Thread resurrection, but I'm hoping to get some help.<snip>
Well, not sure I can help, but I enjoyed reading this "resurrected" thread. Since I didn't post any of the old posts, I cannot say for certain what they mean, but I can tell you what I "think" (for what that's worth). I hope that others who are more knowledgeable will also chime in.
You likely already know this, but as I understand it, the issue with matching cells is closely related to usage. Serious discussion of matched battery packs seems to center on the RC racing world. When racing a small electric car, the ability of your battery pack to drain every possible drop of energy can mean a few extra seconds of runtime which can make the difference between winning or losing a race.
Therefore, in the RC racing world, the closer your cells are matched the better. Serious racers are looking for the highest possible capacity cells that have (if possible) 0% capacity difference between them. As I understand it, having one cell of lower capacity than the others causes the other remaining cells to attempt to recharge the dead cell, and this can both damage cells and greatly reduce performance.
When we're talking about using % to set matched capacity limits, are we talking about using % from nominal? Or are we talking about % from a mean of observed capacities from a set you're looking to match, or some other calculation?<snip>
Since the "ultimate" is having 0% difference in capacity, I am not at all certain that there is any agreed on method to arriving at "% to set matched capacity limits." If it were me, I think I would take the difference between the highest capacity cell and the lowest capacity cell and divide by the lowest capacity.
For example, I broke in 8 of my older Eneloops on my Maha C9000 and recorded the following for each individual cell (I set to 1900 mAh):
1) 1918
2) 1911
3) 1905
4) 1913
5) 1946
6) 1942
7) 1957
8) 1965
Assuming you were going use the first four cells as a matched set: (1918-1905)/1905 = 0.68%
2% from 1900 mAh yields 38 mAh and provides a range of 1862 - 1938 mAh. So cells #1-4 I could mix in use as needed. Cells #5-8 are over the 2% nominal limit, so I couldn't mix them with cells #1-4, yet they have a %CV of 0.54 and would seem to me fine to use in series. Thus my question: If you have cells that are over or under 2% of the nominal, how do you identify outliers from a batch? How am I supposed to calculate the 2-10% match limit?<snip>
Outside the RC racing world, I think the issue is how best to group the cells that you already have, in order to produce the best possible set that you can. (True "outliers" both higher capacity and lower capacity can be relegated to single cell application.)
But, the need for well or "better" matched sets, depends a lot on how you are going to be using those sets. As I understand it, having better sets mean that you will do less damage to the cells should you ever run them "completely" dead. Therefore, if you tend to change your cells before they are "completely" dead, I think having extremely closely matched cells is much less of an issue. If you know you have (or might have) a very poorly matched set in use, try not to run them down too far before recharging.
But, personally, I figure I may as well try to select the best possible sets, when I can. Obviously, the more cells you have to chose from the better. The last time I purchased Eneloop AAAs I purchased 5 x 4 packs, so that I had 20 cells to work with.
It takes "a bit" of time (understatement), but I use my Maha MH-C9000 to run a discharge, break-in, discharge, and analysis. Then, I use the final "analyze" capacity readings to determine my groups. It can be rather mind numbing to look at all the capacity figures and try to determine "which one is most like the other one." So, since I enter all my data into an Excel spreadsheet anyway, I have come up with the following method to "enhance" (make more obvious) the slight capacity differences found in most Eneloop cells:
Using the capacities of your 8 cells, the green area shows the "enhanced" differences. Clearly, given these 8 cells, if your desire was to create two matched sets of 4 cells each, you would want to use the first 4 cells (all negative %) as one group, and the last 4 cells (all positive %) as the other group.
The more cells you have and the closer they are in capacity, the harder and harder it becomes to "decide" which cells to group together. So, I find the above method works well for me. It is also quite simple (if you use Excel). The formulas used in calculation are shown on the right. The idea is to subtract the measured capacity from the rated capacity, to arrive at the "difference" for each cell. Then you average those "difference" values, and you take the standard deviation of those "difference" values. Finally, you take the "difference" value for each cell, subtract the standard deviation, and divide by the average. The resulting "enhanced" difference percentage does not "care" at all about the performance of the cell in relation to its rated capacity, it only "looks at" the differences between cells.
Don't know if any of this has helped you or not. But, I enjoyed writing it. Hopefully, others will join in to both critique my methods, and offer better advice to you.
