But now it's just a reflection of everyday life with a loose biking theme:
Molting after a cold ride in (it's warm now except for a cold snap)
Breakfast. Eggs, oat meal, pineapple.
Rappelling 95 feet down the back side of NI parking garage (shhh...)
On a telephone pole
(Male) cats
Mother Falcon throwing down
...
...
So how's the electric bike? I'm not riding it much at this point (10 times so far in 2011). Instead, I'm riding single speed pretty much full time. I sometimes ride it to work with the mountain bike in tow for our Tuesday night rides, but it mainly sits in the garage, waiting like a sad dog for its owner. I feel bad about it, but I'm just enjoying the exercise of a real bike.
So how's the electric bike? It's holding up quite well, with the exception of a couple of battery banks that are pretty much shot, but 13 out of 16 are still pulling strong.
One bank in particular (bank 9) is pretty much destroyed; banks 5 and 7 aren't that great, either:
Some reflections on the above date (which took days to collect):
A123's nominal rating is quite nominal... the best bank is only 85% of the rated capacity. If you dig through A123's spec sheets, they mention that their claimed life is taking the cell from 3.65 down to 2V, which is not good for the cells. Technically, you aren't supposed to bring them down below 2.9V or above 3.6, so A123's rating will result in premature damage to the cells, which is exactly what's happened:
Above we can see that during a full charge cycle, bank 9 (which has 56.2% of the energy storage of bank 15) was the first cell to empty (down to 2.2V) and the first cell to reach full charge (3.62V). Every single time the battery bank as a whole gets 'empty' it's the 12 cells in bank 9 that reach empty first (the other 180 cells in the other 15 banks still have quite a bit of charge left). Unfortunately, as soon as the first bank in series goes empty, the remaining energy stored in the other 15 banks is unusable without damaging the empty cell... without a battery management system. Right now, without a BMS, there's no way for the bike to tell me the first cell has reached empty; the only indication is that the voltage RAPIDLY drops from 50.0 to ~48.x over about a mile. Thus, by the time I know the first cell has died, I've already pulled it way below 0% and caused permanent cell damage... because bank 9 was originally the lowest energy storer, it always ran out first and always got pulled below 0% state of charge, and thus it's taken quite a beating, hence it stores only 56.2% of the charge of bank 15. Banks 5, 7, and 9 will need to be replaced.
I was curious what the useful voltage range of an A123 cell was, so I plugged a leftover cell into some hardware I borrowed from work and have been repeatedly charging and discharging it for a few days now... my plan is to keep doing this until the cell dies (maybe 200 days or so... A123s used within limits are supposed to last 2000 cycles). Here's the voltage ramp during a single charge/discharge cycle:
Starting on the left side, the cell is sitting at 3.6V and is fully charged. The programmable load (PXI-4130) then begins discharging at 2A (1C), and continues discharging for around an hour (because we're discharging at 1C). As soon as the cell hits 2.95V(1/3 of the way from the left), the 4130 holds and continues to discharge until the cell outputs less than 100mA@2.95V, at which point the 4130 immediately begins charging again at 2A (1C) for an hour, until the cell hits 3.6V(middle), at which point the cell is held at 3.6V until the current falls below 100mA... then the cycle starts over again (right side). I've been doing this for 3 days, so the cycle is repetitive. Here's the last 14 cycles:
Note that below 3.1V is only 5% of the total energy during discharge, so limiting discharge to to 3.1V would result in 5% less distance, but much longer battery life.
I'm logging all of this data to disk so I can see how the cell fails over its lifetime. Data thus far:
I compared the data to the temperature during charge, and there's a strong correlation between temperature and energy stored. I previously noted that the batteries don't heat up much during charging. The data shows that charging 192 cells at 1500W only generates 42W of heat (the cells chemically convert 97.2% of the applied power into useful storage, with only 2.8% wasted as heat).
I've been mentally conceiving a battery management system - the lack thereof is why 3 banks are shot - for around 6 months. The BMS I've come up with is a single unit for each battery bank that scales to any number of banks (for later use with electric cars). Each unit has a 2 bit bus that provide four state conditions: 00 is "bank less than 2.95V", 01 is "bank between 2.95 and 3.6V", 10 is "bank over 3.6V", and 11 is "fuck, something happened and we don't know what to do." 11 is connected to an AND gate and will override the microcontroller by disconnecting the cell from the stack no matter what.
I imagine you'd be shocked to know that this is how must engineering designs are initially conceived :). I've got much more detailed drawings with parts, etc, but the design isn't quite perfect, so it's conceptual. A single microcontroller with a muxed data I/O bus means you can add as many channels as you want without having to add a ton of pins to the controller... you just scan through. As each cell in the series bank reaches full, it's switched out by two SPDT solid state 60A relays... enough to handle 17kW@288V, but if you want more (I don't for now), you can simply drop in higher current relays. The microcontroller then tells the charger to lower down 3.6V for each removed cell.
On the flip side, as each cell is discharged below 2.95V, it is simply removed from the circuit, which simultaneously prevents damage to that bank and allows you to keep on riding until the last bank is depleted... because the motor is current controlled, the voltage drop causes the controller to demand more current... so the cells that are in series are taxed harder, but within spec.
Overall, I've got a bunch of work to do to implement a BMS... for one, all those cells I soldered together will need to be separated into 16 banks, with relays in between each one... I'm ok with this because I originally conceived this bike as a prototype for an electric car.
Yup.
