Thoughts on LightSquared
I know this really isn't relevant to RabbitEars overall, but it's tangentially related in that it's about spectrum policy. I am sick to death of reading stories about LightSquared that do not address the technical issues. I want to republish a post I made on ArsTechnica yesterday to explain the physics of the situation to any who may not know, which seems to be just about everyone who reads these stories because they tend to be so poorly written. Here's a link to the original comment.
The interference problem is due to the laws of physics, which is, in fact, why spectrum next to the GPS band is set aside specifically for satellite usage.
There is no such thing as a perfect filter. You can achieve certain amounts of roll-off over certain amounts of spectrum, but never reach zero outside the pass-band. You have to make trade-offs between the width and sharpness of the filter and the amount of signal the filter ultimately passes through.
Since GPS does not use a high-gain parabolic dish antenna like a satellite TV service would, the receiver needs to be able to gather as much signal as possible. A filter that doesn't have as sharp of a roll-off allows more signal to be passed through in the pass-band to the receiver circuitry, and so the FCC purposely put only other satellite transmissions in the adjacent bands. This means that since adjacent signals are going to be roughly the same strength (0 dB difference), the filter does not have to be as sharp, allowing as much signal as possible through.
Now, LightSquared wants to put a signal on the adjacent band that is about 130 dB stronger than what the FCC allows. And remember, that's not a linear scale, that means that the LightSquared signal would be 10^13 times more powerful than GPS. Not only are the filters in GPS completely unsuited for that, I don't think you can make a filter that will work for that in the size profile of a typical GPS.
My experience with filters comes from the TV broadcast industry. In US digital TV broadcast, before a signal goes from a transmitter out to an antenna, it passes through a "mask filter." For a full-service TV broadcaster, the filter is required to have a 47 dB drop by 500 kHz from the edge of the channel, and a 110 dB drop by 6 MHz from the edge of the channel. The filter can be anywhere from the size of a microwave oven to the size of a hot water heater or a refrigerator, is very sensitive to changes in temperature and humidity, costs thousands of dollars, weighs tens or hundreds of pounds depending on size, and I don't even know what the insertion loss is on it. And that filter might or might not be sufficient to deal with just the interference from LightSquared's proposal, let alone the fact that you now need a parabolic dish to collect enough signal to get something usable to the receiver chip on the output of the filter. Now, of course, this TV filter is designed to handle huge amounts of power compared to the amount of power a received LightSquared signal will be, since this is coming right off the output of the transmitter, but it's to give you an idea of just what an expensive, professional filter is capable of.
So, how does one design a filter with 130 dB off roll-off over 15 MHz (LightSquared ends around 1559 MHz, GPS is at 1575 MHz) without making the filter large and heavy and/or having so much insertion loss that you now require every GPS device to have a motorized parabolic dish? More importantly, how is this inability to overcome the laws of physics the fault of the GPS industry, who built to the specifications the FCC provided them and which were, themselves, based on the laws of physics?