In text book about FM, the bandwidth FM can be calculated by Carson's Rule: BW = 2*(df + fm) where df is the maximum frequency deviation and fm is the maximum bandwidth of modulating analog signal.
But for FSK, some engineers use BW = 2*(df + R/2) where R is the data rate of modulating digital signal and others use BW = 2*(df + R). Which one is more accurate?
1. Could you explain in more details about why choose BW = 2*(df+R/2) or BW = 2*(df+R) depends on "1. The frequency shift compared to the symbol rate."
2. Did you mean that if there is no low pass filtering before modulation, then we should use BW = 2*(df+R)? Normally for unshaped digital signals of data rate R, the maximum fundamental frequency is R/2 and next harmonic frequency is 3*(R/2). So why choose fm=R instead of fm=R/2 or 3*(R/2) where fm is the bandwidth of modulating signal?
You have two types of Frequency shift keying:
1.coherent M-ary C-FSK
bandwidth can be caculated as:
Bc-mfsk=M*Vb/2*ld M=MVm/2
2.nocoherent M-ary NC-FSK
bandwidth can be caculated as:
Bnc-mfsk=M*Vb/ld M=MVm
where Vb is bit rate , Vm symbol rate
Bandwidth for coherent C-FSK is twice smaller than for NC-FSK.