just for information gathering sake
what was the rate of charge you were using?
what are the cell brands/type?
what is the rate of discharge?
and your normal method for charging them is in series?
were they locked into a Pack or could you have seperated them for single charging without much trouble?
Charging:
i only know One way to series charge ni-mhy, and that is slow, for two reasons, termination detection on a series pack is very difficult, and if you get termination properly, you still need to have a top off charge to "balance" out the battereis that are not fully charged, when the first 1 or 2 here and there Vdrop (so the computer can see a reason to terminate).
or of course a tapped pack for balanced charging.
IF you have a smart charger you MUST go fast enough for it to determine Vdrop and terminate, unless it has temperature termination , if you sllllooooowwww charge , you must go Slow enough that it doesnt go past overcharge specs.
my series Dcell ni-mhy lights and power tools, use dumb slow charging, and work for years and hundreds of cycles, and are still working, with stupid stuff like $15 wall warts, and i charge right in some lights in series. many of the consumer ni-cd and ni-mhy charge in device (series) chargers do the same thing, low slow and dum, wait forever for it to charge.
Discharging:
then anytime a single cell in the seires pack is fully depleated, the evil reverse charge comes to play, and damages the already low capacity cell, making it even lower capacity than it was before, which subjects it to even more reverse charge again on full depeletion, it keeps getting worse and worse.
so when this configuration starts drooping in output , you must stop the load (shut off the light) because one cell has depeated, and is about to head to reverse charge.
so you got 2 factors here that are possible for your problem, the charge and the discharge. and the worst is probably the discharge , unless you recharge always before full depleation, or have cutoff curcuits.
Matching cells:
ok 3 problems
because some cells are not very reliable, are not intially balanced in capacity, get flakey over time, or have more self discharge than the items they are teemed up with, and dont handle reverse charge as well.
i dont have a Triton Jr, some because of cost, some because they are a bit to fast for me (mabey i am gettin old) :tinfoil:
the triton Sr , and the newest triton 2 thing, have improved thier alogrythms, and detect cutoff, taper, and top off, much better than earlier models (according to specs and data and user data).
Prefer single channel single cell charging:
IF you can seperate them and singularly charge them for less problems. single cell charging you can go fast without seires charging issues. or dumb slow charge them in series overnight. I would say parellel charge them too but that can be a problem Too, unless you voltage high terminate like li-ion would, no ni-mhy charger works that way.
Cutoff in device:
if you can stuff a low cutoff curcuit in the light that would terminate anytime the batteries are like this 1.1v 1.1v 1.1v .6v or similar, (some math required) the light would stop before destroying the cells. with extreeme loads the cutoff would have to be figured for the load.
Again, i dont have that charger, just know of the ramifications of both series and parellel charging, and series discharging.
it is possible that you could singulary cycle the cells, and recover a few usable ones out of it, but if you remake the SERIES assembly or team, you would prefer the batteries to be matched, matched in age capacity and acting in similarity.