Field Guide · concept

Also known as: TDD, Time-division duplex

Time-division duplex (TDD) is a two-way radio scheme in which the uplink and downlink share one frequency, taking turns in time slots separated by a short guard period.1 Because both directions use the same channel, TDD needs no paired spectrum, and the split between uplink and downlink time can be tuned to match asymmetric, download-heavy traffic. It powers TD-LTE and much of 5G NR in the mid-band and mmWave ranges, and it stands in contrast to frequency-division duplex.

time → one frequency DL UL DL UL DL guard
TDD keeps a single frequency and alternates downlink and uplink in time; a small guard period between them lets one side stop before the other starts. The DL/UL ratio can be skewed toward download.

How it works

A TDD frame is divided into slots that are assigned to the downlink or the uplink on a repeating schedule. At any instant only one direction is transmitting on the channel, so transmit and receive never overlap and no duplexer filter pair is required — the radio simply switches between transmitting and receiving. Between a downlink burst and the following uplink burst the network inserts a guard period: a brief silence that lets the previous transmission’s signal clear the air (accounting for propagation delay across the cell) before the other end begins, preventing the two directions from colliding. This is the same time-sharing idea as TDMA applied to the duplex direction rather than to separating users.

The two big advantages are unpaired spectrum and an adjustable split. A regulator can allocate a single contiguous block, and the operator decides how many slots go to download versus upload — a natural fit for modern traffic, which is heavily download-biased. The costs are tight synchronization (every cell must agree on the frame timing, or one cell’s uplink will clash with a neighbour’s downlink) and the guard time itself, which is spectrum-efficiency lost to the turnaround and grows with cell size.

TDD in cellular

TD-LTE (the TDD variant of LTE) and the majority of 5G NR mid-band and millimetre-wave deployments use TDD, because the prime new spectrum for those systems came in large unpaired blocks and because massive-MIMO beamforming benefits from channel reciprocity — the uplink and downlink share the same frequency, so the base station can infer the downlink channel from what it measured on the uplink. The guard band at the block edges still separates operators in frequency, while the guard period separates the directions in time. The trade-off against FDD is fundamental: TDD wins flexibility and unpaired-spectrum use, FDD wins continuous low-latency symmetric links.

Relevance to SDR

On a spectrum display a TDD signal looks different from an FDD pair: instead of two mirrored bands, a single block pulses on and off as the frame alternates direction, and a waterfall shows the busy downlink slots interleaved with quieter uplink ones on the same centre frequency. Recognising that on/off, single-band pattern is the quickest way to tell TDD from FDD when surveying spectrum. GopherTrunk decodes land-mobile trunking rather than cellular, but the time-sharing principle behind TDD is the same one that lets a single trunked channel carry alternating inbound and outbound traffic, so it is useful context for reading two-way RF.

Sources

  1. Time-division duplex — Wikipedia, for the definition of TDD, the guard period, and the adjustable uplink/downlink split. ↩

See also