The PCI formula
NcellID = 3·N¹ID + N²ID
where N¹ID ∈ {0,1,...,335} is the physical-layer cell-identity group and N²ID ∈ {0,1,2} is the physical-layer identity within the cell-identity group.
Applying this to our example:
From PSS: N²_ID = 1 ← sequence index that gave correlation peak From SSS: N¹_ID = 147 ← candidate index that gave correlation peak PCI = 3 × N¹_ID + N²_ID = 3 × 147 + 1 = 441 + 1 = 442 // Reverse check: N²_ID = 442 mod 3 = 1 ✓ N¹_ID = ⌊442 / 3⌋ = 147 ✓
Why 1008 PCIs?
With N¹_ID ranging over 336 values and N²_ID over 3 values, the total number of unique PCIs is 336 × 3 = 1008. This is not arbitrary — it is a careful engineering trade-off between network planning flexibility and UE processing complexity.
In dense urban deployments, nearby cells must have different PCIs to avoid PSS collision (same N²_ID would confuse UE timing), SSS collision (same N¹_ID causes interference in SSS detection), and reference signal collision (same PCI means identical scrambling seeds — DMRS interfere coherently). With 1008 PCIs and typical cell density, the reuse distance is large enough that interference is negligible.
Each N¹_ID value "owns" exactly 3 consecutive PCIs, one per N²_ID:
PCI seeds everything in the cell
Once PCI = 442 is known, the UE can generate every reference signal in the cell. The PCI is used as a seed for pseudorandom sequence generators that scramble PBCH, PDCCH, PDSCH, and all DMRS signals. Two neighbouring cells with the same PCI would have identical reference signals — they would coherently interfere with each other, making reception impossible.
// PBCH DMRS scrambling (TS 38.211 §7.4.1.4.1) v = PCI mod 4 = 442 mod 4 = 2 ← DMRS subcarrier offset // PBCH scrambling seed (TS 38.211 §7.3.3.1) c_init = PCI = 442 // PDCCH DMRS (TS 38.211 §7.4.1.3.1) n_ID = PCI = 442 (unless overridden by higher layer) // SSS sequence parameters (already used) m₀ = 15×⌊147/112⌋ + 5×1 = 20 m₁ = 147 mod 112 = 35 // CSI-RS, SRS, PUSCH DMRS — all seeded by PCI when not overridden
Interactive PCI calculator
N¹_ID = 147