Powering nRF52810 with 3.6V Lithium battery

[George555]

I have designed some beacons based on nRF52810. The software uses the minimum functionality of the radio to transmit some packets every second and sleep.

In their sleep period the beacon consumes around 2-5uA.

I chose to power the beacons with an ER14250 3.6V Lithium battery, 1200mAh.

During an application testing I deployed around 10 beacons that transmitted a packet to a receiver every second. I noticed some of them drained the battery within days which is not reasonable as they should last over 2 years with that battery.

I tried to investigate the problem further. I programmed a beacon to transmit every 100ms.

A week after I found it dead. The battery was completely drained and also the chip was damaged.

I replaced the battery and it didn't work.

I used a bench power supply at 3.5V to check what was going on and I noticed its current consumption was around 150mA which is an indicator of a faulty chip.

The question is can that battery be used for nRF52810? In the datasheet it says absolute maximum at 3.6V yet in the recommended voltage range it mentions up to 3.6V.

So that's a little confusing as the absolute maximum voltage is the same as the recommended operating voltage.

Can that battery cause damage to the chip?

[helsing]

Looking at chapter 4. and 5.2 in the module's datasheet, the module seems to be quite sensitive. The datasheet also recommends being cautious and "reserve more than 30% of the margin", so this does means operating the module at 3.6V is risky.

[George555]

Indeed. Although it is very strange to me why they state that.

If you look at the picture I posted above you can see that the only active part inside the module is the nRF52810 SoC.

The rest of the circuitries are crystal oscillators, an antenna matching circuit and some capacitors. Nothing that would explain the difference from the SoC datasheet

[helsing]

If I was to guess, I would say that the special capacitors on the DEC-pins are causing the module to be very sensitive. The values for these DEC capacitors are normally not recommendations, they are to be followed strictly. The module manufacturer might have some good reason to deviate though.