LTE-M vs NB-IoT: which should you choose?

Peak figures are standards capabilities, not expected application throughput. Signal, repetitions, network configuration, protocol and module category affect observed results.

Power Saving Mode (PSM) lets a registered device become unreachable while it sleeps, then wake on its own schedule for an update. Extended Discontinuous Reception (eDRX) lets a device periodically listen for paging over longer intervals than conventional LTE. PSM usually gives deeper sleep; eDRX provides bounded windows for downlink.

Timers are requested by the device but granted by the network, and implementation differs. Battery estimates must include attach events, Tracking Area Updates, poor-signal repetitions, application retries, temperature and self-discharge. Measure current with production firmware and a realistic network rather than multiplying a module data-sheet sleep current by years.

Both technologies include coverage-enhancement mechanisms using repeated transmissions. NB-IoT's narrow channel and repetition design can support challenging link budgets, but “deep indoor” is not a guarantee for a particular basement or enclosure. LTE-M can also use coverage enhancement.

Coverage maps may not distinguish every band or feature. A network can advertise LTE while not enabling LTE-M or NB-IoT in the relevant area, and roaming support may be narrower than domestic access. Confirm the exact operator, band, module certification and roaming arrangement, then test the final antenna inside the enclosure.

LTE-M is generally the appropriate choice for moving trackers because it supports connected-mode handover. NB-IoT devices can reselect cells, but session continuity and behavior at speed are less suitable for continuous mobility. A stationary asset that moves only occasionally might still use NB-IoT if interruptions are acceptable.

LTE-M can support Voice over LTE (VoLTE), including use cases such as alarms, but support must exist across module firmware, network, subscription and certification. Do not infer voice from LTE-M data access. Test call setup, emergency requirements, audio path and fallback behavior with each target operator.

NB-IoT favors compact, infrequent, delay-tolerant telemetry. Large firmware downloads can remain possible but consume time and energy, especially in weak coverage. LTE-M's throughput makes updates and richer exchanges easier. Neither removes the need to optimize protocol overhead.

Use store-and-forward queues, message identifiers and duplicate handling. Avoid aggressive retries when a network is unavailable. Decide how long commands may wait while a device uses PSM and whether the application can tolerate changed IP sessions. Calculate normal, update and fault traffic separately with the data usage calculator.

Eliminate options unavailable across required operators and countries. Next score mobility, downlink reachability, payload, update size, latency, voice and power. Select candidate modules with correct bands and regulatory/operator approvals. Prototype both technologies if requirements sit near the boundary.

Cat-1 bis is a useful comparator where LTE-M and NB-IoT footprints do not align internationally. It uses standard LTE coverage and higher power for moderate throughput. The recommendation is to choose for fleet-wide lifecycle risk, not the lowest current in one laboratory transaction.

Build a test matrix for the actual countries, operators, subscription and roaming arrangements, supported bands, antennas, module revisions and production firmware. Confirm LTE-M or NB-IoT access and required features with each provider; ordinary LTE coverage or a module data sheet is not proof of usable LPWA service.

Measure attach success, time to first data, round-trip latency, delivery failures and energy per delivered message at representative installation sites, including indoor and weak-signal locations. Use production payload sizes, reporting intervals, protocol overhead and firmware-update downloads. Test stationary operation and the intended movement pattern, including cell changes and loss of service, against written delivery and latency thresholds.

For each operator, record requested and granted PSM and eDRX timers, measured sleep current, wake-up behavior and downlink reachability. Include repetitions, retries, reattachment and outages in the battery budget. Power-saving features in a standard do not guarantee that a network grants the desired settings or that a particular firmware implementation achieves the expected battery life.

Compare measured results rather than peak rates or a single laboratory current trace. Retain the module firmware, network configuration and test conditions with the acceptance evidence so a later hardware, firmware or operator change can trigger focused retesting.

Link each deployed device to its radio technology, supported bands, module and firmware revision, subscription, operator and installation location. Assign owners for delivery-rate, latency and power-budget monitoring. Set silence thresholds around the reporting schedule and negotiated sleep behavior so an intentionally sleeping meter is not treated like a failed alarm.

Monitor repeated registration attempts, retransmissions, delayed commands and unexpected data or battery consumption. Keep enough device-side evidence to distinguish weak coverage, changed sleep settings, roaming access failures and application outages. Use bounded retries and store-and-forward behavior, and test recovery after extended loss of coverage.

Stage firmware updates and recheck attach, mobility, timers and energy use before fleet-wide rollout. Track target operators' LPWA availability and lifecycle plans; another radio technology is a recovery option only if the hardware, bands, subscription and approvals support it. Document when remote recovery is possible and when a site visit or device replacement is required. Use the IoT deployment checklist to record launch and ongoing responsibilities.