stevend wrote: ↑
Sun Dec 03, 2017 3:04 pm
If using one MCP3008, its ground needs to be connected to the negative of the Pi's power supply, and to the negative of the complete battery assembly.
You then have several choices:
1. You can use separate potential dividers to measure the voltages of 1, 2, 3 and 4 batteries in series, and subtract to get the voltage of each individual battery. If you don't need great accuracy, this may be good enough (but remember that the potential dividers take current at all times).
2. You can use several A-D converters, electrically isolated from each other and the Pi, each measuring the voltage across a single battery. This gets a bit more complicated hardware-wise, but is probably the most accurate solution.
3. You can use an A-D converter in differential mode. Not sure that the MCP3008's 'pseudo-differential' mode is good enough for this; you may need to either use a better A-D converter, or use op amps in differential mode to generate a single-ended voltage input.
4. Sure I've seen specialist chips designed to do exactly what you want (it's a fairly common problem); suspect these won't be available on easy to use PCBs. Have a search if you've got the hardware building skills.
Thanks so much for your detailed reply.
#1 Looks simple, but like you said, a crap shoot.
#2 Interesting, but not sure if I can get the Pi to talk to 4 MCP3008s
#3 Thought about this route from some search results I've found - a few examples out there
#4 Would like to go with this one, the Maxim 14920 chip looks ideal, but hard to tinker with on a breadboard. I've queried for an eval kit. Don't know if they are avail still. Going to look into fully built products, but they may be $$$
I'll post some progress