What "Antenna Port" Actually Means
Per TS 38.211 §4.4.1, an antenna port is "defined such that the channel over which a symbol on the antenna port is conveyed can be inferred from the channel over which another symbol on the same antenna port is conveyed." A port is a logical channel — not a physical antenna. One port may map to 1, 4, 8, 16, 32, or 64 physical antennas via a precoder.
The QCL Concept
Two ports are QCL'd when their large-scale channel properties are inferable from each other. The UE estimates these properties from the source RS (typically a CSI-RS or SSB) and reuses them for the target port (typically PDSCH DMRS) — saving compute, battery, and time.
Six Large-Scale Properties
Doppler shift · Doppler spread · average delay · delay spread · average gain · spatial Rx parameter. The first five are the time-frequency footprint of the channel; the sixth (spatial Rx parameter) is the UE's receive beam direction — critical for mmWave.
QCL Type A · The Workhorse
Type A shares Doppler shift, Doppler spread, average delay, delay spread (4 of 6 properties). Source: TRS or a CSI-RS configured for tracking. Target: PDSCH DMRS. ~95% of sub-6 GHz PDSCH uses Type A.
QCL Type B · Doppler-only
Type B shares only Doppler shift and Doppler spread. Used when the source has only Doppler info (e.g., a long-period CSI-RS). Rare in operator deployments.
QCL Type C · Tracking Bootstrap
Type C shares Doppler shift + average delay. Source: SSB. Target: CSI-RS-for-tracking. Used to bootstrap tracking RS from the broadcast SSB after handover.
QCL Type D · The mmWave Magic
Type D shares the spatial Rx parameter — i.e., "use the same UE receive beam direction." This is what makes mmWave practical. Without Type D, the UE would have to re-search for the best beam every time it received a new transmission. With Type D, the UE locks beam #N from SSB, then reuses that beam for every PDSCH/CSI-RS that the gNB has marked QCL Type D with that SSB.
TCI State · The Container
A TCI-State is a labeled QCL tuple: tci-StateId, qcl-Type1 (mandatory), qcl-Type2 (optional, typically Type D for FR2). The UE receives a pool of up to 128 TCI-States via RRC, MAC-CE activates 8 of them, DCI 1_1 selects 1 of those 8 (3-bit field) for the current PDSCH.
Beam Failure Recovery (BFR)
When the UE's active TCI state degrades below a threshold, the UE detects beam failure, scans for new beams from a candidate list, sends a beam-failure-recovery request on a dedicated PRACH, and the gNB updates the TCI state.
Frequently Asked Questions
What does QCL mean in 5G NR?
Quasi-Co-Location means two antenna ports share large-scale channel properties — Doppler, delay, average gain, and (for mmWave) receive beam direction. The UE estimates these properties once and reuses them across QCL'd ports, saving compute and battery.
What is the difference between QCL Type A and Type D?
Type A shares 4 properties (Doppler shift/spread, average delay, delay spread) — used for sub-6 GHz time-frequency channel estimation reuse. Type D shares the spatial Rx parameter (UE receive beam direction) — used only in FR2 mmWave for receive-beam selection.
How does the UE know which TCI state to use for a PDSCH?
RRC configures up to 128 TCI states. MAC-CE activates 8 of them. DCI 1_1 contains a 3-bit "TCI" field that selects 1 of the 8 active states for the current PDSCH transmission.
Why is QCL Type D only used in FR2?
Type D shares the spatial Rx parameter — the UE's receive beam direction. In FR1 sub-6 GHz, UEs typically use omnidirectional or weakly-directional antennas, so beam selection is unnecessary. In FR2 mmWave, narrow beams are essential, so Type D is required.