IWR6843 Troubleshooting¶
Start with the symptom shown in the terminal. Avoid changing estimator settings until power, ports, firmware, config, and geometry are verified.
| Symptom | Likely cause | Action |
|---|---|---|
no IWR6843 CLI found |
Wrong USB interface, board still in flash mode, missing functional RESET, stale serial owner, or unstable power | Set functional switches, press RESET, verify interface 00, stop serial processes, then retry with explicit --iwr6843-port |
GPIO busy |
Another kiosk, calibration, or shot-test process owns BCM17 | Stop the old process; use pgrep -af and sudo fuser -v /dev/gpiochip* to locate it |
captureFormat or phaseCaptureCfg rejected |
Older firmware is flashed | Flash the configurable release, reset in functional mode, and retry either supported profile |
| Bootloader probe returns no response | Wrong CP2105 port or RESET occurred before the script opened UART | Use Enhanced/UARTA, rerun the probe, type READY, then RESET only when prompted |
Flash fails after Erasing existing SFLASH |
Transfer was interrupted after the old image was erased | Leave the board in flash mode and rerun the complete flash; the ROM bootloader is still available |
| Server starts only after unplugging TI | Board was not reset cleanly, a prior dump was still streaming, or USB/power wedged | Stop the old process, press RESET in functional mode, wait for the port, then reconnect USB only if needed |
short IWR6843 dump |
Interrupted UART transfer, process shutdown during dump, or wrong firmware format | Let the active dump finish, restart, and confirm 732,812 bytes for wide or 549,764 bytes for dense |
Dense stats accumulates hwa_missed or iq8_overrun |
The 2 ms processing budget is not being sustained | Stop using the capture for measurements, reset, and return to the wide profile while investigating |
Clap produces rejected_by_ball_tracker |
A clap has no moving ball range track | Expected for trigger testing; confirm the dump completed, then hit a ball |
rejected_track_quality |
A ball-like track was found but it was too thin, noisy, inconsistent, or net-contaminated | Verify geometry and aim; inspect the debug dump before relaxing acceptance gates |
rejected_missing_tdm_sign |
The ball track was usable, but the TX timing evidence did not resolve a trustworthy correction sign | Keep the estimated UI angle, inspect the debug dump, and verify signal quality before changing gates |
| All UI angles are estimated | TI captures are absent, unmatched to OPS, or rejected by LCMF | Run with --debug, inspect iwr6843_capture, and check the reported rejection reason |
| OPS reports no data | Wrong OPS port, missing power, WiFi OPS connected through unsupported receive-only J3 UART, or non-WiFi UART wired incorrectly | For a WiFi OPS use the externally powered USB hub; otherwise verify /dev/ttyAMA0, power, shared ground, and crossed TX/RX |
| Either radar disconnects when both run | Insufficient USB power or unstable cabling | Use OPS GPIO power or a hub with its own external supply; verify the hub supply is connected and sized for both radars |
| Angles are consistently shifted | Tilt, antenna orientation, radar height, ball height, or tee distance is wrong | Re-measure all geometry from the antenna center and common floor reference |
| Dump file is missing from the session | OpenFlight was not launched with --debug |
Re-run in debug mode when raw capture retention is required |
If the firmware itself must be rebuilt rather than flashed from the checked-in binary, continue with the firmware developer guide.