Phoenix — Replacement Board Restored and Recalibrated

The replacement ordered after the unresolved temperature warning arrived and was fitted. It reports the same BETAFPVG473_V2 target, already on the correct BETAFPV dev firmware (e92c10887, matching the pre-replace backup) with no reflash needed. The gyro on this unit is BMI270, one of the substitutes the board ships with when ICM42688 is scarce — see the gyro firmware notes.

Restore

Config restored from backup_2026-09-17_195000_crafty_prereplace.txt, taken the day the old board was pulled. Verified after the reboot:

Setting	Value
`craft_name`	`Crafty`
`vcd_video_system`	`AUTO` — correct for this craft
`osd_displayport_device`	`MAX7456`
Crashflip	AUX5, mode 35, `crashflip_rate = 30`, `crashflip_auto_rearm = OFF`
PID profile 0	`GF 1614`, unchanged
PID profile 2	the snappier Horizon settings

The carried-over calibration was wrong

acc_calibration restored at -63,17,-35,1 — the old board’s offsets, replayed onto a different physical chip. That value is a property of one sensor’s mounting and manufacturing tolerance, not of the airframe, so it does not transfer across a board swap even though everything else in the backup does.

Recalibrated level and still, over MSP (MSP_ACC_CALIBRATION, code 205):

	Old board	New board
`acc_calibration`	`-63,17,-35,1`	`-4,-44,-15,1`
Attitude at rest	—	roll 0.0°, pitch 0.0°

A config backup restores everything except the accelerometer calibration correctly after a board swap. Recalibrate on a level surface before the first flight on new hardware, even when restoring an otherwise-known-good config.

Bench temperature

A first reading, 37 °C on USB with a fan on the board, was not a usable baseline — that is exactly the confound that sent the old board away in the first place. A no-fan reading followed it, tracked over time rather than as a single number, since the previous investigation found no-fan readings keep climbing for minutes before flattening:

Time	Core temperature
30 s	42 °C
60 s	49 °C
90 s	54 °C
120 s	57 °C
150 s	63 °C
180 s	64 °C
210 s	65 °C

Levelling off around 63–65 °C at 3.5 minutes, USB only, no battery. That lands close to the Air65 control reading from the same investigation (36 → 60 °C at 3 minutes, still rising), and well under the old board’s 70–72 °C plateau. This is the real baseline for the new board.

Blade orientation

Same as the Air65: default QUADX mixer, no custom mmix, yaw_motors_reversed = ON on both craft. The restored config reproduces that, so prop rotation direction did not need to be touched on the new board.

Status

Motors confirmed connected. Config, blade orientation, calibration and a clean thermal baseline are all in place — the board is ready to fly.

Still open: watch for the propwash-era level drift noted in the 13 September review on the first hover pack, and confirm the core temperature stays near this baseline in flight rather than repeating the 09-17 warning.