Dead speakers due to SWJ release failure

[Chase-san]

For whatever reason the STM never disables SWJ debug interface during running, preventing PA_EN from going high and preventing the speakers from working.

Symptoms

• Both speakers completely silent from first boot.
• Headphone jack works normally.
• Volume wheel has no effect on speakers.

Hardware / Firmware

• PicoCalc
• Pico2W
• Picocalc BIOS v1.4, v1.6 (likely older as well)

Test Progression

1. Clean PicoMite install produces audio at the jack. So PWM (GP26/27), filter, and volume pot are good.
2. HP_DET at the R507/R508 node: ~0 V unplugged, ~3.2 V plugged, so the jack detect switch works.
3. Speakers verified working when driven directly from an external source.
4. AW8010 output pads read 0 V (high-Z). So that means the amps never enable.
5. STM32 pin 49 (PA14 / PA_EN) stays low even with firmware modified to write it HIGH every 10 ms. Reflowing pins 49–53 changed nothing.
6. I noticed while testing that pin 50 (PA15) sits solidly high with nothing driving it. PA15 is JTDI, which has an internal pull-up only while the JTAG/SWD matrix owns the pin. PA14 is SWCLK (internal pull-down). PA13 is SWDIO (internal pull-up). Which is why PICO_EN is high and the Pico boots anyway.
7. Bridging pin 49 to pin 50 (JTDI pull-up vs SWCLK pull-down) raises PA_EN above the AW8010 EN threshold causing the speakers to work.
8. Firmware patched to explicitly write AFIO->MAPR SWJ_CFG = 100 (JTAG+SWD disabled) and reconfigure PA14 which lets the speakers work with no hardware modification.

Cause and Fix

On the chip I have in my PicoCalc the implicit SWJ release performed by the Arduino core during pinMode(PA13/PA14, ...) does not take effect, so PA13/PA14/PA15 remain owned by the debug interface. PA14 (PA_EN) is therefore held low by SWCLK's internal pull-down and firmware GPIO writes never reach the pin. A direct/explicit SWJ_CFG write works however.

In picocalc_keyboard.ino, force the release explicitly:

// in setup(), before any PA13/PA14 pinMode:
__HAL_RCC_AFIO_CLK_ENABLE();
__HAL_AFIO_REMAP_SWJ_DISABLE();

// In check_hp_det(), re-assert each poll:
__HAL_AFIO_REMAP_SWJ_DISABLE();

[PasswordIs0000]

I think I had the same problem. My speakers didn't work out of the box. The BIOS 1.7 from yesterday fixed the problem for me and now my speakers are working.