The "dangers" of paralleling 18650 cells of different capacity

Spurred by multiple posts I have seen regarding safety or danger of paralleling 18650 cells of different capacity, I am starting this thread.

There are many variables that could be thrown into this discussion. I would like to limit it to 18650 cells for the sake of simplicity. The cells to be connected in parallel must be assured to be at a reasonably close voltage to each other to consider that actually connecting the cells together, prior to connecting a load, is safe. One method is to charge them in a charger known to put the same voltage on all cells, then to confirm with a DVM that the cells are the same voltage prior to connecting them in parallel.

Now that we have two cells (or more) that are connected in parallel, we consider what happens during a discharge if they are not the same capacity. Again, for the sake of simplicity, the cells are not physically damaged, all have an ESR of less than X ohms. The ESR may or may not play into the discussion, I would argue initially not.

Next consider using these parallel cells with a load, (say a flashlight or lantern :wink: ) that demands less then 3A from the cells. I’m imagining that all 18650s of a reasonable ESR can provide somewhere close to 3A by themselves.

In this scenario, what danger, if any is there in using the cells this way. Cells could be from different manufacturers, different capacities, even different chemistry that take a full charge to 4.2V and are considered completely discharged at a voltage of 3.0V or lower.

If you believe this dangerous, please cite the sources of information which leads you to this conclusion.

It seems series is more dangerous then parallel

The article sees problems when two parallel cells are charged together, the OP is about charging them independent but discharge them in parallel.

Good addition Jos, i must have missed that part

post 2 is an excellent example of providing a reference. the one above discusses the possible degradation of cell life when parallel cells are used at a “extremely high C rate”.

I would argue for the test case listed in the OP this is not terribly applicable, and even if it was there is no talk of specific safety issues with the cells. I guess one needs to define “extremely high C rate”, in this case I am assuming 3A is not in that category for even a single 18650 cell. If it is for the cell one chooses, then I would agree with the danger.

Further, the article is about LiFePo cells, and the measured problems are largely because LiFePo cell have a very flat discharge curve, it is suggested that li-ion cells with a more sloped discharge curve suffer much less from this problem.

Indeed misleading, since the title specifically says Li-ion

Would there not still be an inrush into cells that may or may not be the same or close internal resistance? Parallel would not be as bad as serial by any means, but cells would still try to equalize as they are all common at that point. I charge my 4C parallel cells in the same 4 bay charger at the same charge rate and wait for each cell to reach 100% before removing it from the charger.

Well, the problem with cells in parallel is that they are linked together. So if they are at different states of charge, they are directly wired together to fix that as fast as possible.
In theory, cells of different capacity will drain at an uneven rate to keep the state of charge equal. I don’t know how well this holds up in practice - I’d be especially curious to compare high drain cells with very low internal resistance that are in parallel with low-drain cells. Say, a VTC5A with a 35E. Or maybe a 30T with a 50T.
In my multi-cell lights, I only use married cells. My S6 and MF02 actually share the same set of 4 married cells. It makes me feel better so I do it that way.

They would presumably only have to “equalize” if their voltages were different at the start.

Correct, but by keeping the cells in a family and charging them at the same time and rate the voltage differential is very small. I do not think one would ever have a problem with decent cells within a small handful of tenths of volts apart. Now it you start mixing cells or have large voltage spreads, I am not so sure you would not have a problem. Also, if mixing good low resistance cells with higher resistance cells, the question becomes (in my mind) am I charging the high resistance cells above the manufacturers charge rate?

The issues pointed out in that article are that a single cell may end up absorbing most of the charge current at the ends of the charge curve. Seems like charging the light with a single cell would be worse than having multiple mismached cells installed.

Series connections of different capacities is definitely not recommended. It is really important when mixing different cell age/condition as under loads it can unnecessarily stress the stronger cells to the point of over discharge or overheating and potential failure/venting or fires. That’s why you have some kind of protection and monitoring board for series li-ions.

Parallel connections are pretty benign since the total capacity of the cells as a whole is dependent on the lower capacity cell and the cells don’t need to be individually monitored.