Receiver configuration persistence confirmed from the flash layer Read back with UBX-CFG-VALGET rather than inferred. Also records why power-cycling the board cannot test this: the firmware configures on every boot, and a power-on boot's log cannot be captured.
GNSS works: 20 Hz accepted fixes on hardware The receiver is configured at boot and delivering accepted fixes. The critical path has moved from hardware to software: nothing consumes the fix queue yet. Also records the two bugs bring-up found, including the UTC sub-second field that made measurement time run backwards and the test that asserted it.
The receiver is connected and talking: one configuration step from a fix The Qwiic cable arrived and the receiver is on the I2C socket. The device talks to it, the board was not damaged, and the receiver already holds a valid fix indoors. It emits NMEA where the firmware parses UBX, so nothing parses yet - that is #16, and it is now the only step between this device and a lap time. Also: the #9 socket hazard is retired but not answered, the boot log's stage lines are documented, and the UART0 open decision is replaced by the one that is actually open.
Write the Hardware page, and link it from Home Covers the parts and why those parts, the shared I2C bus and its deadline, the UART socket conflict between the console and the receiver, how to get diagnostics off the board, what is verified against what is assumed, and the order to bring the receiver up in.
Created Hardware (markdown)