Selfbuilt, do you have a good charger to test your eneloops?Your AA Lion results look much better compared to the eneloop ones.As far as i am concerned, good eneloops against good energizers won't have that much of difference. I think it is good to check your batteries! Else, it clearly seems that the circuit of the flashlight works better with higher voltage batteries!
Although it depends on the circuit, L91 does typically give much better performance than Eneloop. I am not surprised by these results - they are consistent with my other multi-cell testing on NiMH/L91.
But it is a good question to ask anyone who does a lot of testing. Damaged batteries would skew rechargeable results. Here's my regimen:
All my Eneloops are charged on a Maha C9000 charger - one of the best you'll find. I always start new cells on a break-in cycle,and I periodically do discharge/charge cycles to confirm capacity (plus additional break-in cycles at least one or twice a year to maintain optimal health).
New out of the package, most Eneloops show up as 2000-2200 mAh capacity on my charger (most ~2100-2150 mAh). Within a few discharge cycles, typical capacity is ~1900-2000 mAh, where it stays stably for quite some time. If the cells drop below ~1800mAh capacity on a discharge/charge cycle, I "retire" them from active use.
Before starting a run, all batteries are topped on the Maha charger to full capacity. They are pulled from the charger and immediately placed in the light to start the run.
Note also that I manually stop all my Eneloop runs shortly after the light falls out of regulation (or drops to ~25% of original output), and immediately begin a charge cycle on the Maha. This is to insure the long-term health of the batteries. NiMH (particularly LSD like Eneloop) are very susceptible to the damage from over-discharge. Although the main effect of over-discharge is damage to the LSD characteristics of the cell (and not total capacity), I don't want to chance it. This is why I won't do Lo-mode Eneloop runtimes - I can predict when the run will end, and I will damage the cells if I let fully discharge.
I rotate cells through testing, to make sure all my sets of Eneloops have consistent number of discharge/recharge cycles. All batteries are labelled, and I record which cell ran on which test. I also insure that cells from the same batch are consistently used in any given light. And again, because I periodically check capacity, I toss any cell that drops below ~90% rated capacity.
I thus believe my cells give you a fair representation of what a "typical" Eneloop in the hands of a regular user would be like. Probably better in fact, since I treat the cells with more care than most do.
:wave:
EDIT: I've just crunched some numbers from my max output 1xAA and 2xAA tests for comparison:
On 1xAA, 1xL91 very consistently gives ~50-55% extra runtime over 1xEneloop (as measured by time to 50%). On 2xAA, the results get a lot more variable - 2xL92 gives anywhere from ~40-80% more runtime on 2xEneloop. I presume the circuit design explains part of this variance, but there are also difference in how heavily driven the cells are on different lights. In any case, you will note that much greater runtime is possible when you increase the cell number. Thus, the 85% increase in runtime for 4xL91 over 4xEneloop on the M40A doesn't seem so unreasonable - especially when you consider the M40A is driven fairly hard (i.e. the absolute runtime on 4x here is typically less than that seen 1x or 2x on other lights).
