I have been in this business for a long time and worked with Lightsquared's predecessors on the concept of a combined satellite/terrestrial network. The frequencies then involved were not the same as they are now.
A satellite system is a very good proposal when combined with a terrestrial system.
Terrestrial base stations need to put out a much stronger signal because they are generally eclipsed from the mobile phone, unlike the satellite signal, which is line-of-sight. Furthermore, the satellite and terrestrial components don't interfere with each other even in the same frequency band, because you only use the satellite when you can't receive a terrestrial base station signal. But that only applies between cooperating systems. GPS is not a cooperating system, and was there first. It is not on exactly the same frequency, but is darn close. GPS needs interference free use of 1574.9 - 1576 MHz for the C/A code, and 1570 - 1580 for the precision code.
The precision code I believe is still unavailible to civil users but is used by the military and aircraft.
Lightquared wants to put a network of terrestrial base stations up that would transmit wideband LTE OFDM signals in the 1525-1559 MHz band. That is pretty darn close to GPS. Furthermore, aircraft would receive the base station's signals line-of-sight,
and therefore uneclipsed by the clutter that their power was raised to deal with. The aircraft would therefore receive an enormous interfering signal, and not just from one base station, but from every base station within a circle of about 250 miles radius. That's a lot of base stations.
GPS receivers have to have very high sensitivity and cannot tolerate the use of lossy front-end filters. They never had to, because only other weak satellite signals were originally in this band. Therefore they would be prone to interference from excessively strong nearby signals. Lightsquared's idea of retrofitting them with a filter is total rubbish. If you looked inside your Garmin or Trimble or even GPS equipped I-Phone, you would rightly ask "where the hell do you start to modify them to add an additional component? And who pays? It's a total non-starter and Lightsquared needs to give up making that stupid suggestion. A precedent from a previous, similar conflict suggests that what Lightsquared would have to offer would be to replace, at their cost, every GPS in existence (in North America at least) with one that was not suceptible to interference from their base stations.
There are two interfering problems they have to deal with: They have to stop GPS devices being interfered by the strong signals they would ope to transmit on their allocated frequencies of 1525-1559 MHz, and they have to refrain from transmitting substantially anything whatsoever in the GPS band of 1575.42 +/-5 MHz. I would say, 50 yards from their tower, the power at 1575.42 MHz needs to have dropped to -106dBm. If they are starting with 100 watts and an 18dB antenna gain, that's 174dB down. 70dB of this comes from the 50 yard separation, leaving a spec of -104dBC for transmitter out-of-band products. That will be tough and they will need a very clean transmitter and some filters to get there. That may be doable. The former problem may not be doable - that is retrospectively modifying all GPS receivers in North America.
That does not mean Lightsqared is dead in the water. They just need a different spectrum plan. They can use their downlink spectrum for satellite transmissions, and might be able to use it for very low power terrestrial indoor picocells for example.
Maybe in the 1 milliwatt range, like Bluetooth. You can't get a GPS signal indoors anyway, and who needs GPS if you are indoors - you know where you are then, I suppose!
So they have to move the frequency of their terrestrial macrocells somewhere else - away from GPS. They should look at going time-duplex using part of their uplink band. That has its own problems, but there is a Chinese version of LTE that uses time duplex.
Whatever, networks of huge towers are no longer current technology. They are what the satellite replaces. LTE range to cellphones in the 100MB/s datrate region is less than 1km anyway, so the name of the game is very dense infrastructire of micro-and picocells.
They need a system architect who can start with a blank sheet of paper to save their asses.
