significant Battery voltage drop

[nirmal@nordic]

Hi, Everyone,

I am using NRF51422 chip for my custom board. with transmission interval set to 2000ms and Tx power to -20 dBm, I transmit 32 Bytes BLE data and Scan and response packet with scan request.

Since, battery lifetime is biggest concern we measure the average current consumption with multi meter logging data over period of 16 hours and avg current measurement measured is 0.022 mA.

With this measurement our battery(1000mAh) is good for 3 years, with following calculation

no of hours = battery capacity(mAh) / current consumption(mA) * 0.7 no of hours = 1000/0.022*0.7 = 31818.18182 hours

Ref: www.digikey.ca/.../conversion-calculator-battery-life

**But somehow, beacons chip stopped working in 1 month as battery voltage dropped from 3.1890 to 2.48. I don`t see any advertised data from beacon on Master control Panel.

Even same chip do not go into debug mode with keil with 2.48v battery as keil says "No cortex device found"

Can somebody explain me the reason of this drastic voltage drop and NRF chip not working at all ??? I am using S130 NRF SDK 11.0.2 - alpha for NRF51422 chip. ** Thank you.

[Stefan Birnir Sverrisson]

Hi

Since you are measuring normal current consumption, but still your battery drains in the long run, something must be happening in the meantime that causes higher current consumption. Two suspicions come to my mind:

• The device enters debug mode at some point. See this thread for the cause and solution.
• The S130 softdevice you use is prior to v1.0.0 (or prior to S110 v8.0.0), then the bootloader consumes a lot of current, see this thread. I doubt anyhow that you have this old softdevice as you use SDK 11.0.0-2.alpha, which should go together with S130 2.0.0-7.alpha, which should not have this bug.

[Stefan Birnir Sverrisson]

1. Sorry for the confustion. The S130_nrf51_2.0.0-7.alpha_softdevice.hex is a softdevice which is totally independant of a bootloader. So I conclude that you do not have a bootloader, so we can rule this bootloader thing out as cause of the problem.
2. I would think the safest approach is to leave all unused GPIOs as unconfigured, which means they are not configured as input or output, i.e. not with drive connected and input buffer also disconnected. That will make every GPIO have impedance of >10Mohm which will not create any noticeable leakage current.
3. Ok, good to know that the problem is consistent.

My highest suspicion is still that the device is entering debug mode at some point, since you measure low current consumption in the lab but still the battery drains quickly in the field. When in debug mode, the current consumption increase is around 1mA compared when not in debug mode. The expected battery lifetime with 1mA current consumption is 1000/1*0.7 = 700 hours = 29 days, which is pretty much on par with what you are seeing.