%XVBAT Accuracy

[ProfHindsight]

We're looking to use the nRF91 to measure the LiPo battery voltage to determine battery health. We would want to know what the accuracy and range of %XVBAT is. Looking at the module, it is either able to use 0.6V as a reference or VDD / 4 as a reference. My original thought was that the 0.6V internal reference was used to measure the voltage of VDD. However, the maximum range using the internal reference is 3.6V. We need to report when the battery is below 85% and below 10%, so it seems that the 85% voltage would be too high to use this method.

This is also assuming VDD is connected to VDD and not VDD_GPIO.

Thanks!

Colton

[awneil]

ProfHindsight said:

measure the LiPo battery voltage to determine battery health

See: devzone.nordicsemi.com/.../171088

[ProfHindsight]

Thank you for the quick response awneil!

It doesn't quite answer my question. The nRF91 is able to take an input voltage of 3.0 - 5.5 on the VDD pin. The LiPo battery will have a range of roughly 3.3 - 4.5V.

The AT command in question can be found under section 4.23 of nrf91_at_commands_v0.7

The full-scale resolution note of 3.6V is found in section 6.12.8 of nRF9160_OPS_v0.7.1

One question is whether VDD as referenced by the SAADC is VDD, VDD_GPIO, or some other internal regulator on the nRF91.

We currently have the STC3117 in the design but testing it has showed it to be not that accurate in current-voltage mode and the voltage mode seems to measure the voltage and compare against a lookup table. It's possible it does more calculations under the hood - we would include both methods in the design and run a comparison.

Thanks!

[awneil]

ProfHindsight said:

It doesn't quite answer my question.

True.

The point is that the whole idea of using battery voltage to gauge remaining capacity is deeply flawed - so you may be better pursuing a better approach rather than messing about with trying to perfect the voltage reading ...