrum analyzer while the rooftop workers reconfigured the antennas. We found a configuration that got the Nextel signals below -50 dBm instead of the -30 to -40 dBm Nextel had been at before. This was enough to eliminate audible interference on the GTX radios.
The bus system’s radio fleet was elderly and breaking down more and more. Eventually they went to a digital 450MHz system on a tower more centrally located in their service area. This is probably a great idea, as there was really no reliable way to make their low-spec GTX receivers work in a modern RF environment. This problem would have kept popping up like whack-a-mole until they got > 80 dB rejection 800 MHz radios on a closed repeater system.