Re: why FSK encoding?
Guys, I think he is trying to ask why most FSK systems use a Manchester code, or in more advanced system some sort of bit injector.
FSK systems transmit data based on what frequency is transmitted. In a simple two-level FSK system, an example of this could be: "0" might be 9.999 GHz, and a "1" might be 10.001 GHz. In other words you deviate +/- 1 MHz in modulation.
The local government is pretty particular about what transmit frequency you use. They do not want you to transmit on a frequency outside of your authorized band. In the above example, lets say you got a license to transmit at 10.000 GHz. So you would use a phase locked loop (PLL) to control the "average" transmit frequency to be 10.000 GHz. Your PLL would be designed to have a slower response time than your modulation, so it would look at the various random combinations of 0's and 1's being transmitted, and try to keep the average right at 10.000 GHz.
Now lets say you send a block of data that is momentarily all 0's (could be a header, or just no data for a period of time). Your transmitter would send out a continuous carrier of 9.999 GHz for that period. You PLL would not like to be stuck at 9.999 GHz, so it would retune the transmitter back to the center frequency of 10.000 Ghz. Then, eventually, you would start sending randomized data again, but the first "1" that you send would jump all the way to 10.002 GHz, a full 1 MHz above your authroized bandwidth. After some time of random 1's and 0's, the PLL would retune the average frequency to again be at 10.000. But for some period of time, you were transmitting outside of your authorized band, and either the government is now mad at you, or you could not detect those first few 1's and ended up with a lot of bit errors.
Manchester coding breaks up the data to kind of look like it is always 1's or 0's, even if a block of 0's hit it. So FSK systems often use Manchester coding, or some sort of randomizer/bit inserter to overcome this problem.