Troubleshooting

SDR timing, clock, and gain problems

Fix the SDR-layer problems that produce inconclusive verdicts: sample-rate mismatch, clock drift, gain saturation, USB throughput, and calibration.

7 min read

SDR problems masquerade as protocol problems. Learn the four signatures and you stop chasing NAS bugs that do not exist.

SignatureMeaningFix
'U' / 'O' spam in logsUnderrun / overflow — host cannot keep upLower sample rate, use a hosted lane build
'L' / 'D' charactersLate packets / dropped burstsTiming problem; re-check master_clock_rate
Corrupt SIB decode at high RSRPReceiver saturationReduce rf.rx_gain by 10-15 dB
Slow drift then loss of syncClock offset (no GPSDO)Request a GPSDO-equipped lane

Correct clock rates

texttext
B210, 20 MHz NR, 30 kHz SCS : master_clock_rate=23.04e6
B210, 10 MHz LTE            : master_clock_rate=11.52e6
LimeSDR USB, 20 MHz NR      : sample_rate=30.72e6, oversample=2

Lane calibration

bashbash
aeon lanes describe --id eu-west-1/lime-03
# sdr: LimeSDR-USB  fw: 22.10  gpsdo: yes  calibrated: 2026-07-13T22:04Z

aeon lanes calibrate --id eu-west-1/lime-03   # ~90 s, free

noteLanes auto-calibrate every 24 h and after any firmware change. Every report records the calibration timestamp, which is what makes a cloud verdict defensible.

Run this against a real SDR lane

The tester is a service, not a box. Push a build, reserve a lane, get a verdict.