Originally Posted by Captain Canuck
I need some opinions, guys.
The Particle photon runs on 3.3-6.5 vdc. 5.5 is recommended.
The only solution I can come up with to keep it powered 100% of the time is to run it off of the battery itself through a DC-DC converter. Is there a better way?
I'd like to run each battery on it's own BMS, but I'm having a hard time finding relays that work off of the particle's 5v 100mA output. There are tons of them for 120V, but when I start looking for 48VDC 200A relays, I only find relays that the particle has no hope of pushing closed. If I use a 120vac relay, that means either one battery shuts the whole pack down, or I need a charger for each battery. It would be a lot less expensive if I could just find a relay for each battery. While I'm sure using one BMS for the whole system would work, It's more optimal to do a bms for each battery.
As always, I'm open to ideas.
any relay running at 5v/100ma is a tiny little relay that doesn't have the spark gap to stop the other 46v + inductance of the whole circuit (Read: You will get a voltage surge when your relay trips, and that will weld the contacts the first time it opens under load)
The only thing your individual cell monitor needs to do is bleed the cell for balancing and alert for under voltage/over voltage. The alert lines should be opto isolated for safety to the main BMS unit, and engineer it to it keep the alert lines at the same relative voltage to each other (since they'll have a lot less insulation on them than the primary conductors). There's no point in putting a relay on each cell since the cells are all in series...any one relay trips you shut the whole bank anyways. You cannot realistically take one cell out of the circuit, since each set of relays to do that would have to be rated at max system current and voltage. That's expensive.
Better and far less expensive to use one big assed fat contactor to shut the bank down and just trigger it on HV or LV events.
The individual cell monitor is probably better off a very simple low power 1.2v digi device with a dedicated AD, or a pure analog device, though pure analog devices are hard to calibrate/maintain calibration over temperature ranges. A commercial solution that has had decades of sound engineering behind it coupled with practical application experience is really the only safe way to go esp. considering your desire to use less forgiving LiIon vs. LiFePO4.
If I remember correctly there are some BMS projects on https://endless-sphere.com/forums/
if you want to look at people's design thoughts on the subject. After you become familiar with the sheer magnitude of what you are trying to do I think you may reconsider either the chemistry or the bms(Though again, i recommend reconsidering both.) I looked into BMS design figuring it can't be that hard. And I'm really familiar with eletronic design, have had my own circuit boards printed, have etched them myself, built transmitters, antenna tuners, antenna tuner control circuits, programmed Microchip PIC and Atmel microcontrollers, I code in a plethora of languages c c++ python etc... and while I enjoyed learning about BMS design, I decided it would be hubris to pursue it.
YMMV but I don't recommend it.