Root cause seems to be the axp192 interrupt, which on some boards fires during sleep. I'm not sure why, but we don't need this yet, so leaving masked during sleep.
249 lines
8.2 KiB
C++
249 lines
8.2 KiB
C++
#include "configuration.h"
|
|
#include "rom/rtc.h"
|
|
#include <driver/rtc_io.h>
|
|
#include <TinyGPS++.h>
|
|
#include <Wire.h>
|
|
#include "BluetoothUtil.h"
|
|
#include "MeshBluetoothService.h"
|
|
#include "MeshService.h"
|
|
#include "GPS.h"
|
|
#include "screen.h"
|
|
#include "NodeDB.h"
|
|
#include "Periodic.h"
|
|
#include "esp32/pm.h"
|
|
#include "esp_pm.h"
|
|
#include "MeshRadio.h"
|
|
#include "main.h"
|
|
#include "sleep.h"
|
|
|
|
#ifdef T_BEAM_V10
|
|
#include "axp20x.h"
|
|
extern AXP20X_Class axp;
|
|
#endif
|
|
|
|
// deep sleep support
|
|
RTC_DATA_ATTR int bootCount = 0;
|
|
esp_sleep_source_t wakeCause; // the reason we booted this time
|
|
|
|
#define xstr(s) str(s)
|
|
#define str(s) #s
|
|
|
|
// -----------------------------------------------------------------------------
|
|
// Application
|
|
// -----------------------------------------------------------------------------
|
|
|
|
/**
|
|
* Control CPU core speed (80MHz vs 240MHz)
|
|
*
|
|
* We leave CPU at full speed during init, but once loop is called switch to low speed (for a 50% power savings)
|
|
*
|
|
*/
|
|
void setCPUFast(bool on)
|
|
{
|
|
setCpuFrequencyMhz(on ? 240 : 80);
|
|
}
|
|
|
|
void setLed(bool ledOn)
|
|
{
|
|
#ifdef LED_PIN
|
|
// toggle the led so we can get some rough sense of how often loop is pausing
|
|
digitalWrite(LED_PIN, ledOn);
|
|
#endif
|
|
|
|
#ifdef T_BEAM_V10
|
|
if (axp192_found)
|
|
{
|
|
// blink the axp led
|
|
axp.setChgLEDMode(ledOn ? AXP20X_LED_LOW_LEVEL : AXP20X_LED_OFF);
|
|
}
|
|
#endif
|
|
}
|
|
|
|
void setGPSPower(bool on)
|
|
{
|
|
DEBUG_MSG("Setting GPS power=%d\n", on);
|
|
|
|
#ifdef T_BEAM_V10
|
|
if (axp192_found)
|
|
axp.setPowerOutPut(AXP192_LDO3, on ? AXP202_ON : AXP202_OFF); // GPS main power
|
|
#endif
|
|
}
|
|
|
|
// Perform power on init that we do on each wake from deep sleep
|
|
void initDeepSleep()
|
|
{
|
|
bootCount++;
|
|
wakeCause = esp_sleep_get_wakeup_cause();
|
|
/*
|
|
Not using yet because we are using wake on all buttons being low
|
|
|
|
wakeButtons = esp_sleep_get_ext1_wakeup_status(); // If one of these buttons is set it was the reason we woke
|
|
if (wakeCause == ESP_SLEEP_WAKEUP_EXT1 && !wakeButtons) // we must have been using the 'all buttons rule for waking' to support busted boards, assume button one was pressed
|
|
wakeButtons = ((uint64_t)1) << buttons.gpios[0];
|
|
*/
|
|
|
|
// If we booted because our timer ran out or the user pressed reset, send those as fake events
|
|
const char *reason = "reset"; // our best guess
|
|
RESET_REASON hwReason = rtc_get_reset_reason(0);
|
|
|
|
if (hwReason == RTCWDT_BROWN_OUT_RESET)
|
|
reason = "brownout";
|
|
|
|
if (hwReason == TG0WDT_SYS_RESET)
|
|
reason = "taskWatchdog";
|
|
|
|
if (hwReason == TG1WDT_SYS_RESET)
|
|
reason = "intWatchdog";
|
|
|
|
if (wakeCause == ESP_SLEEP_WAKEUP_TIMER)
|
|
reason = "timeout";
|
|
|
|
DEBUG_MSG("booted, wake cause %d (boot count %d), reset_reason=%s\n", wakeCause, bootCount, reason);
|
|
}
|
|
|
|
void doDeepSleep(uint64_t msecToWake)
|
|
{
|
|
DEBUG_MSG("Entering deep sleep for %llu seconds\n", msecToWake / 1000);
|
|
|
|
// not using wifi yet, but once we are this is needed to shutoff the radio hw
|
|
// esp_wifi_stop();
|
|
|
|
BLEDevice::deinit(false); // We are required to shutdown bluetooth before deep or light sleep
|
|
|
|
screen.setOn(false); // datasheet says this will draw only 10ua
|
|
|
|
// Put radio in sleep mode (will still draw power but only 0.2uA)
|
|
service.radio.rf95.sleep();
|
|
|
|
nodeDB.saveToDisk();
|
|
|
|
#ifdef RESET_OLED
|
|
digitalWrite(RESET_OLED, 1); // put the display in reset before killing its power
|
|
#endif
|
|
|
|
#ifdef VEXT_ENABLE
|
|
digitalWrite(VEXT_ENABLE, 1); // turn off the display power
|
|
#endif
|
|
|
|
setLed(false);
|
|
|
|
#ifdef T_BEAM_V10
|
|
if (axp192_found)
|
|
{
|
|
// No need to turn this off if the power draw in sleep mode really is just 0.2uA and turning it off would
|
|
// leave floating input for the IRQ line
|
|
|
|
// If we want to leave the radio receving in would be 11.5mA current draw, but most of the time it is just waiting
|
|
// in its sequencer (true?) so the average power draw should be much lower even if we were listinging for packets
|
|
// all the time.
|
|
|
|
// axp.setPowerOutPut(AXP192_LDO2, AXP202_OFF); // LORA radio
|
|
|
|
setGPSPower(false);
|
|
}
|
|
#endif
|
|
|
|
/*
|
|
Some ESP32 IOs have internal pullups or pulldowns, which are enabled by default.
|
|
If an external circuit drives this pin in deep sleep mode, current consumption may
|
|
increase due to current flowing through these pullups and pulldowns.
|
|
|
|
To isolate a pin, preventing extra current draw, call rtc_gpio_isolate() function.
|
|
For example, on ESP32-WROVER module, GPIO12 is pulled up externally.
|
|
GPIO12 also has an internal pulldown in the ESP32 chip. This means that in deep sleep,
|
|
some current will flow through these external and internal resistors, increasing deep
|
|
sleep current above the minimal possible value.
|
|
|
|
Note: we don't isolate pins that are used for the LORA, LED, i2c, spi or the wake button
|
|
*/
|
|
static const uint8_t rtcGpios[] = {/* 0, */ 2,
|
|
/* 4, */
|
|
#ifndef USE_JTAG
|
|
12, 13, /* 14, */ /* 15, */
|
|
#endif
|
|
/* 25, */ 26, /* 27, */
|
|
32, 33, 34, 35, 36, 37, /* 38, */ 39};
|
|
|
|
for (int i = 0; i < sizeof(rtcGpios); i++)
|
|
rtc_gpio_isolate((gpio_num_t)rtcGpios[i]);
|
|
|
|
// FIXME, disable internal rtc pullups/pulldowns on the non isolated pins. for inputs that we aren't using
|
|
// to detect wake and in normal operation the external part drives them hard.
|
|
|
|
// We want RTC peripherals to stay on
|
|
esp_sleep_pd_config(ESP_PD_DOMAIN_RTC_PERIPH, ESP_PD_OPTION_ON);
|
|
|
|
#ifdef BUTTON_PIN
|
|
// Only GPIOs which are have RTC functionality can be used in this bit map: 0,2,4,12-15,25-27,32-39.
|
|
uint64_t gpioMask = (1ULL << BUTTON_PIN);
|
|
|
|
#ifdef BUTTON_NEED_PULLUP
|
|
gpio_pullup_en((gpio_num_t) BUTTON_PIN);
|
|
#endif
|
|
|
|
// Not needed because both of the current boards have external pullups
|
|
// FIXME change polarity in hw so we can wake on ANY_HIGH instead - that would allow us to use all three buttons (instead of just the first)
|
|
// gpio_pullup_en((gpio_num_t)BUTTON_PIN);
|
|
|
|
esp_sleep_enable_ext1_wakeup(gpioMask, ESP_EXT1_WAKEUP_ALL_LOW);
|
|
#endif
|
|
|
|
esp_sleep_enable_timer_wakeup(msecToWake * 1000ULL); // call expects usecs
|
|
esp_deep_sleep_start(); // TBD mA sleep current (battery)
|
|
}
|
|
|
|
/**
|
|
* enter light sleep (preserves ram but stops everything about CPU).
|
|
*
|
|
* Returns (after restoring hw state) when the user presses a button or we get a LoRa interrupt
|
|
*/
|
|
esp_sleep_wakeup_cause_t doLightSleep(uint64_t sleepMsec) // FIXME, use a more reasonable default
|
|
{
|
|
//DEBUG_MSG("Enter light sleep\n");
|
|
uint64_t sleepUsec = sleepMsec * 1000LL;
|
|
|
|
Serial.flush(); // send all our characters before we stop cpu clock
|
|
setBluetoothEnable(false); // has to be off before calling light sleep
|
|
|
|
// NOTE! ESP docs say we must disable bluetooth and wifi before light sleep
|
|
|
|
// We want RTC peripherals to stay on
|
|
esp_sleep_pd_config(ESP_PD_DOMAIN_RTC_PERIPH, ESP_PD_OPTION_ON);
|
|
|
|
#ifdef BUTTON_NEED_PULLUP
|
|
gpio_pullup_en((gpio_num_t) BUTTON_PIN);
|
|
#endif
|
|
|
|
gpio_wakeup_enable((gpio_num_t)BUTTON_PIN, GPIO_INTR_LOW_LEVEL); // when user presses, this button goes low
|
|
gpio_wakeup_enable((gpio_num_t)DIO0_GPIO, GPIO_INTR_HIGH_LEVEL); // RF95 interrupt, active high
|
|
#ifdef PMU_IRQ
|
|
// FIXME, disable wake due to PMU because it seems to fire all the time?
|
|
// gpio_wakeup_enable((gpio_num_t)PMU_IRQ, GPIO_INTR_HIGH_LEVEL); // pmu irq
|
|
#endif
|
|
assert(esp_sleep_enable_gpio_wakeup() == ESP_OK);
|
|
assert(esp_sleep_enable_timer_wakeup(sleepUsec) == ESP_OK);
|
|
assert(esp_light_sleep_start() == ESP_OK);
|
|
//DEBUG_MSG("Exit light sleep b=%d, rf95=%d, pmu=%d\n", digitalRead(BUTTON_PIN), digitalRead(DIO0_GPIO), digitalRead(PMU_IRQ));
|
|
return esp_sleep_get_wakeup_cause();
|
|
}
|
|
|
|
#if 0
|
|
// not legal on the stock android ESP build
|
|
|
|
/**
|
|
* enable modem sleep mode as needed and available. Should lower our CPU current draw to an average of about 20mA.
|
|
*
|
|
* per https://docs.espressif.com/projects/esp-idf/en/latest/api-reference/system/power_management.html
|
|
*
|
|
* supposedly according to https://github.com/espressif/arduino-esp32/issues/475 this is already done in arduino
|
|
*/
|
|
void enableModemSleep()
|
|
{
|
|
static esp_pm_config_esp32_t config; // filled with zeros because bss
|
|
|
|
config.max_freq_mhz = CONFIG_ESP32_DEFAULT_CPU_FREQ_MHZ;
|
|
config.min_freq_mhz = 10; // 10Mhz is minimum recommended
|
|
config.light_sleep_enable = false;
|
|
DEBUG_MSG("Sleep request result %x\n", esp_pm_configure(&config));
|
|
}
|
|
#endif
|