Maximum Reload time for WDT

[Shagan]

Hi Guys, 

We have a need to send data every 4 hours. 

In order to do this, we sleep 1 hour at a time (1-hour timer - single shot) and increment counter, and when the count is 4 we send the data. This is working well. 

Due to unknown reasons the device sometimes can get hung when trying to send the data. 

Our plan was to set the WDT_CONFIG_RELOAD_VALUE to 120000000, to make sure we reset the device after 1 hour if it got stuck. 

But we are unable to provide values more than 600000 (which triggers at 100 seconds). 

Can u suggest a way to configure such a large interval? 

regards

Shankar

[Kenneth]

I am not sure what is the problem here, but there is watchdog example in the SDK that should be possible to use as reference. 

As a side comment, have you done any calculation on the current consumption?

The system ON idle current with an RTC running is ~1.7uA depending on how much RAM you retain during sleep.

If you wakeup every 100 seconds for ~50us then the average increase is ~0.002uA. In other words there is no difference in battery life whether you wakeup every 100 seconds or 1hour, because the idle current is about ~1000x larger.

Best regards,
Kenneth

[deknoob]

Hi Keneth,

I also want to put device to sleep for 1 hour, but with current even lower than ~1.7uA. Please let me know if this is feasible:

- disable all peripherals except WDT
- disable interrupts
- reload WDT with 1 hour wakeup
- Power off all RAM blocks
- halt the CPU

After one hour there will be reset, right. I can't do that with wakeup every 100sec. Is that correct?

Best regards

Kerim

[Kenneth]

You can find some current consumption values listed here:
https://infocenter.nordicsemi.com/topic/ps_nrf52840/_tmp/graviton/autodita/CURRENT/parameters.i_sleep.html

The current consumption with an RTC clock (LFCLK) running can get as low as I_{ON_RAMOFF_RTC} , typical 1.5uA. That equal about 13mAh a year.

Best regards,
Kenneth

[deknoob]

Hi Kenneth,

Yes, but 0.97uA in that link sounds better than 1.5uA 

- System ON, no RAM retention, wake on any event  =>  0.97uA

In my design, battery has to last 10 years, hence nanoamps count. And my device spends 90% of time sleeping.

Thing is that with no RAM retention only reset wakeup will work, where WDT comes handy.
You didn't answer my question, is it feasible what I propose?

Regards

Kerim

[Stian Røed Hafskjold]

The WDT needs the LF clock to be running. LFRC consumes 0.3uA in ULP mode (LFXO is at 0.25uA, but requires much longer startup time)

0.97uA (sys on, no RAM) + 0.3uA = 1.27 uA. (typical value)

I did some measurements here and I actually got a bit lower; 972 nA in sleep:

In my test the WDT is resetting every 10 seconds.

The wakeup charge is 5.9uC:

This gives an additional average current of 5.9uC / 3600 sec = 1.6 nA if the WDT is resetting the chip every 1 hour. Which is of course negligible, but with more advanced initialization code, this should maybe be taken into consideration. Here is my test code:

int main(void)
{
    NRF_CLOCK->LFCLKSRC = 0; // LFRC
    NRF_CLOCK->LFRCMODE = 1; // ULP mode
    NRF_CLOCK->TASKS_LFCLKSTART = 1; // Start LF clock
    while(!NRF_CLOCK->EVENTS_LFCLKSTARTED){
    }
    NRF_CLOCK->EVENTS_LFCLKSTARTED = 0;
    NRF_WDT->CRV = 327680; // WDT wakeup every 10 sec
    NRF_WDT->TASKS_START = 1; // start WDT
    for(int i = 0; i < 8; i++){
        NRF_POWER->RAM[i].POWERCLR = 0x3; // Turn off RAM block 0-7
    }
    NRF_POWER->RAM[8].POWERCLR = 0x3F; // Turn off RAM block 8
    
    for(;;){
        __WFE();
    }
}

[deknoob]

That sounds so good, thank you so much for having a look at this. I will test this as soon as possible and report results.

regards

Kerim

[deknoob]

Hi Stian,

I am actually using nRF52833 which doesn't have ULP mode. Best result is 1.4uA when I select LFXO. As you can see I use 1000 second WDT reload timeout, to enable longer startup time for LFXO. But I don't see variation if I wait for longer. With RC oscillator I am measuring 1.9uA.

I use NRF Power Profiler for measurement, VDD is 3V. 

Is this the best I can get with 52833? Or maybe my consumption is higher because of process variation?

1.4uA is still an improvement compared to 2.3uA sleep current measured in the past when I simply use timer to wakeup and with full RAM retention

    NRF_CLOCK->LFCLKSRC = 1; // LFXO
    //NRF_CLOCK->LFRCMODE = 1; // ULP mode
    //NRF_CLOCK->LFCLKSRC            = (CLOCK_LFCLKSRC_SRC_RC << CLOCK_LFCLKSRC_SRC_Pos);

    NRF_CLOCK->TASKS_LFCLKSTART = 1; // Start LF clock
    while(!NRF_CLOCK->EVENTS_LFCLKSTARTED){
    }
    NRF_CLOCK->EVENTS_LFCLKSTARTED = 0;
    NRF_WDT->CRV = 32768000; // WDT wakeup every 100 sec
    NRF_WDT->TASKS_START = 1; // start WDT
    for(int i = 0; i < 8; i++){
        NRF_POWER->RAM[i].POWERCLR = 0x3; // Turn off RAM block 0-7
    }
    NRF_POWER->RAM[8].POWERCLR = 0x3F; // Turn off RAM block 8
    
    for(;;){
        __WFE();
    }

[Stian Røed Hafskjold]

Hi, the base current is a little bit higher on 833.

ION_RAMOFF_EVENT

System ON, no RAM retention, wake on any event

1.1

µA

So 1.1 + .23 = 1.33 uA, which is the expected typical current. And yes, if you get higher (or lower) current it's because of process variation.