Unkown return from low power mode using sd_app_evt_wait()

[Matheus Bordin]

Hi,

I'm using the function sd_app_evt_wait() to go into low power mode, but it returns sooner than expected. 

I tried to figured out which event that is the cause for that with sd_evt_get() before and after sd_app_evt_wait(), but it always return (NRF_ERROR_NOT_FOUND No pending events). I also checked the NVIC registers and, apparently, it is something related with the RTC1 or GPIOTE (but, I don't think it is the GPIOTE). But I've tried to do 
NVIC_ClearPendingIRQ(GPIOTE_IRQn) and NVIC_ClearPendingIRQ(RTC1_IRQn) before sd_app_evt_wait() and it didn't work. 

Here is the NVIC state before sd_app_evt_wait():

And here it is after sd_app_evt_wait():

Here is my main loop:

void loop(void) 
{  
    nrf_drv_wdt_channel_feed(m_channel_id);
    app_system_tasks(); 
    app_ble_tasks();
    
   __set_FPSCR(__get_FPSCR() & ~(FPU_EXCEPTION_MASK));      
   (void) __get_FPSCR();
   NVIC_ClearPendingIRQ(FPU_IRQn);
   bsp_gpio_led_high();   
   APP_ERROR_CHECK(sd_app_evt_wait());
   bsp_gpio_led_low();
}

But, what makes me confuse is that when I put a second sd_app_evt_wait() in sequence, it works just fine. 

void loop(void) 
{
    nrf_drv_wdt_channel_feed(m_channel_id);
    app_system_tasks(); 
    app_ble_tasks();

   __set_FPSCR(__get_FPSCR() & ~(FPU_EXCEPTION_MASK));      
   (void) __get_FPSCR();
   NVIC_ClearPendingIRQ(FPU_IRQn);
   bsp_gpio_led_high();   
   APP_ERROR_CHECK(sd_app_evt_wait());
   APP_ERROR_CHECK(sd_app_evt_wait());
   bsp_gpio_led_low();
}

I used one LED to see the sleep time in comparison with the GPIOTE interruptions, as you can see below. The 'Channel 1' is the led state (1 - Power Down / 0 - Normal mode), while the others are GPIOTE interruptions. 

Finally, here is the graph when I utilized two sd_app_evt_wait() in sequence. It is just as I expected, the microcontroller just waked with the interruptions. 

How can I discover which event causes the return from sd_app_evt_wait()? Can it be something related with app_timer or with the watchdog as they use the RTC? But, I don't have any app_timer with a frequency comparable with the return from sd_app_evt_wait().

[Susheel Nuguru]

Hi Matheus,

NVIC_ClearPendingIRQ(GPIOTE_IRQn) and NVIC_ClearPendingIRQ(RTC1_IRQn)

 After clearing pended interrupts just after the wakeup, if you chip wakes up and the pend bits are again enabled, this means that the interrupt source has againi generated the interrupt. In this case you need to understand why RTC1 and GPIOTE interrupts are generating so frequently. Probably these interrupts are disabled in NVIC level but the peripherals are still generating the interrupt as they are enabled in NRF_RTC1 and NRF_GPIOTE.