So @dynamicdave creates a nice Single Sensor Node PCB board and both he and I ordered them and built them. Dave is clever and was able to put two boards on each PCB one using a USB jack to provide and the other a two pin header to connect a battery to it. They can run using a Wemos-D1-mini or ESP32-S2-mini (our choice). They talk to Node-RED via wifi and MQTT.
So I had a couple batteries from an old UPS and gave them a charge and decided to run the board off of it. Here is what it looks like (a nice layout if I says myself):
I put it outside in my shed and let it run. It was a cold night and got below freezing and the battery voltage dropped to 5.44V but was only providing 2.4v to the circuit and so it stopped working about 8AM so I brought it inside.
This morning I decide to charge up the battery and then see how long it would run in the house. I attached my 12V auto battery charger to the battery (which I used to originally charge it) and then decided to flip the switch to turn on the circuit so it could start reporting data.
That was a bad...bad idea! In less than a 1/2 second there was a POP and a smell in the air. I hit the switch to turn the circuit board off and decided I should walk away for a while.
When I came back after a bit, I took the battery off the charger and removed the PCB board and took a look and this is what I saw:
What you see is the remains of a 100uF 15V capacitor. After cleaning the board up and replacing the capacitor, I discovered that the surge fried the ESP32-D1-mini too (sob).
So learn from my lesson...don't test a circuit while connected to a 12V auto battery charger.
The good news is that with a new ESP32-S2-mini, the board is working again.
If it is a dumb charger it probably has very little smoothing on the output so it could well have spikes higher than 15v, and also a large amount of ripple that your cap was trying to absorb. Hence the bang.
Ouch. I've only had one capacitor go on my like that. An old high voltage job in an aluminum can (for use with vacuum tubes). Thankfully the pressure relief valve worked. Boy did it stink though!
As Apprentices, there were those among us that would plug an Electrolytic Capacitor into a Mains switched socket. At switch on they would go bang, the Capacitor that is, and hopefully not the Apprentice. You would find two wires in the socket and large amount of fine confetti floating around the Workshop!
I remember a spectacular "failure" of an analogue oscilloscope I was working with at Uni (yes, no digital scope then) - thankfully determined not to be my fault!
More spectacular still was violently releasing the blue smoke on a very expensive customer Dell server after fitting more RAM. That gave me some grey hairs as I was in the act of pressing the on-switch as it blew up. Also thankfully determined not to be my fault but rather the fault of the customer who wouldn't pay for OEM memory, the 3rd-party memory was slightly out of spec and shorted the main power bus. No more server!
Not too bad a track record for over 40 years of dealing with expensive electronics I guess.
You do of course realise that ALL electronic and electrical devices work on smoke, if the smoke escapes from them, they don't work any more! (Although some can limp along)
Also thankfully determined not to be my fault but rather the fault of the customer who wouldn't pay for OEM memory, the 3rd-party memory was slightly out of spec and shorted the main power bus. No more server!