Waking up on right button - doesn't happen.

[Sarah-C]

I see that the right button puts the device to sleep - which works fine.
However I note that esp_sleep_enable_ext0_wakeup is used to set the wake button to the same right button.
This doesn't work either via a build in Arduino (latest regular build) nor in PlatformIO - the device stays off.

I've no idea where the error is, it doesn't look like it's how it's being used here - perhaps Espressifs own code?
I'm just raising this issue for others to know it's a "known issue" - and not this codes fault.

main.cpp:

void go_to_sleep(void) {
  Serial.println("sleep");
  digitalWrite(PIN_LCD_BL, HIGH);
  gpio_hold_en((gpio_num_t)PIN_LCD_BL);
  digitalWrite(PIN_LCD_RES, LOW);
  esp_sleep_enable_ext0_wakeup((gpio_num_t)PIN_BTN_R, 0);
  esp_light_sleep_start();
}

What pins can wake the device.

The pins labelled RTC compatible.

The left button is GPIO 0, and the right is GPIO 47, so I originally thought it wasn't waking due to GPIO47 not being a RTC pin (and thus unable to wake the device) as shown in the article above. I swapped the pin numbers in pin_config.h, because GPIO 0 IS connected to the RTC. But this failed to wake the device either.

#define PIN_BTN_L            47
#define PIN_BTN_R            0

[Gargantuanman]

Im having the same issu thats because for some reason the PULL UP of the pin 0 fall when you send the controller to deepsleep that shouldnt be happening i believe there is some error on the schematic because supposely the button isn pulled up directly to regulator out, so it should be 3.3v always.

[mmMicky]

As far as I know ESP32-S3 has this problem. You can use esp_sleep_enable_ext1_wakeup instead.

[afpineda]

Im having the same issu thats because for some reason the PULL UP of the pin 0 fall when you send the controller to deepsleep that shouldnt be happening i believe there is some error on the schematic because supposely the button isn pulled up directly to regulator out, so it should be 3.3v always.

Despite being externally pulled-up, maybe the GPIO is disabled before entering deep sleep mode. Check this code:

void enterDeepSleep()
{
  // disable radios
  esp_bluedroid_disable();
  esp_bt_controller_disable();

  // Disable pins to avoid current drainage through pull resistors
  // Enable proper pull resistors for wake up (if available)
  for (int p = 0; p < GPIO_NUM_MAX; p++)
    if (GPIO_IS_VALID_GPIO((gpio_num_t)p) && !((p >= 6) && (p <= 11))) // Exclude flash memory pins
    {
      uint64_t pinBitmap = BITMAP(p);
      if (pinBitmap & ext1_wakeup_sources)
      {
        if (ext1_wakeup_mode == ESP_EXT1_WAKEUP_ANY_HIGH)
        {
          gpio_pullup_dis((gpio_num_t)p);
          gpio_pulldown_en((gpio_num_t)p);
        }
        else
        {
          gpio_pulldown_dis((gpio_num_t)p);
          gpio_pullup_en((gpio_num_t)p);
        }
      }
      else
      {
        gpio_pulldown_dis((gpio_num_t)p);
        gpio_pullup_dis((gpio_num_t)p);
      }
    }

  // enter deep sleep
  if (ext1_wakeup_sources != 0ULL)
  {
    ESP_ERROR_CHECK(esp_sleep_pd_config(ESP_PD_DOMAIN_RTC_PERIPH, ESP_PD_OPTION_ON));
    ESP_ERROR_CHECK(esp_sleep_enable_ext1_wakeup(ext1_wakeup_sources, ext1_wakeup_mode));
  } // else reset is required for wake up
  esp_deep_sleep_start();

  // should not enter here
  log_e("power::powerOff(): Deep sleep not working");
  delay(5000);
  ESP.restart();
}

ext1_wakeup_sources and ext1_wakeup_mode are global variables. ext1_wakeup_sources is a bitmap of configured GPIO numbers to be used for wake up.

// Power management
static uint64_t ext1_wakeup_sources = 0;
static esp_sleep_ext1_wakeup_mode_t ext1_wakeup_mode = ESP_EXT1_WAKEUP_ANY_HIGH;

Source code

[lewisxhe]

I discovered this issue during yesterday's reorganization and has now fixed it

[Sarah-C]

lewisxhe - this was a PCB routing change?