MSK Modulation

Minimum Shift Keying (MSK) is a special form of Continuous Phase Frequency Shift Keying (CPFSK) where the frequency deviation is chosen to be the minimum value that allows the two frequencies to be orthogonal. MSK achieves this with a modulation index of exactly 0.5, making it highly bandwidth-efficient while maintaining a constant envelope property — crucial for use with nonlinear power amplifiers.

Core Concept of MSK

MSK belongs to the family of Continuous Phase Modulation (CPM) schemes. Unlike conventional FSK where the carrier frequency switches abruptly between symbols (causing phase discontinuities), MSK ensures the phase of the carrier is always continuous across symbol boundaries. This continuity is enforced by constraining the modulation index to exactly h = 0.5.

When bit '1' is sent, the carrier uses frequency f1 = fc + 1/(4Tb). When bit '0' is sent, it uses f0 = fc - 1/(4Tb). The frequency difference between the two is delta_f = 1/(2Tb), which is the minimum separation that guarantees orthogonality over a bit period Tb. This is the origin of the name Minimum Shift Keying.

The constant envelope property (amplitude never changes) is critical for practical systems. When a signal with amplitude variation is passed through a nonlinear power amplifier, spectral regrowth and intermodulation distortion occur. MSK avoids this, making it suitable for power-efficient amplifier configurations.

Mathematical Expression

The MSK signal is expressed as: s(t) = A cos(2π fc t + φ(t)), where the instantaneous phase is φ(t) = ±(π/2Tb)t + φ0. The positive sign is used for bit '1' and the negative sign for bit '0'. The phase changes linearly within each bit period by ±π/2.

The modulation index h is defined as h = 2 * delta_f * Tb = 2 * (1/(2Tb)) * Tb = 1/2 = 0.5. For conventional BFSK, h is typically 1.0 or greater. MSK uses h = 0.5, which is the minimum value that maintains orthogonality.

MSK can also be viewed as Offset QPSK (OQPSK) with half-sinusoidal pulse shaping. The I and Q components each carry alternating bits, and the bit streams are offset by Tb/2 (half a bit period). This interpretation is useful for understanding the receiver structure and the bandwidth properties.

Practical Understanding

The null-to-null bandwidth of MSK is approximately 1.5/Tb, which is narrower than conventional FSK (which requires at least 2/Tb for h=1). The main lobe of MSK's power spectral density is more compact, but the sidelobes are still significant. This is why Gaussian MSK (GMSK) was developed — to further reduce the sidelobe energy using a Gaussian prefilter.

The BER performance of MSK in AWGN is identical to BPSK and QPSK: Pb = Q(sqrt(2Eb/N0)). This is because MSK achieves the same minimum distance as coherently detected BPSK, even though it uses two frequencies rather than two phases.

MSK can be demodulated coherently (requires phase reference) or differentially (no phase reference needed but slight BER penalty). Coherent MSK achieves optimal performance, while differential MSK (DMSK) trades off about 3 dB in SNR for implementation simplicity.

Solved Numerical Example

For an MSK system with a bit rate of 10 Mbps, find the carrier frequencies f1 and f0 used for bit '1' and bit '0' respectively, given the carrier frequency fc = 900 MHz.

Given:
Bit Rate Rb = 10 Mbps
Bit Period Tb = 1/Rb = 100 ns
Carrier Frequency fc = 900 MHz
Modulation Index h = 0.5

Why this formula applies:
In MSK, frequency deviation delta_f = 1/(4Tb) from center frequency.
f1 = fc + delta_f, f0 = fc - delta_f.

Formula:
delta_f = 1/(4Tb)
f1 = fc + 1/(4Tb)
f0 = fc - 1/(4Tb)

Substitution:
delta_f = 1 / (4 x 100 x 10^-9) = 1 / (400 x 10^-9)

Calculation:
delta_f = 2.5 MHz
f1 = 900 + 2.5 = 902.5 MHz
f0 = 900 - 2.5 = 897.5 MHz

Final Answer:
f1 = 902.5 MHz (for bit '1')
f0 = 897.5 MHz (for bit '0')
Frequency separation = 5 MHz = 1/(2Tb)

Exam Tip: For GATE, remember that MSK has h = 0.5, and its BER is the same as BPSK in AWGN. A common trap is assuming MSK is inferior to BPSK in noise performance because it uses FSK — this is wrong. The minimum frequency separation in MSK is 1/(2Tb), NOT 1/Tb. Do not confuse MSK with conventional FSK where h = 1.

MSK vs FSK vs BPSK Comparison

• MSK uses h = 0.5 — minimum modulation index for orthogonality, resulting in frequency separation of 1/(2Tb).
• Phase is always continuous at symbol boundaries, unlike conventional FSK where phase jumps occur.
• Constant envelope allows use with nonlinear (efficient) power amplifiers without distortion.
• BER of MSK in AWGN equals that of BPSK: Pb = Q(sqrt(2Eb/N0)).
• MSK can be interpreted as OQPSK with sinusoidal pulse shaping — useful for receiver design.

Quick Revision

• MSK is CPFSK with h = 0.5. Frequencies: f1 = fc + 1/(4Tb), f0 = fc - 1/(4Tb).
• Minimum frequency separation delta_f = 1/(2Tb) ensures orthogonality with continuous phase.
• BER = Q(sqrt(2Eb/N0)) — same as BPSK and QPSK in AWGN.
• Constant envelope: amplitude never changes, ideal for nonlinear amplifiers.
• Null-to-null bandwidth = 1.5/Tb, narrower than conventional FSK.
• Exam trap: h = 0.5 does NOT mean the signal has half the noise immunity. It refers to the modulation index, not power. BER of MSK is optimal.
• MSK is the precursor to GMSK used in GSM — GMSK adds a Gaussian filter to further reduce sidelobe energy.