LiFePO4 system design considerations and sources for off-grid battery bank replacing lead acid

Dave said:

I originally posted the above comment on a different thread and was advised to start new thread so I copied this comment from mcgivor over here for follow-up:

mcgivor Solar Expert Posts: 3,180 ✭✭✭✭✭

November 22 edited 5:19AM #9

Don't be intimidated it's not that complicated to build a DIY bank once the fundementals are understood.

I appreciate that!

For a BMS I contacted a manufacturer, they requested specifications, maximum load current, maximum charge current, battery capacity and charging source etcetera, their engineering department then recommend one based on my reply. 

Both the batteries and BMS I  ordered were from China, the support has been excellent from the supplier's I used. No experience with Electriccarparts.

Would you mind sharing the name/contact info for who you worked with for batteries and BMS?

Shallow cycles will extend the life expectancy, they can be used at partial states of charge without negative impact as well.

The only time they really need to be charged to 3.65 VPC is when initial ballancing is performed and perhaps every couple of years, under normal cycling it's best to only charge to ~3.40 VPC, somewhere in the 90% region, likewise discharge to 20% maximum which results in 70% of nominal capacity used.

ElectricCarParts (ECP) advised initial balancing by hooking all batteries up in parallel first for 8 hours or so before charging. Are you referring to that balance or the balancing that the BMS does, which I thought would be occurring all the time?

The cells would be arranged parallel first then series, for 600Ah, 6 100Ah in parallel to form a cell block, the BMS sense wire for each cell will be connected to the cell block, 8 for 24V, 16 for 48V nominal. The ballancer would be redundant because the maximum voltage would be below it's operating range using  3.4 VPC , to understand this you need to consider how the cells are ballanced in the first place.

This is where I am most confused. The BMS I was planning on would have one attachment (sense wire I guess) to each 6-cell parallel bank and the assumption (via ECP) would be that all cells in that group were at the same voltage. Then, I thought the BMS would distribute the charge current among the 16 groups as needed to keep them balanced and in range. I thought this would occur any time charging was happening even if the voltage was below max V. There is a more expensive BMS that monitors every cell but that seemed over the top, not so?

    This sounds very different than the balancing process you described so I could use some more tutoring! This aspect is important because I like the idea of setting the max voltage at 90% of max V but not if that means the BMS won't be doing its job automatically and constantly.

In EV applications the cells would typically be bottom ballanced to allow maximum capacity availability, in solar applications the cells are typically top ballanced, in both cases all cells or cell blocks are connected in parallel, then either charged to maximum voltage or discharged to a minimum voltage value. The cells can be top or bottom ballanced, but not both, with top ballancing a 3.65VP LiFePo4 single cell charger can be used to safeguard aginst voltage runaway which happens extremely fast without a proper charger, likewise with discharge things happen extremely fast, both over or under voltage can result in permanent damage to the cells.

The links below are a great source of general information, not nesesarally specifically for solar but useful for general information.
http://liionbms.com/php/bms_options.php#7
http://nordkyndesign.com/assembling-a-lithium-iron-phosphate-marine-house-bank/

The attached pdf is the BMS I ordered, it costs ~$100

One suggestion, start a new discussion to concentrate response to your particular project.