MD-UV380 bring-up checklist

The MD-UV380 port (make BOARD=mduv380, and POWER=10w for an MD-UV390 Plus 10 W) builds and passes its host tests. Nothing has run on a radio yet. This page is the order to try it in, what each stage proves, and what to change when it doesn't work. The stage numbers match the unknowns table in hardware-mduv380.md section 14. Every row names the knob in the source that the likely failure points at.

The radio is a TYT MD-UV380, MD-UV390 or Retevis RT3S, with or without GPS. The port is the UV380 board family's code (boards/uv380_family, shared with the DM-1701) with this radio's data in boards/mduv380/board.c (uvf_variant, uvf_key_matrix). Most stages check one entry of that data.

The bench is remote: nobody presses keys or looks at the screen. Every check below reads something over the debug link (tools/rfoundry_debug.py) or on the SDR bench, and the diagnostics page lines (diag) carry the raw readings.

The feature checklist, with every feature's required HIL tests and the last recorded result, is tools/bringup.py list --radio mduv380. tools/bringup.py run --radio mduv380 records a run (tests/hil/README.md). This page keeps the hardware steps and what to change when one fails.

Status

Not run on a radio. The codec builds were checked on a host only (untested on the radio): TYT 19.26; TYT 19.74 and 219.74, 5 W and 10 W; TYT and Retevis 18.11 and 18.16; Retevis RT3S 19.06 and 219.60. A first DMR test should use a 19.26 file and one whose code goes to the start of sector 10 (RT3S 219.60, any 219.74, S19.74).

0. Before anything

Back up the radio: the codeplug with the stock CPS, then the whole SPI flash and the internal flash through the stub (rfoundry backup). The factory calibration is in the SPI flash security registers; RadioFoundry only reads it.
Build: make BOARD=mduv380 (5 W) or make BOARD=mduv380 POWER=10w (Plus 10 W). With a supported TYT or Retevis file in codecs/ (codec-mduv380.md, "Stock files"), the build also writes _codec.bin with the codec merged in.
Flash with rfoundry flash through the TYT bootloader (PTT and SK1 held at power on) or the web flasher.

1. It boots

Check	Pass	If not
USB enumerates as `RadioFoundry MD-UV380`	the debug link answers `info`	HSE mode (unknown 1): `diag` line "hse": `ext` or `xtal` says which started; `HSI` means neither (`boards/uv380_family/system_uvf.c`)
Boot reason, crash log empty	`rfoundry_debug.py crash` empty	`tools/crash_decode.py` on the report

2. Display

Check	Pass	If not
Panel type	`diag` "panel type" 0-3	unknown 2: the factory byte (security register 3, 0x1D) maps through `uvf_variant.panel_type` / `panel_rgb`
Picture upright, not mirrored, colors right	the screen capture (`rfoundry_debug.py screen`) matches the gallery	the MADCTL of the type (`drivers/display/st7735`), `madctl_xor` in `display_uvf.c`
All 128 rows	no blank or wrapped band at the bottom	`UVF_LCD_HIDDEN_ROWS` (board.mk)
Backlight dims	the brightness setting changes the current draw	PD8 (5 W) / PC6 (10 W): `pins.h`

3. Keys, channel knob, power, flash

Check	Pass	If not
Each key reports the right name	the debug link's key view, with each key pressed by the HIL key test or a helper at the radio	unknown 4: `uvf_key_matrix` in `boards/mduv380/board.c`; `diag` shows raw rows
Encoder: clockwise is the next channel	`rfoundry_debug.py` encoder counters	unknown 5: the decoder's rest state (`input_uvf.c`)
Battery voltage within 0.1 V of a meter	`diag` "bat raw"	unknown 6: `uvf_variant.bat_mul` / `bat_div`
Volume 0 at the stop, 100 at the top	`diag` "vol raw" at both ends	unknown 7: `vol_min`, `vol_break`, `vol_max`
SPI flash JEDEC 0xEF4018	`info`	`flash_uvf.c` (`device`)
Calibration read	`diag` "cal" factory or working	the k factors `uvf_variant.cal_k`, the reader `calibration_uvf.c`
10 W marker	read security register 3 byte 0x1C on a 5 W and on a Plus 10 W radio (`rfoundry backup` keeps the security registers); `diag` "TX off: 10 W radio, 5 W image" means a 5 W image found the marker	unknown 3: the TYT images never read the byte, so the marker is unconfirmed; a 10 W image doesn't require it. `MDUV380_DEFS=-DMDUV380_10W_CHECK=0` skips it

4. AT1846S, HR_C6000

Check	Pass	If not
Both chips answer	`diag` "C6000 4:10=80", "errs 0"	`radio_uvf.c` bus set-up (the same pins as the DM-1701)

5. FM receive

tests/hil FM receive cases on 461.4625 against the SDR bench (--fm-mhz).

6. FM transmit (dummy load)

The lowest power first. tests/hil FM transmit cases with --allow-rf.

Check	Pass	If not
Carrier on frequency, deviation	the bench's FM check	the TX offset in `radio_uvf.c`, the calibration's I/Q rows
Power per level	a power meter at each level, both band edges	unknown 10: `uvf_variant.frac`, `cal_k`

7. DMR receive

The codec needs the merged image (_codec.bin). diag DMR counters, tests/hil DMR receive cases from the bench HAT.

8. DMR transmit (dummy load, then a hotspot or repeater)

Check	Pass	If not
The bench decodes our bursts and voice	`tests/hil -m radio_tx` DMR cases	unknown 9: AT1846S 0x44 AF gain; unknown 12: the codec descriptor (`vocoder_mduv380.c` profile)

9. GPS

G models only. diag and the APRS position: the module is powered on PA9 for 5 s at start-up, kept on when NMEA is heard on USART1 (PA10, 9600 baud), switched off otherwise (unknown 8, boards/mduv380/board.c).

10. After the bench

Record each result in hardware-mduv380.md (unknowns table) and in this page's Status, and move the radio's [status] in boards/mduv380/board.toml on.

View the source of this page