nrf52 TWI excess current consumption

[cedric]

Hello, I use an I2C device to wake up a nrf52 microcontroler. If I just use the init function

static ret_code_t twi_master_init(void)
{
    ret_code_t ret;
    const nrf_drv_twi_config_t config =
    {
       .scl                = 6,
       .sda                = 5,
       .frequency          = NRF_TWI_FREQ_400K,
       .interrupt_priority = APP_IRQ_PRIORITY_LOW
    };

    do
    {
        ret = nrf_drv_twi_init(&m_twi_master, &config, NULL);
        if(NRF_SUCCESS != ret)
        {
            break;
        }
        nrf_drv_twi_enable(&m_twi_master);
    }while(0);
    return ret;
}

and then

int main(void)

{

	softdevice_init();

	scheduler_init();

	twi_master_init();

	sd_power_system_off();

}

the system has a 0.1mA (or less, but that's not the point today) sleep current.

But if I add a line (to disable the TWI peripheral, and it's the same when I programm a data transfer)

int main(void)
{

  softdevice_init();

  scheduler_init();

  twi_master_init();

  nrf_drv_twi_uninit(&m_twi_master);

  sd_power_system_off();

}

the system current is 0.6mA. I've checked the SCL and SDA lines for pull ups issues, but there are at a high level (there are 10k pull ups outsde the nrf52), so that I don't think it's the case. I've added

NRF_TWI0->ENABLE=TWI_ENABLE_ENABLE_Disabled << TWI_ENABLE_ENABLE_Pos;

To force the peripheral off, but nothing better happens. Iv'e set the Pins SDA and SCL as input, outputs or float, but nothing more happens too. I've added those lines too :

nrf_twi_task_set(m_twi_master.p_reg,NRF_TWI_TASKS_STOP);
while(!nrf_twi_event_check(m_twi_master.p_reg,NRF_TWI_EVENTS_STOPPED ));

But again, nothing. There must be something I've missed, but I don't know what...

[Susheel Nuguru]

I should have been able to solve this case earlier but I missed to see one errata. Errata 89 tells that this is an anomaly inside the chip. And the work around is given. Just add below code when you want to disable TWI

For TWI0 for number 89 anomaly on nRF52

*(uint32_t *)(0x40003000 + 0xFFC ) = 0;
*(uint32_t *)(0x40003000 + 0xFFC ) = 1;

For TWI1 for number 89 anomaly on nRF52

*(uint32_t *)(0x40004000 + 0xFFC ) = 0;
*(uint32_t *)(0x40004000 + 0xFFC ) = 1;

I tested this personally and it works.

[cedric]

ok that works. I thought I4ve tried this errata workaround in all the tries I've made, but it actually works.

[Paul Dunn]

I tried the above as well, and it does indeed solve the problem of excess current with the TWI disabled. However, once I've issued these instructions, I cannot get the TWI to work again. Errata 89 says "Reconfiguration of TWI is required before next usage". I tried re-initializing (nrf_drv_twi_init) and just re-enabling, but both return error codes, and the TWI does not work afterwards. I'm using SDK 11. Can you provide more details on what it takes to "reconfigure the TWI". The errata also mentions "Turn the TWI off and back on after it has been disabled". I assume that means to issue a nrf_drv_twi_disable prior to the above instructions; which is what I tried. Maybe provide a complete sequence of instructions to disable, then re-enable.

[Susheel Nuguru]

but both return error codes

What error codes do you get and for which API?

[Paul Dunn]

Aryan - I get an error code 17 (busy). I initialize using nrf_drv_twi_init and nrf_drv_enable, then when I want to sleep I issue nrf_drv_twi_disable prior to the two instruction in Errata 89 (using the volatile keyword, as the current does not reduce without it, presumably because the compiler optimizes out the instructions), then when I wake up I nrf_drv_twi_enable, but get error 17 when I try to use the TWI.

[Paul Dunn]

I'm still having the same trouble I described above. Can you confirm you think this solutions works when running a soft device? (I am using SD 332.) If I do a 'sd_app_evt_wait();' after the above, then a 'nrf_drv_twi_enable', I get error code 17 when I next try to use the TWI. Can you be more clear on what "Reconfiguration of TWI is required before next usage." (from errata 89) means.