LightSquared's national rollout has shifted frequencies to avoid interfering with GPS kit, taking early advantage of a deal with Inmarsat for 10MHz of bandwidth further down the dial.
The wannabe operator, which aspires to cover the USA in an LTE network for wholesale leasing, has also offered to cut its permitted transmission power in half, back to what was permitted before the FCC started mucking about with the licences. That should avoid interfering with GPS, but it will also make building the network a good deal more complicated.
LightSquared doesn't even have access to the frequency in which it now plans to deploy a national network - that spectrum is owned by Immersat who had agreed that LightSquared could use it in a few years for increasing capacity.
LightSquared admits it will now have to "accelerate the schedule... to begin using the frequencies", while the band in which it originally intended to launch will sit fallow until needed, and even then it will be at much-reduced power.
That should avoid interfering with all but the most sensitive of GPS receivers, though we'll have to wait until 1 July to see how sensitive as that's when the FCC-mandated investigation into the problem publishes its results. LightSquared asked for, and got, a two-week delay on those results, presumably in order to get this plan announced.
The problem is that LightSquared's original frequency is right next to that used by GPS. That was OK when it was only used for satellite-to-earth communications which are, by necessity, very low power, but when the FCC changed the licence to allow LightSquared to build a national mobile network in the band alarm bells began to ring.
One generally talks about radio transmissions being square - starting at full power and remaining at that point for the width of the signal before dropping to nothing - but in reality they are bell-shaped with full power only being achieved at the centre point and the tails of the bell extending into the neighbouring bands. A great deal of work is put into making signals square, and they are much squarer than they used to be, but neighbouring spectrum users still have to watch each other's toes.
LightSquared thought it could create a signal that was square enough to avoid leaking significantly into the neighbouring band, despite the comparative weakness of the GPS broadcast. That has obviously proved impossible, so LightSquared has little alternative but to shift bands and is lucky to have another place to call home.
That shift will increase the complexity of the network, and the LTE handsets needed to support it, and will increase pressure on the company to confirm its deal with Sprint if it's going to restore some credibility.
Adding another frequency to LTE's roster might seem problematic, but LTE is already expected to be deployed in more than 40 different frequencies, and it's very questionable how many bands will be supported by different handsets - the tri-band handsets of the GSM days aren't going to cut it this time around. ®