Moteino M0 RFM95 LoRaWAN battery life test ends with dead battery [solved]

[mharizanov]

Hi,
I just completed a durability test using a Moteino M0+RFM95 and a 1400mAh LiPo. The goal is to see how many messages (20 bytes payload that includes battery reading and internal CPU temperature @ SF7) can it send before the battery dies. The node is sitting at room temperature. It managed to send 68323 messages out before failing. The battery discharge curve is as follows, there is an interesting bounce before it dies completely.

So now the problem is that the battery is completely dead, and plugging into USB won't charge it. Any tips to revive it?

[Felix]

Sounds like you have a bad battery?
The charger (MCP73831) should charge it unless the battery somehow is blocked by its own circuit board which it should include to protect the cell and prevent:
- over charging
- over discharging
- over load / shorts

Otherwise the MoteinoM0 and everything on it, have no preference about what batteries are used, even if they are chinese
This also means that this test is not indicative of performance of the M0 itself or the radio, but more of the battery. Often times cheap asian batteries (assuming yours is one) are pretty overrated and their capacity is far less than stated.

Did you unplug before attempting to recharge through the M0?
Do you have any other board with a similar charger that you can try?

[mharizanov]

Thanks for the Feedback,
It might be that the battery protection circuitry is either missing entirely or just bad. I am still trying to revive the battery, and therefore still reluctant to rip it open and check.
I don't recall where I procured that particular battery - probably I got what I paid for. Anyway, the lesson learned for me is to either shut down everything by software means when the battery goes below 3.5V or apply external protection circuitry for peace of mind.

The battery connections are fine, yes - I plugged and unplugged a few times.
So I wired the battery to a NiMH charger for a couple of minutes, so it measures roughly 3.6V now. Plugging back to Moteino still changes nothing.

Next, I remembered the Current Ranger has the same LiPo charger IC and plugged the bad battery there. It started charging, or at least the CHG LED is on. I'll leave it to see what happens.

Still curious as to why the Moteino won't charge it. I don't exclude shorting something sometime, as this particular (and only one I own) Moteino is abused heavily for tests.  Might order a replacement MCP73831 to test this possibility.

[mharizanov]

Just to follow up, I swapped the perfectly good battery I had in the Current Ranger into the Moteino, and it still won't charge. Obviously, I damaged something on the Moteino somehow.

[Felix]

Good catc, divide et impera always solves a lot of problems fast
The SOT23-5 MCP73831 should be easy enough to replace, flood with some solder and lift up, or even better if you have a dual wand solder station. I would not recommend hot air due to the plastic JST nearby, and delamination if too much heat is applied or for too long.

[mharizanov]

I just replaced the MCP73831 and all works as expected now.
Topic closed.