One thing I've noticed with a123 cells is that even on the same charger, two cells might end up with different voltages, right off the charger. I've seen variances from 3.58 to 3.72 with 8 cells that I charged individually with a TP-1010C set for one cell. If you put a slight load on these, however, even for just a few seconds, the "resting" voltage for each cell settled in at a voltage of about 3.53-3.55V, so it doesn't mean much to check variances right off the charger(s). Healthy cells will be "full" at somewhere around 3.55V resting voltage. I've seen some cells that have been stressed that will lose some capacity and end up only having a resting voltage (i.e. -- not right off the charger, but after a few seconds of being under a load...) of around 3.42-3.45V. These cells still work fine, but they won't have a full 2.3Ah of usable capacity, maybe only 2.0-2.1 Ah.
The VoltPhreaks chargers I have all seem to have slightly different cutoff voltages that vary from about 3.80 to 3.85V. After a ride yesterday that consumed about 1.7Ah, I recharged both packs. One ended up with voltages per cell of 3.80V, 3.83V, 3.81V, 3.80V, 3.86V, 3.80V, 3.83V, 3.84V, 3.83V and 3.81V. The other pack had similar variances.
Once you put any sort of load on the cells, however the voltages will drop down to about 3.5-3.6V. This is illustrated in the graphs below, which are from the short test ride yesterday. The data comes from a new version of the EagleTree MicroPower eLogger that I have been testing for them, and there is more info in this thread:
http://endless-sphere.com/forums/viewtopic.php?t=2599. Anyway, here's the charts:
In the first one, you can see that the pack voltage starts out at around 73V, or about 3.65V per cell, but this drops quickly when a bit of a load is applied. In the second graph, which is simply a zoomed in section of the first, you can see that this initial drop settles down to about 66V, or about 3.3V per cell, where it remains, even after multiple 80A+ peak loads. This is why these a123 cells are rated as having a "nominal" voltage of 3.3V per cell.
Anyway, it has been my experience that the largest variance between cells will be when they are fresh off the charger. You should always check the balance between cells after they have been under load a bit.
Another thing I've noticed about a123 cells is that if you do have a cell that has been stressed to the point it has lost some capacity, it will always have a much lower voltage right off the charger. I've got a few that will only be at 3.36V, some at 3.48V and some at 3.40V. This is the case even if they are paralleled with 3-4 good cells. The bad one will always pull the good ones down to this level. They still work, but the capacity is reduced to about 1.9-2.0Ah.
-- Gary