4. FDD and TDD: How Wireless Links Share Uplink and Downlink

The Duplexing Problem

A cellular link needs two directions:

• downlink (DL): base station to user equipment, such as a video stream or webpage;
• uplink (UL): user equipment to base station, such as acknowledgements, voice, sensor data, or a camera upload.

If the transmitter and receiver use the same resource without coordination, the much stronger local transmission can overwhelm the wanted received signal. A duplexing method separates the two directions in frequency, time, or a controlled combination of both.

FDD and TDD describe this duplex separation. They do not specify the modulation, coding, antenna count, scheduler, or application by themselves.

FDD: Separate the Directions in Frequency

Frequency Division Duplex (FDD) assigns one frequency band to downlink and a different paired band to uplink. Both directions can be active at the same time.

If the downlink is centred at $f_{\mathrm{DL}}$ and the uplink at $f_{\mathrm{UL}}$, then

Where:

• $f_{\mathrm{DL}}$ is the downlink carrier's centre frequency;
• $f_{\mathrm{UL}}$ is the paired uplink carrier's centre frequency;
• $\neq$ emphasizes that FDD separates the directions onto different frequencies.

A simplified paired-spectrum budget is

Where:

• $B_{\mathrm{paired}}$ is the total illustrated span occupied by the paired arrangement;
• $B_{\mathrm{DL}}$ is the allocated downlink bandwidth;
• $B_{\mathrm{UL}}$ is the allocated uplink bandwidth;
• $B_{\mathrm{gap}}$ is the duplex separation or unused spectrum between them.

The exact arrangement is defined by the applicable band specification and is not necessarily one contiguous operator allocation.

How FDD Works

1. The base station transmits in the downlink band.
2. The UE receives that band while it may transmit in the separate uplink band.
3. Duplex filters isolate the transmit and receive paths.
4. The scheduler still decides which users receive time–frequency resources inside each band.

The two directions are separated continuously in frequency, so FDD does not need a DL-to-UL switching guard between every scheduled transmission interval.

Where FDD Is Useful

• wide-area coverage layers where paired spectrum is available;
• services with persistent traffic in both directions;
• mobility scenarios where rapid DL/UL switching coordination would be inconvenient;
• lower-frequency cellular layers used for coverage and dependable signalling;
• systems that value continuous uplink availability.

FDD's two carriers experience different radio frequencies. Their instantaneous channel responses are therefore not identical. Channel information learned from uplink is not automatically the exact downlink channel, although geometry and long-term properties may still be related. Downlink beamforming commonly needs reference signals, feedback, prediction, or other channel-acquisition methods.

TDD: Separate the Directions in Time

Time Division Duplex (TDD) uses the same carrier frequency for both directions but assigns different symbols, slots, or time intervals to downlink and uplink.

For a simplified repeating cycle,

Where:

• $T_{\mathrm{cycle}}$ is the duration of the simplified repeating TDD pattern;
• $T_{\mathrm{DL}}$ is the time assigned to downlink;
• $T_{\mathrm{UL}}$ is the time assigned to uplink;
• $T_{\mathrm{guard}}$ is the flexible or guard time used around a direction transition.

and the scheduled time fractions are approximately

Where:

• $\eta_{\mathrm{DL}}$ is the fraction of the cycle scheduled for downlink;
• $\eta_{\mathrm{UL}}$ is the fraction of the cycle scheduled for uplink;
• the remaining fraction includes guard/flexible time and any other excluded interval in this simplified model.

The guard interval allows propagation and RF switching before the direction changes. Real NR configurations operate at symbol and slot level and may include downlink, uplink, and flexible symbols.

How TDD Works

1. The base station and UEs share one unpaired carrier.
2. During a DL interval, the base station transmits and scheduled UEs receive.
3. A guard or flexible interval supports switching and timing uncertainty.
4. During an UL interval, selected UEs transmit and the base station receives.
5. The DL:UL pattern may be semi-static or adapted under supported network procedures.

Where TDD Is Useful

• unpaired mid-band or high-band spectrum;
• traffic that is strongly asymmetric and often downlink-heavy;
• dense urban capacity layers and fixed wireless access;
• massive-MIMO systems that can exploit calibrated UL/DL channel reciprocity;
• private or local networks that need a task-specific DL:UL ratio.

Because UL and DL use the same carrier, a TDD base station can estimate the uplink channel and use reciprocity for downlink beamforming after correcting the different transmit and receive RF-chain responses. Propagation reciprocity does not remove the need for hardware calibration.

FDD and TDD Compared

Neither method is universally better. Deployment spectrum, cell size, traffic asymmetry, device capability, mobility, synchronization, interference, and antenna architecture determine the useful choice.

What Does One FDD or TDD Frame Look Like?

In NR, a radio frame lasts $10\ \mathrm{ms}$ and contains ten $1\ \mathrm{ms}$ subframes. A subframe contains

Where:

• $N_{\mathrm{slot/subframe}}$ is the number of slots in one $1\ \mathrm{ms}$ subframe;
• $\mu$ is the NR numerology index;
• $2^\mu$ shows that doubling subcarrier spacing doubles the number of slots per subframe.

Therefore, the number of visible slots changes with subcarrier spacing even though the radio-frame duration remains $10\ \mathrm{ms}$. Each normal-cyclic-prefix slot contains 14 OFDM symbols.

Reading the FDD Frame

The FDD panel shows two aligned $10\ \mathrm{ms}$ radio frames:

• the upper frame exists on the downlink carrier, so the base station transmits toward scheduled UEs;
• the lower frame exists on the paired uplink carrier, so scheduled UEs transmit toward the base station;
• the two rows represent different frequencies at the same time, not alternating directions on one frequency;
• inside each time interval, the DL and UL schedulers independently assign PRBs, users, beams, MCS, control, reference signals, and HARQ resources.

Consequently, an FDD frame does not normally contain a pattern such as DL–DL–UL on one carrier. A DL time resource can be idle when no downlink transmission is scheduled, but that does not turn it into an UL resource. UL remains on the paired UL carrier.

Reading the TDD Frame

The TDD panel shows one $10\ \mathrm{ms}$ frame on a single carrier frequency. Its slots or symbols can be classified as:

• DL: the base station transmits and UEs receive;
• UL: UEs transmit and the base station receives;
• flexible or mixed: symbol directions are configured or indicated according to the supported procedure; part of a slot may carry DL symbols, a guard/flexible region, and UL symbols.

The illustrated sequence—DL, DL, DL, mixed, UL, UL, DL, DL, mixed, UL—is only a teaching pattern. Real deployments select supported periodicities and symbol allocations based on traffic ratio, numerology, synchronization, coverage, switching time, and interference coordination. Several adjacent UL slots are valid; several adjacent DL slots are also valid. What is prohibited in classical half-duplex TDD is treating the same time–frequency resource as ordinary DL and ordinary UL simultaneously.

A Numerical TDD Example

Suppose a simplified $5\ \mathrm{ms}$ TDD cycle contains

• $3.5\ \mathrm{ms}$ for downlink;
• $0.25\ \mathrm{ms}$ for guard/flexible operation;
• $1.25\ \mathrm{ms}$ for uplink.

Then

Where:

• $\eta_{\mathrm{DL}}=0.70$ means $70\%$ of the example cycle is assigned to downlink;
• $\eta_{\mathrm{UL}}=0.25$ means $25\%$ is assigned to uplink;
• the denominator $5$ is the full cycle duration in milliseconds;
• the remaining $5\%$ is the stated guard/flexible interval.

The remaining $5\%$ is switching/guard overhead in this teaching example. These fractions are time shares, not guaranteed user throughput. Control channels, pilots, coding, retransmissions, user competition, and radio quality further reduce delivered data.

If an application suddenly generates large uplink camera traffic, a supported TDD network may allocate more uplink opportunities. The change must be coordinated: one cell transmitting DL while a neighbouring cell expects weak UL signals can create severe base-station-to-base-station cross-link interference.

Can FDD and TDD Be Used Together?

Yes, in several distinct ways.

What Changes in Real Time?

The word switching can refer to three different actions, and separating them avoids a common misunderstanding.

1. FDD scheduling changes: FDD does not normally flip a carrier from DL to UL. The paired DL and UL carriers remain separate, while their schedulers independently change the UE, resource blocks, beam, modulation/coding, and retransmission assignment from one scheduling interval to the next. Either direction can be temporarily idle.
2. TDD direction changes: the same carrier is permitted to carry DL, UL, or flexible symbols according to a configured or indicated pattern. A direction transition needs an appropriate guard/flexible opportunity. The scheduler can only place a DL transmission in a DL-capable interval and an UL transmission in an UL-capable interval.
3. Carrier or cell changes: a network may add, release, or move traffic between an FDD coverage carrier and a TDD capacity carrier using supported carrier aggregation, dual-connectivity, mobility, and bearer-control procedures. This is a network decision, not merely recolouring the next radio slot.

A TDD pattern is therefore not changed independently for every arriving packet. The network considers supported configurations, UE signalling, neighbouring-cell alignment, cross-link interference, guard requirements, traffic measured over time, and implementation limits. Fast scheduling happens inside the currently usable opportunities; a larger pattern reconfiguration is a coordinated control action.

Example: Interactive Video While Walking Out of a Hotspot

Consider a user on a video call while downloading shared presentation material.

1. A low-band FDD cell carries control traffic, voice/video packets, and mobility signalling with simultaneous UL and DL availability.
2. Inside a busy-building hotspot, the network adds a wider TDD carrier. Its DL opportunities carry the large presentation and remote-video bursts, while UL slots carry the local camera and acknowledgements.
3. When the user starts sharing a high-resolution camera, the scheduler fills more existing UL opportunities. If sustained UL load justifies it and the deployment supports adaptation, the network can coordinate an UL-richer TDD pattern.
4. As the user leaves TDD coverage, the capacity carrier is released and the session continues through the FDD layer, subject to available capacity and mobility procedures.

At every stage, the radio scheduler still selects the actual user and resources. Duplexing states where a direction is allowed; scheduling states who receives the allowed resource.

1. Different Cells or Coverage Layers

An operator can use a low-band FDD layer for broad coverage and mobility, plus a wider mid-band TDD layer for local capacity. A UE may camp on or maintain control through the coverage layer and receive extra data capacity from the TDD layer when coverage and device support allow.

Real-life scenario: a commuter moves from a rural road into a busy city centre. The FDD layer maintains robust wide-area service. Near the city site, the network adds a TDD capacity carrier for fast downloads and dense-user scheduling.

2. Carrier Aggregation or Multi-Connectivity

A capable UE can be configured with component carriers or serving cells that use different duplex modes. This does not merge them into one carrier; the network coordinates multiple carriers.

Important conditions include:

• the UE and network support the specific FDD–TDD band combination;
• the spectrum licences and regional band plans permit the deployment;
• timing, control signalling, and scheduler decisions are coordinated;
• the UE RF front end can maintain adequate isolation between simultaneous transmit and receive operations;
• power, thermal limits, and maximum uplink capability are respected;
• inter-band and cross-link interference are controlled.

3. Dual Connectivity Across Different Nodes

A UE may receive service from two coordinated nodes or cell groups, for example an FDD macro layer and a TDD small cell. This can improve coverage robustness and capacity, but mobility, bearer mapping, latency, backhaul quality, and UE capability matter.

4. Advanced Time–Frequency Duplexing

Classical TDD does not transmit DL and receive UL simultaneously on the same carrier. Subband full duplex (SBFD) divides a TDD carrier into non-overlapping frequency subbands so a base station can transmit DL in one subband while receiving UL in another during part of the same time interval.

This resembles a time–frequency combination, but it is not simply ordinary FDD placed inside TDD. It introduces demanding self-interference, adjacent-subband leakage, base-station-to-base-station interference, UE-to-UE interference, RF dynamic-range, scheduling, and legacy-coexistence problems. 3GPP studied NR duplex evolution including SBFD and flexible/dynamic TDD; practical support depends on the relevant release, band, network implementation, and device capability.

Real-World Decision Examples

Downlink-Heavy Stadium

Thousands of spectators stream video and download media. A wide TDD carrier can devote more intervals to DL. The operator still reserves UL capacity for acknowledgements, social uploads, calls, and control signalling. Neighbouring cells need aligned or carefully coordinated patterns.

Industrial Robot Cell

The factory sends control commands downlink and receives sensor/vision data uplink. A near-balanced or uplink-rich TDD pattern may be preferable to a consumer-style DL-heavy pattern. The cycle length, guard time, deterministic scheduling, interference, and failure recovery must fit the control deadline.

Rural Coverage and Emergency Calling

Paired low-band FDD offers a dependable coverage layer with simultaneous UL and DL. A higher-frequency TDD layer can be added near towns or fixed-wireless sites for capacity, while the FDD layer continues to provide reach and mobility continuity.

Cloud Gaming or Interactive XR

The rendered scene may be DL-heavy, but head pose, controller state, and acknowledgements require timely UL. A TDD configuration must provide frequent enough UL opportunities; a long DL run can increase waiting time for an urgent UL update even if average DL throughput is excellent.

What Determines the Best Choice?

1. Spectrum: paired spectrum naturally supports FDD; unpaired spectrum supports TDD.
2. Traffic ratio: estimate UL and DL load over time, including control and retransmissions.
3. Latency: TDD direction switching creates a waiting opportunity that depends on the pattern.
4. Cell size: larger propagation delay can require more switching margin.
5. Synchronization: neighbouring TDD cells need coordination to limit cross-link interference.
6. Massive MIMO: TDD reciprocity can reduce explicit downlink-channel feedback when calibrated.
7. Device RF capability: simultaneous multi-band operation may be limited by filters, power, heat, and supported combinations.
8. Mobility and reliability: measure handover, outage, UL coverage, and application-level performance rather than only peak DL rate.

Takeaway

FDD separates uplink and downlink in frequency and can carry both directions simultaneously. TDD separates them in time on the same carrier and can adapt the direction ratio more flexibly. Modern networks can use both through layered deployments, carrier aggregation, or multi-connectivity when spectrum, UE capability, scheduling, timing, and interference conditions are satisfied.