Eeeewww, Battery Death

OK so here's my problem. The car came with some batteries that were not maybe real well known for reliability from a company called HAZE in China and were nearly a year old when I got the car plus who knows how they were treated while EcoMotion (the original dealer) had them.
There was also this one single DEKA Intimidator battery from the East Penn Co. in the US. Added to bring the pack voltage up to 84V. Supposed to be top quality. Same 100 Amp-Hour rating as the HAZE and fairly expensive, like $220 or $240 each. Guess which one dies? Yup, the DEKA lasted about thirty cycles - despite - the high bucks BattEQ battery management system.
So, we get ahold of the dealer, who turns us over (quite appropriately) to the battery maufacturer's rep, who eventually says "Yeah, we'll get you a new one." and after a week or two they pick up the failed battery and a week or two later (after some nudging) they drop off a new one. Total elapsed time is like TWO MONTHS.
I wonder how the battery guy would feel if his new car developed a gas tank leak and the gas tank supplier took two months to get him parts (which he would then have to install himself...) I don't really know him, but I bet he'd be really pissed-off to be without his new car for two months.
Fortunately I had a spare battery.

In the mean time the BattEQ system gets recalled and I get involved in the design of the VoltBlocher boards as noted earlier. I didn't have enough boards for the whole pack. Battery #5 was behaving really well on it's own so I didn't put a board on it for like two weeks after all the others had their's. Guess maybe it wasn't doing so well after all because it started to fail about two weeks later. Right about the time the new DEKA arrived. What a stroke of luck!
Just to make sure it's OK (and to break it in a little) I ran a couple cycles of charge test. Discharge at 5 Amps until the voltage drops to 10.8 and then charge up (Bulk charge phase to 14.5V - etc.) Funny thing happens: at 5 Amps (actually 4.93) disharge rate, a 100 Amp-Hour battery should last 20 hours. The most this one will do is 16 hours. 20% low. Hmmm.
Maybe it takes a few more charge cycles to come up to full capacity? We'll see. Swap it into the car in battery slot #5. A fully charged pack should be capable of going over 20 miles. This pack would only do about 18 as the original #5 was a little weak by the time I tried this. Still with a worst case commute of 12.8 miles we should be fine, eh?
Not for long. The new DEKA seems almost fine for a while. It was 0.3V lower than the others pretty much all the time. Seems a little odd, but I have one other that runs a little low and it's holding up fine. I set the VoltBlocher to limit it to 14.4-14.5V during the high current bulk charge phase and a little higher during the later (generally lower current) phases.
All goes well for about 20 charge cycles. I limited the initial miles/charge as the battery got broken in but it was running the longer commute (12.8) within a week. About two weeks later I have a 'short' commute of 7 miles (with charging a the other end) which I had been doing for a couple days at that point. I ended up going an extra 2 miles that day and the new battery dropped way off in mile 8. The next day I fired up ye olde data-collection system and captured the voltage spikes down as the new battery crapped out at just about the same place. By around the 26th or 28th cycle on the new battery it was effectively dead.
Me and the battery dude are going to have another dance. You watch, they'll try to blame it on the charging system. I OTOH have actual data. We'll see. Meanwhile I'm cycling one of the other 2+ year old batteries to see if I can swap it in while we wait. Again.
Moved the VoltBlochers further forward between the seats so I can keep a better eye on things. That part seems to be working good.

Yea! Pictures!

Da car, nestled into it's "Motorcycle Parking" spot at work. You can't see it but the charging outlet is on the post behind the car. Thanks again to my workplace for providing the charging facilities.

So, the warrantied DEKA 'Intimidator' battery was delivered about ten days ago (just over two months after the first one failed) and I stuffed it in to replace battery#5 that was failing pretty rapidly. I discharged and recharged 'cycling' the new battery several times to get it broken in. Yes you do have to do that, tossing a new horse into the team and running them hard can cause stumbling and crashing problems...
I pulled off all the Volt Blocher boards before swapping in the battery and did a little testing and tweaking and then remounted them in a more managable arrangement. I got a few resistors swapped around and now all the boards operate within about a tenth of a volt of each other. I put the boards that operate a little low on the batteries that operate a little low, makes everybody happier.

"What, the batteries aren't balanced? Heresy!! "
There are batteries of three different ages and two different manufacturers in the pack. Of COURSE they aren't 'balanced' Don't be silly.
In fact I think the battery balancer that was on the pack (until it was recalled by the manufacturer) was part of the reason for some of my battery failures. Even when 'new' both the original and replacement DEKA batteries seem to run 2-3 tenth's of a volt lower than the HAZE units (which have a couple tenths of variation between themselves) and having something that's trying to bring them all to the same voltage seems to cause the units that run naturally lower to overcharge just enough to vent a bit. The venting causes a reduction in capacity, which increases the tendency to overcharge, and vent, and 20-30 cycles later you have a failing battery.
The traditional fix for this problem is to throw money at it. Just replace the whole pack. Then they'll be balanced...if you're lucky. Being cheap, I don't much like the $1500 solution.
My fix is to have the Voltage/Overcharge limiting boards set so they match the battery they're attached to. The 0.3V difference between the switch settings on this version of the VoltBlocher boards turns out to be just about right. (Brian, take note!)
As a result, when it's nearing the end of the 'bulk charge' part of the charging cycle with the pack sitting at around 60deg.F I'll see voltages of 14.7, 14.3, 14.5, 14.5, 14.3, 14.5, 14.5
#2 is a two year old HAZE and #5 is the new DEKA. If I take the boards off, some of the voltages will shoot up to 15.0 or 15.2 wich will kill them in short order.
I'm not sure what the deal is with #1, it always runs a bit high, though part of that is maybe measurement error. Anyway, we're now at 2000 miles...and with only four (total) battery failures...Huzzah!!