LiFe might be different, but with the sketchy quality of LiPo from Hobby King, I test every pack on a cell level before building into a parallel group.
I do a few things specifically. I first do a light cycle on the pack, maybe charge up to 70% (easily set on the Hyperion 1420) and discharge to 50%. I'm looking for gross errors with cell voltages.
If that looks good I do a 95% charge, resulting in something like 4.15 cell voltages, followed by a disharge to 30%. Here, I'm looking at cell spread. Does one cell drop faster than the others? Does one hit the cliff before 20%?
Next I'll discharge to the cliff. Here I'm closely watching the cell voltages and noting if one drops off way early. I.e. is one dropping to 3.5 while the others are at 3.7. This would have a significant impact on usable pack capacity.
Then I'll recharge back to 95%. This time I'm watching if one cell charges much faster than the others. If a cell that dropped in voltage early is now charging faster than the others, its a good indication of a weak cell.
Once I get them charged back up, I let the Hyperion balance the pack. Then the pack gets labeled, all cell voltages get noted, and the pack gets set aside for a couple of days. After that I'll check the voltages again to look for 'leakers.'
Any packs with leakers or signicantly weak cells will get set aside (then most likely mailed to Amberwolf:he loves separating the good cells from the weak ones!
)
Then, I parallel up the mains for the packs going into parallel groups. But I don't parallel the balance leads at this point. I'll let this group go through several use cycles, each time keeping a close eye on each cell after discharge and charge. I still charge each pack solo.
Finally after 5 - 10 cycles, once the cells are proven to be good, I charge and balance each pack. Then I parallel the group at the mains and the balance taps and consider it a solid group.