Synchronization

In spread-spectrum communication systems, the receiver must be perfectly aligned with the transmitter in both time and frequency before any data can be decoded. This alignment process is called synchronization, and it is one of the most critical operations in CDMA and other spread-spectrum systems. Without synchronization, the locally generated PN code at the receiver does not match the incoming signal, and the despreading process fails entirely.

Core Concept of Synchronization

Synchronization in spread-spectrum systems means aligning the receiver's locally generated PN (pseudo-noise) code with the incoming spread signal in both time (chip timing) and frequency (carrier). If the code is off by even one chip, the cross-correlation output drops sharply, and the signal cannot be despread.

The synchronization process is split into two distinct stages. The first is code acquisition, which finds the approximate timing offset of the incoming PN sequence. The second is code tracking, which continuously maintains and refines that alignment in the presence of channel variations, Doppler shifts, and noise.

Acquisition is inherently a search problem. The receiver tests different timing hypotheses one by one (or in parallel) until the correlation output exceeds a detection threshold. Once a candidate is found and verified, the system transitions to the tracking loop, which operates continuously during data reception.

Mathematical Expression

The core of acquisition is the cross-correlation between the received signal and the local PN replica. If the received chip sequence is c(t) and the local replica is c(t - tau), the correlation function is:

R(tau) = (1/T) integral from 0 to T of c(t) * c(t - tau) dt

For a maximal-length PN sequence of length N chips, R(tau) = 1 when tau = 0 (perfect alignment) and R(tau) = -1/N for all other chip-aligned offsets. This near-zero off-peak correlation is what makes CDMA work. The acquisition search covers all N possible timing offsets, and the decision is made by comparing R(tau) to a threshold.

Once acquisition is complete, a Delay-Locked Loop (DLL) or Tau-Dither Loop handles tracking. The DLL generates early and late versions of the local PN code and computes a discriminator error signal. This error drives a voltage-controlled clock to keep the local code phase aligned continuously.

Practical Understanding

Acquisition time directly affects how quickly a phone can connect to a base station. In IS-95 CDMA systems, the acquisition search can take hundreds of milliseconds in worst cases. Modern systems use aided acquisition (GPS time, network assistance) to dramatically reduce search time.

Tracking must compensate for Doppler frequency shift when the user is moving. A carrier-tracking loop (Costas loop or PLL) runs alongside the DLL to handle frequency offset. In GNSS receivers, this dual-loop architecture is the heart of every satellite signal processor.

The mean acquisition time depends on the number of cells N in the search space and the dwell time Td per cell. For a serial search:

Mean acquisition time = Td * N * (2 - Pd) / (2 * Pd), where Pd is the probability of detection per dwell.

Given:
PN sequence length N = 1023 chips
Chip rate Rc = 1.2288 Mcps (IS-95 standard)
Dwell time Td = 1 ms per hypothesis
Probability of detection Pd = 0.9

Why this formula applies:
Serial search tests each timing hypothesis one at a time.
Mean acquisition time accounts for misses requiring re-search.

Formula:
T_acq = Td * N * (2 - Pd) / (2 * Pd)

Substitution:
T_acq = 0.001 * 1023 * (2 - 0.9) / (2 * 0.9)

Calculation:
T_acq = 0.001 * 1023 * 1.1 / 1.8
T_acq = 0.001 * 625.5

Final Answer: T_acq = 0.6255 seconds (approx 625.5 ms)

Exam Tip: For GATE, remember that DLL tracking uses early-minus-late correlation as the error signal, and the PN code peak correlation equals 1 only at zero offset. Off-peak correlation for m-sequence of length N is -1/N, not zero.

Mechanism: Acquisition and Tracking Steps

• Serial search acquisition: receiver shifts local PN code by half-chip steps and checks correlation magnitude at each offset until threshold is crossed.
• Verification stage: once a candidate offset is found, a longer dwell confirms it is a true detection and not a false alarm.
• DLL takes over after acquisition: generates early (advanced by half chip) and late (delayed by half chip) PN replicas.
• Discriminator computes E - L difference: positive error means local code is late, negative means it is early, zero means perfect alignment.
• Loop filter smooths the error and drives the voltage-controlled clock to advance or retard the local code phase.
• Carrier tracking runs in parallel via a Costas loop to remove Doppler frequency offset before despreading.

Quick Revision

• Synchronization = acquisition (coarse timing) + tracking (fine, continuous alignment).
• PN cross-correlation peak = 1 at tau = 0, off-peak = -1/N for m-sequence of length N.
• Mean serial search acquisition time: T_acq = Td * N * (2 - Pd) / (2 * Pd).
• DLL uses early-minus-late discriminator; positive error = local code is late.
• Costas loop handles carrier frequency (Doppler) tracking; DLL handles code phase tracking.
• GATE trap: acquisition time increases linearly with PN sequence length N and dwell time Td.
• Parallel acquisition reduces search time but increases hardware complexity proportionally.