I've been offline for 4 weeks in the Eastern San Blas...so a couple of general comments on all the posts for that time to catch up....
I've been using 48 100Ah cells connected 12P/4S making a 1200Ah house bank for about 6 months now. Several posts on my install are way back in this thread. The batteries have been great so far, some issues with the BMS, but mostly resolved now after a replacement part.
Initial Balance/Charging
--------------------------
Unless you have perfectly capacity matched cells (unlikely), its impossible to have a bank where cells remain at the same voltage at both top and bottom of charge. For balancing on initial
installation, I believe a
marine installation must be top balanced. That means all cells are equalized at full charge since that's a condition we see in
marine every time we
plug in at a marina, or have bright sunny days with
solar. We should not see below 20% SOC normally...and if we do under normal operation...more battery capacity is probably needed.
Many of the EV guys bottom balance, but that's because their requirements are relatively high discharge rates and an occasional need to go below 10% SOC in order to get home.
As already mentioned, the only way to do this is to initially balance by paralleling all the cells for a few hours and then either charge them with low voltage
power supply to 3.6V or connect in series and tinker with each cell until they match at 14.4V or 3.6V for each cell. I found a 100 foot piece of 16-14 gauge wire (poor mans resistor) worked well with an isolated 6V golf cart battery to top off an individual cell. The same wire can be used to short a cell and remove charge.
My personal opinion is you should limit the initial balance charge to a maximum of 3.6V per cell and to allow for some possible cell imbalance during operation, only charge to 14V or 3.5V per cell. I'd prefer even lower, 13.8V, but my charging systems are not adjustable enough. There is some evidence that suggests cycle life is extended by avoiding the extreme ends of the charging curve by keeping the cell voltage within the range of say 3.0-3.5V. It's possible that the Winston/Thundersky chemistry is more tolerant of higher voltages based on their specs, but I'm not convinced yet.
Even with the bank perfectly top balanced at 3.6V, the voltage of the cells do not track together above 3.35V even though they all reach 3.6V at the same time. For example, I've seen the cells range from 3.45 to 3.5V during charge, but lining up exactly at 3.6V at the end.
One question I have is if anyone using them has found a need to rebalance after some period of time? I don't have enough cycles on my bank since I replaced a bad cell monitoring board that was drawing down one cell more than the others.
Bank Disconnect/Separate Charge and Discharge Paths
-----------------------------------------------------------
I'm in the camp that believes you should have some sort of cell based monitoring system to disconnect the bank if any cell reaches too low or too high a voltage. I'm using >3.6V and <2.8V now as limits.
The problem for many is inverter/chargers which don't support different paths for charging and inverting. Also, if you want to use the Junsi Celllog 8 as a basic
monitor, it only has one connection for a disconnect.
To solve this problem, I leave my starting batteries, a pair of AGM's connected in parallel to my Li house bank. It's not elegant, but it works. Should the Li bank disconnect, the AGM's are still connected. While the Li are discharging, it's basically a "float" voltage for the AGMs at around 13.2V, and during charge I normally only get to 13.6-13.8 before I turn off the genset and this does not seem to overcharge the
AGM. The cost is the trickle charge to the AGM's at under 1 amp, about 10-15 Ah per day.
Eventually I plan to hook a solenoid up to the
engine switch such that the parallel connection is only made underway, since at
anchor, I'm not as concerned about a sudden disconnect. Not sure how to do that for a sailboat.
Safety Issues
--------------
I don't know the battery chemistry in the Boeing case, but I think Cessna also used lithium in a certified design in production. They also had a fire, on the ground in a customer aircraft and subsequently issued an AD directive to remove the Li from all their aircraft.
I doubt that either was lithium ferrous,
LiFePO4. But I don't know that. Testing has shown that LiFePO4 is very unlikely to explode into a fire on their own, unlike other Li cells can. Under most any condition. They are undoubtably the safest Li chemistry available today. But they can create quite a bit of heat, especially if overcharged. There is also a tremendous amount of energy potential in a fully charged bank that deserves a lot of respect, just like any other battery. But even more so since it will deliver it almost all at once.
I believe that cell voltage monitoring/disconnect and appropriate bank fusing offers the needed protection to reduce the risk to acceptable levels. Just like we live with
propane tanks,
fuel tanks .... even
fiberglass boats instead of
steel, but take steps to mitigate those risks. But this is virgin territory and there is more to learn and probably even safer versions of the batteries themselves as the technology is adopted.
We know they will vent gas if severely overcharged or shorted internally. See the YouTube videos of the Sinopoly testing. This venting is not a likely, but is a possible event. How possible and under what "normal" condition is the first unanswered question since I don't plan on shooting one with a 45 caliber bullet. And second, nowhere can I find
documentation of the volume or composition of the gas vented, specifically for a LiFePO4 cell. I'm pretty sure it will vary based on the electrolyte, which may not be the same for every manufacturer. It may turn out some sort of active/passive/contained ventilation to the outside is desirable? I don't know. If anyone has seen data on this vent gas, please post it. I think this area represents the most significant unanswered safety questions.
Bob