How to Diagnose One-Way LoRa Links

A methodical way to find why packets travel in one direction but not the other.

Prove that the link is really one-way

First remove ambiguity. A node appearing in a list does not prove that current two-way application messages work. Send numbered packets A to B and B to A while logging at both endpoints. Confirm whether acknowledgements are requested and whether any delivery used relays. Repeat with the nodes close together. If both directions work locally, the fault is linked to the installed path or environment.

Check configuration symmetry

Compare region, frequency band, modem preset, custom modem values and frequency slot. Then compare messaging channel indexes and keys for the traffic being tested. A node may hear or relay packets it cannot decrypt, which can make “I can see it but cannot message it” look like an RF problem. Also inspect roles, rebroadcast settings, hop limits and whether one device is sleeping or power-saving during the test.

Compare the two RF systems

Part	Questions
Transmit path	Same configured power? Legal limit? Damaged output or connector?
Antenna	Correct band, polarization, pattern and installation?
Feed line	Different length, loss, adapters, water ingress or loose connector?
Receiver	Local noise, desensitization, board fault or enclosure coupling?
Power	Voltage collapse or reset during transmission?

Do not assume identical product labels mean identical systems. One site may have ten metres of lossy coax while the other has the radio beside the antenna.

Look for asymmetric noise and terrain

Propagation paths can be asymmetric in practice because the endpoint antennas, heights and local environments differ. One receiver may sit beside a noisy power supply or another transmitter. A directional or obstructed antenna pattern may favor one bearing. Buildings and hillside clutter near one endpoint can dominate the path.

Compare the failing receiver at a quiet temporary location or on battery power. Swap known-good antennas or radios one item at a time. If the failure follows a component, the component is suspect; if it remains at the site, investigate the site.

Account for the mesh

An apparent A-to-B success may actually travel A–relay–B while the reverse packet uses a different route, expires at its hop limit or collides under different traffic. Run the test when the path can be identified. Traceroute and packet metadata can help, but results themselves generate traffic and may vary as nodes contend to relay.

Use a disciplined isolation order

Confirm repeated directional loss.
Verify compatible radio and channel configuration.
Test both nodes at short range.
Inspect power, connectors, coax and antennas.
Substitute one known-good component at a time.
Measure or reduce local receiver noise.
Repeat at the installed locations with relay paths identified.
Document the cause and corrective action.

Avoid compensating immediately with more transmit power. It may be illegal, does not repair the weak receive direction and can increase interference.

Use packet identity, not app impressions

Write a small test matrix listing packet number, origin, destination, time, expected direction, receiver and path. Screens from different phones can display cached NodeInfo, relayed packets or old positions and create false conclusions. Logs from both radios, synchronized approximately in time, are stronger evidence.

If the path includes public relays, ask their operators before generating sustained diagnostic traffic. A quieter controlled test often reveals more than increasing the message rate.

Typical corrective actions

Corrective work may be as simple as replacing a mismatched adapter, drying a connector, reducing coax length or moving a switching supply away from the receiver. Other cases require raising one antenna, correcting polarization, changing a damaged radio or relocating the endpoint out of a terrain shadow. After the repair, repeat the original directional sequence under the same preset and placement. Do not close the incident merely because a different nearby test now works.

Related diagnostics

Official sources

Firmware, applications and documentation change over time. Confirm current setting names and applicable regional rules before commissioning or changing live infrastructure.