Building custom battery packs from salvaged cells?

I have a surplus of used laptop batteries. Disassembly yielded 18 usable cells, 8 are 2,550 mah, the rest are 2,400 mah. All lithium-ion.

I'd like to build a custom battery pack or two to power a prototype a DIY bicycle light.

In the battery packs themselves, the cells were arranged in sets of two or three cells in parallel. (2 parallel cells for the 6 cell pack, 3 parallel cells for the 9 cell pack.) There were no fuses between the cells. The set of cells then were connected to the laptop circuit board, which presumably was the PCM. (The circuit looks the same as a 11.1V PCM, but there were three cells in parallel instead of one.)

I'm thinking about doing something similar in designing my battery pack. Instead of using one cell, I was going to use three in parallel. There are 7.4V CPMs that are rated for 3.5A, using two cells. If I used 6 cells instead, I'd get roughly 50 watt-hours in the battery. That's 7.4V out, enough that most drivers would be happy. I could draw 3A for a xm-l emitter, and get several hours of runtime.

So, question #1: Is this a good design? I can't see it being any less safe than a laptop battery, which also has three cells in parallel. A laptop has to be drawing a similar amount of current from it's battery.

Question #2: Would I have any problems with using an off the shelf universal smart charger rated for 7.2 or 7.4V li-ion batteries? I can't think of a reason.

Question #3: Would a different configuration be more optimal? Say 3.7V instead of 7.4V. Or perhaps 11.V. 2 or 4 cells in parallel instead of 3? I'd preferrably like to use a maximum of 8 cells per pack, just because I'd have enough to build two battery packs then.

I think you need a hobby charger or some method to test the actual capacity and the discharge rates of the cells before even thinking about doing this. I always number the pairs of batteries as I remove them and test them as different pairs can and do vary even within the same pack. If you were to get cells with significantly different discharge rates, at the very least it would make for a poorly performing system, at the worst you might have created something you don't want anywhere near anything living.

My first step on weeding out the bad cells was a quick voltage test and tossing out the odd ones.

I'll charge them in a smart charger and do a DC load test then. That should give me impedance. I'll toss any of the cells from the same pack that are unusually low. I haven't done this yet because I'd rather keep an eye on the charger while charging old cells for the first time.

I thought about building a testing rig using a single-battery CPM, and just letting each fully charged cell run through a discharge cycle on a dummy load. Then I could calculate the real capacity of the used cells in mah.

As far as I can tell, if the CPM doesn't allow more than the rating of an individual cell, I should be okay from a safety perspective even if a cell goes bad. But the custom battery pack performance will suffer.

Aloha and welcome to BLF dasunt!

I'm debating changing the cell configuration slightly.

Using 2 parallel 7.4 PCMs, then having 4 cells per PCM (2 parallel cells x 2). That should allow me to draw a few amps without a problem, and increase battery life.

I'll know more once the charger comes and I test the cells I have. But 65 watt-hours is a nice size for a pack.