Question regarding nrf_fuel_gauge_idle_set behavior and SoC estimation during charger connection

[Tcunha]

I am using the nrf_fuel_gauge_idle_set function to handle the fuel gauge state when the system enters a low-power mode. When the charger is connected, I exit this state and call nrf_fuel_gauge_process.

However, connecting the charger causes a voltage spike, and this spike is fed into the nrf_fuel_gauge_process call. The SoC estimated by nrf_fuel_gauge_idle_set does not seem to update, as the SoC value remains unchanged.

Is this behavior expected? Would it be better to use the battery voltage measured just prior to connecting the charger instead?

[Kazi Afroza Sultana]

Hello,

This seems a consistence behavior with how the library reacts to transient measurement.

According to this blog post (+) Advanced power management and fuel gauging for designs with non-Nordic MCUs - Blogs - Nordic Blog - Nordic DevZone, ''When calling the fuel_gauge_update() function, this will in turn call  nrf_fuel_gauge_process() function, you must provide a t_delta value (as a float) [nRF Fuel Gauge API]. This value must be the time difference between the previous gauge update and the current update, not the time elapsed since the system started.''

You can see the parameters of the nrf_fuel_gauge_process function here nrfxlib API: nrf_fuel_gauge_process. 

For example: if you call the update every 1s, the t_delta should be approx 1.0 and if you skip update and the next call is after 10s, the value should be approx 10.

You can also look look at this previous case (+) nPM1300 charging battery and getting a correct SOC with the Fuel Gauge lib from nRF Connect SDK V2.6.0 - Nordic Q&A - Nordic DevZone - Nordic DevZone.

[Tcunha]

Thanks for the info. I have one more question: do you think a constant 65 µA current draw justifies the use of nrf_fuel_gauge_idle_set() ?

[Kazi Afroza Sultana]

Hello,

The sleep current configuration instructions described here have been followed:

https://docs.nordicsemi.com/bundle/nan_045/page/APP/nan_045/fg_configuration_sleep.html

nrf_fuel_gauge_idle_set() already accounts for the sleep period. On wake-up, the fuel gauge will not update the SOC based on a single transient VBAT sample (such as a charger-attach spike), so seeing no immediate SOC change is expected. After a few stable process() iterations, the SOC should begin updating normally.

Could you also share a short log from the fuel-gauge library output covering the following?