It's been a couple of months but I just run into this thread.
I was on a path near yours, planning on using individual charging circuits fed by separate isolated DC/DC converters (like MEJ2S1205SC) for charging the battery pack. I'm using LifePO4 batteries to start with, which is about the most common "safe" rechargeable Lithium Ion chemistry, or at least safer than many other and not that hard-to-get.
They have the great advantage of using the same charger as Lead-acid batteries when handled in packs with an even number of cells. LiFePO4 are charged up to a maximum of 3.65V. One 12V lead-acid battery is charged up to 14.5V (= 4 x 3.625V) with the same current limited constant voltage profile. AND, some of them can be charged at 2C or 3C, while supporting a discharge current of up to 30C (yes! thirty C!) Of course, safety comes at the cost of less capacity than other lithium chemistries, but still better than NiMHs. Plus a typical endurance of 2000 (two thousand) charge/discharge cycles.
While searching the internet I came across a few boards sold on ebay for protecting multicell battery packs. They looked simple enought, but these boards were too big for my pack. A little more searching turned up the bq77910a, which combines pack overcharge and overdischarge protection with (hardware-configurable) 50mA cell balancing. It's highly configurable for any type of lithium chemistries I've come across.
Given their very flat discharge voltage of 3.2V, gaunging was being a challenge until I discovered TI's bq34z100, a "battery fuel gauging" chip designed to be used in conjunction with the aforementioned chip, and capable of tracking the battery internal impedance by monitoring charge and discharge voltage and current, and self-adjusting its parameters to always keep a good estimation of battery capacity untill its EOL.
There are Evaluation Modules for both, but you have to order the programming hardware separately, and they use different ones.
@Duane, I read an article on charging Lithium chemistries and it seems that current limiting to less than the maximum charge current while restricting the maximum open terminal voltage is all they need. This means that if you have a load cell in parallel with each cell that bleeds current to avoid charging above ~3.6V per cell (depends on exact chemistry) is the way to go. Then you can put the cells in series. The current will either flow through the bleed circuit (above the cell voltage) or charge the cell (below the cell voltage. When the cell voltage == the charge voltage no current can flow so you can't overcharge the battery.
I used a precision low current opamp and reference driving a power transistor across each cell that would bleed at 3.7V (for my chemistry) and then charged with an 11.1V 5A supply.
I hadn't considered that type, but with a bit of looking I can see that they seem to have good availability and reasonable pricing. Not as dense as LiPoly, but still pretty good. I really like the fact that it's more robust. LiPoly make me nervous.
Check out the ams AS8506C autonomous balancer: passive or active mode up to 7 cells. Integrated w/ temp sensor +2 temp sensor inputs, all necessary analog functions & switches built in. No error prone code or micro needed. Only 3 resistors needed: one for discharge 2 for high and low thresholds. Balancing & Monitoring now simplified in the real analog world.
Duane -- the LiFePo4 Batteries are used in Power Tools for Motor Drive (Will take more abuse than Li-Poly Battery) -- Something to think about for a mobile robotics device -- also they will take about 350C heat rather than 100-150C heat(failure point for LI-Poly)
I will be designing a charger for a series stack of LiFeO4, to reach a high voltage. I plan to make individual, Isolated chargers for each sub-pack. This is for my regular day job. However, I will share my general approach, so you all can have some fun.
What are the engineering and design challenges in creating successful IoT devices? These devices are usually small, resource-constrained electronics designed to sense, collect, send, and/or interpret data. Some of the devices need to be smart enough to act upon data in real time, 24/7. Specifically the guests will discuss sensors, security, and lessons from IoT deployments.