Comparison of Power Consumption between PPI and APP_TIMER

[GUO-dong]

I am working on a SAADC project using SDK 15.3 & nRF52832, and I learned from the forum that using PPI can allow the CPU to be in an IDLE state while still measuring the ADC, theoretically saving power. I compared two projects: one using APP_TIMER and the other using PPI with nrf_drv_timer.

However, I found that the power consumption of nrf_drv_timer itself is higher than that of APP_TIMER, which contradicts the concept of using PPI to save power. Is this really the case, or am I doing something wrong?

Here is my function code for reference:
1. APP_timer：

static void BLE_ctrl(void* p_context){adc_action=true;}

APP_ERROR_CHECK(app_timer_create(&ADC_timer, APP_TIMER_MODE_REPEATED, ADC_ctrl));
APP_ERROR_CHECK(app_timer_start(ADC_timer, APP_TIMER_TICKS(detect_duration), NULL));
while(1){
    do
    {
        idle_state_handle();
    } while (!adc_action);
    for(int i=0;i<SAMPLES_IN_BUFFER;i++){
        APP_ERROR_CHECK(nrf_drv_saadc_sample());
    }
}

2. ppi with nrf_drv_timer：

static void timer_handler(nrf_timer_event_t event_type, void* p_context)
{}
static void saadc_sampling_event_init(uint16_t time)
{
    ret_code_t err_code;

    // err_code = nrf_drv_ppi_init();
    // APP_ERROR_CHECK(err_code);

    nrf_drv_timer_config_t timer_cfg = NRF_DRV_TIMER_DEFAULT_CONFIG;
    timer_cfg.bit_width              = NRF_TIMER_BIT_WIDTH_32;
    err_code                         = nrf_drv_timer_init(&m_timer, &timer_cfg, timer_handler);
    APP_ERROR_CHECK(err_code);

    /* setup m_timer for compare event every 400ms */
    uint32_t ticks = nrf_drv_timer_ms_to_ticks(&m_timer, detect_duration);
    nrf_drv_timer_extended_compare(&m_timer, NRF_TIMER_CC_CHANNEL3, ticks,
                                   NRF_TIMER_SHORT_COMPARE3_CLEAR_MASK, true);
    nrf_drv_timer_enable(&m_timer);

    uint32_t timer_compare_event_addr =
        nrf_drv_timer_compare_event_address_get(&m_timer, NRF_TIMER_CC_CHANNEL3);
    uint32_t saadc_sample_task_addr = nrf_drv_saadc_sample_task_get();

    /* setup ppi channel so that timer compare event is triggering sample task in SAADC */
    err_code = nrf_drv_ppi_channel_alloc(&m_ppi_channel);
    APP_ERROR_CHECK(err_code);

    err_code =
        nrf_drv_ppi_channel_assign(m_ppi_channel, timer_compare_event_addr, saadc_sample_task_addr);
    APP_ERROR_CHECK(err_code);
}
// main loop
while(1){
    idle_state_handle();
}

Here is the measurement chart(using PPK2):

[Vidar Berg]

Hello,

PPI can be used to offload the CPU, which can result in more precise timing and reduced power consumption, but this is not always the case. As you observed, the TIMER's current consumption is approximately 500 uA, compared to around 0.1 uA for the RTC used by the app timer library.

Best regards,

Vidar

[GUO-dong]

Thank you for your response.

Does this mean that I should use the APP_TIMER solution?
Or are there other things I can do with PPI and nrf_drv_timer to help reduce power consumption?

Currently, I only enable the low power mode for SAADC and disable it when not in use.

Power consumption is a big concern for me...

[Vidar Berg]

Yes, using the app timer is what will give you the lowest power consumption. The time it takes to wake up and service the RTC interrupt should have minimal impact on the total current consumption. 

Are the measurements you posted from a DK or a custom board? The base current seems a bit high. It should be closer to 2uA if you compare it to the online power profiler:  Online Power Profiler for Bluetooth LE 

[hmolesworth]

Instead of using a Timer for PPI instead maybe try using a spare RTC channel as the timer for PPI; the power consumption may not be as low as app timer but will be lower than using a Timer ..

[Vidar Berg]

You can use PPI with the RTC as hmolesworth  suggests, but you would need to select RTC2 since RTC1 and RTC0 are being used by the app timer and softdevice.