Well at least I think it's solved. The mystery I'm referring to is why did some of the batteries suddenly start accepting more charge than they should have? My first assumption was that they'd moved in their state of charge relative to the others. But really I had no reason to believe that other than I couldn't think of any other reason.
It was then suggested to me that perhaps those particular cells simply have a lower capacity than the others so they reached their full state of charge first. I think that's likely the case, but truthfully it could be a combination of both reasons for all I know. But that really doesn't answer the fundamental question as to why, when they'd been behaving in a very predictable way for 5 months, did they suddenly start showing the tendency to run away?
Well I think I figured it out today, and I'll explain, but I need to give you some background first. I'd mentioned before that the Manzanita Micro charger has a little potentiometer screw that you use to set the cut off voltage for charging. When the pack reaches that voltage, it switches from constant current to constant voltage and starts ramping down the current according to a timer you set. I'd also mentioned that it can be a finicky thing to set.
Some time at the beginning of August, I was charging the car and I happened to stay in the garage for a while. Normally I charge it at night during off peak hours and I just let it finish on it's own. Well this time I hung around, cleaning up a bit and I saw the charger start flashing, warning that it was overheating. When I set the charger up last February, it was nice and cool in the garage. Not so much in August. To make a long story short, I realized I didn't want to shorten the life of the charger by overheating it every time I charged up, so I decided to turn down the current a bit. Turned up all the way and it sends 28 amps to the batteries. I decided I'd turn it down to 15. Problem solved.
Well, little did I know that I'd just introduced different problem. Apparently the point at which the charger changes from constant current to constant voltage is also dependent on the current out. I now have it set to start it's ramp down when the charger reaches 164.7 volts. But that was set when I was pushing 15 amps. Today was a beautiful cool day, so I decided to give it the full 28 amps and low and behold, the charger switched to constant voltage when the pack hit 163.4 volts, leaving the pack about 3.2 kWh's short of it's full capacity. I stood there for a second thinking "what the hell?"
I started experimenting with different charging schemes and trips around the neighborhood to draw some power from the pack. I left the potentiometer screw alone and just tried different charging currents and I found a curve. Send the batteries less current and the cutoff voltage is higher. Send it higher current and the cutoff voltage dropped. Here's what it looked like:
15 amps = 164.7 cutoff
20 amps = 164.4 cutoff
25 amps = 163.9 cutoff
28 amps = 163.4 cutoff
So back in August I simply dropped the current I was pushing through the charger to keep it from over heating, and the result was that the cut off voltage I had so carefully tuned for 28 amps went up. Suddenly it was too high and ultimately ended up pushing more current to the pack than it should have. I don't know the exact date this happened, so I don't know exactly how many times it happened. I think it's more than 3 but less than 7 times.
But there's a lesson to be learned here. Don't change any thing, be it a setting or a process, and expect everything else to remain the same. Always check and double check when making a change to a procedure or setting. Also, for any of you that have a Manzanita charger, or are thinking of getting one, remember this potential issue. Don't get caught by it. I don't consider this to be a major flaw in the charger, it's minor at worst. But forewarned is forearmed.