What are the configurable values for bwp-InactivityTimer?

3GPP TS 38.331 defines the following timer values in milliseconds: 2, 3, 4, 5, 6, 8, 10, 20, 30, 40, 50, 60, 80, 100, 200, 300, 500, 750, 1000, 1200, 1600, and 2000 ms. The network selects the value based on expected traffic patterns. For bursty web browsing, 100-200 ms prevents excessive BWP toggling. For always-on IoT monitoring, 10-20 ms aggressively saves power between sparse uplink reports. Setting the timer too short causes frequent BWP switches that add latency; too long wastes power by keeping the wideband BWP active during idle gaps.

How does the BWP timer interact with DRX in 5G NR?

The BWP inactivity timer and DRX (Discontinuous Reception) inactivity timer operate independently but their interaction determines UE power state transitions. If bwp-InactivityTimer is shorter than drx-InactivityTimer, the UE first narrows its BWP (reducing ADC power) while still monitoring PDCCH, then enters DRX sleep when the DRX timer expires (powering down the RF chain entirely). If the BWP timer is longer, DRX sleep occurs before BWP fallback, and the UE wakes in the wide BWP. Optimal configuration sets the BWP timer shorter than DRX to cascade power savings: first reduce bandwidth, then sleep.

Does BWP timer fallback cause latency for incoming data?

Yes, but the impact is manageable. When the UE falls back to the default narrow BWP, incoming data triggers a DCI-based BWP switch back to the wider BWP, adding 1-3 slot delays (0.5-3 ms depending on SCS). For latency-sensitive applications like VoNR, the network can either set a longer BWP timer to avoid fallback during active calls, or configure the default BWP wide enough to carry voice traffic directly. The 1-3 ms switching penalty is typically acceptable for enhanced mobile broadband (eMBB) traffic but may be configured out for URLLC services requiring sub-1 ms latency.

What is the BWP Inactivity Timer? | 5G NR Power Saving Timer

Definition & Function

The BWP Inactivity Timer (bwp-InactivityTimer) is a UE-side timer defined in 3GPP TS 38.321 that governs automatic fallback from a wideband active Bandwidth Part to a narrower default BWP. The timer starts (or restarts) each time the UE successfully decodes a PDCCH addressed to its C-RNTI on the active BWP. When the timer expires without any new scheduling grant, the UE autonomously switches to the configured defaultDownlinkBWP-Id, which is typically the narrowest BWP configured for the cell.

This mechanism provides the primary power-saving benefit of BWP switching without requiring explicit network signaling for every transition. By narrowing the active bandwidth, the UE reduces its ADC/DAC sampling rate, narrows the analog front-end filter bandwidth, and decreases baseband processing load. Measurements from commercial 5G modems show 30-60% reduction in modem power consumption when falling back from a 100 MHz BWP to a 20 MHz default BWP. The timer value is configured per serving cell via RRC signaling and can be updated dynamically through RRC reconfiguration messages.

Key Formulas

Timer Duration (slots):

N_slots = T_timer / T_slot

100 ms timer at 30 kHz SCS: N_slots = 100 / 0.5 = 200 slots

Power Savings Estimate:

P_save = (T_idle / T_total) × (P_wide − P_narrow)

80% idle time, 500 mW wide vs 200 mW narrow: P_save = 0.8 × 300 = 240 mW average

Optimal Timer (latency-power trade-off):

T_opt ≈ T_inter-burst × 0.3

For 500 ms burst intervals: T_opt ≈ 150 ms

5G NR Timer Interaction

Timer	Trigger	Action on Expiry	Power Impact	Typical Value
bwp-InactivityTimer	No PDCCH grant	Fall back to default BWP	30-60% modem savings	100-200 ms
drx-InactivityTimer	No PDCCH grant	Enter DRX sleep	70-90% RF chain savings	10-100 ms
drx-onDurationTimer	DRX cycle start	Monitor PDCCH window	Brief wake period	1-10 ms
sr-ProhibitTimer	SR sent	Allow next SR	Minimal	1-128 ms
t310 (RLF)	N310 OOS events	Declare radio link failure	Recovery overhead	1000-2000 ms

Practical Application

A 5G NR operator configures a dense urban small cell on n41 (2.5 GHz, 100 MHz carrier) with bwp-InactivityTimer = 100 ms, defaultDownlinkBWP-Id = 0 (20 MHz), and an active BWP 1 (100 MHz) for high-throughput scheduling. A user browsing social media generates bursty traffic with 200-500 ms gaps between page loads. After each content burst, the 100 ms timer expires and the UE falls back to BWP 0, reducing modem power from 480 mW to 190 mW. When the next page load triggers a scheduling grant, the gNB sends a DCI activating BWP 1, and the UE switches back within 1 ms. Over a typical browsing session (70% idle), this saves approximately 200 mW average power, extending battery life by 15-20% compared to always-on 100 MHz operation.

Frequently Asked Questions

What values can bwp-InactivityTimer take?

Per TS 38.331: 2, 3, 4, 5, 6, 8, 10, 20, 30, 40, 50, 60, 80, 100, 200, 300, 500, 750, 1000, 1200, 1600, and 2000 ms. Too short causes excessive switching overhead; too long wastes power keeping the wide BWP active during gaps.

How does BWP timer interact with DRX?

They operate independently. Optimal setup: BWP timer shorter than DRX timer so the UE first narrows bandwidth (saves ADC power), then enters DRX sleep (powers down RF chain). This cascades power savings from 30% (BWP narrow) to 80%+ (DRX sleep).

Does timer fallback add latency?

Yes, 1-3 slots (0.5-3 ms) for BWP re-activation via DCI. Acceptable for eMBB but may be disabled for URLLC. For VoNR, operators set longer timers or configure the default BWP wide enough to carry voice directly.

BWP Inactivity Timer