When you hook Li-Ion cells in parallel, they will equalize depending on their voltage differential. There has been some concern that the initial surge pulse of current may be a problem. Because of this concern, it is usually recommended to keep the voltage of the cells you are paralleling within 0.5 volts of each other.
To study this surge current pulse, I took a 26500 3200 mAh Li-Ion cell and over discharged it to a resting voltage of 2.097 volts. I then took a Sony US18650V 1600 mAh cell, which is rated to 10C, and fully charged it to 4.193 volts.
The 26500 cell has a 1kH impedance of 0.085 ohms and the 18650 has 0.025 ohms. A short piece of 12 gage stranded wire was used to complete the connection. It appears that the wire will add around 0.0008 ohms of resistance.
A Craftsman 82062 DC Amp Clamp Meter was used to measure the current. Voltage was monitored with a Fluke DMM.
I hooked the cells up in parallel, and observed a maximum of 6.5 amps for under a half of a second. The current quickly dropped to 3 amps over the next 2 seconds, then leveled out at 2.4 amps. It was still dropping at this point, but was not dropping as fast as it initially had been. I was watching the current, but when it leveled out at 2.4 amps, the voltage was around 3.104 volts. As the voltage rose, the current tapered off.
I disconnected the battery and the 18650 cell was at 4.012 volts, and the 26500 cell was at 3.121 volts. This gives me around a 1 volt differential. Hooking things back up in parallel again saw the current peak out at just over 3 amps and quickly settle down to around 2.4 amps and was steadily dropping.
With a 2 volt difference between the cells, I measured a surge current pulse of around 6.5 amps. When the difference was 1 volt, the current dropped to a little over 3 amps.
A similar test was run by others, and they observed that with a 1.4 volt difference, the surge pulse was 5 amps which quickly settled down to around 2.5 amps in roughly 10 seconds.
I measured that 93 mAh of capacity had been added to the empty cell. The duration of the test was about 2 minutes, so that makes the average equalization current about 2.79 amps.
It would be very unusual to have two individual cells with this much of a voltage spread, but if you were charging parallel packs, it is possible. Keep in mind that Li-Ion cells are OK with up to 10C pulses when they are empty, but they should not be charged at over 1C. In the example above, my 3200 mAh cell was subjected to a pulse of 6.5 amps, or roughly 2C, then the transfer rate dropped to around 1C and continued to drop from there.
Although I can see no problems with this, it is still recommended to make sure that you have less than 0.5 volts difference when hooking up in parallel.
Now, lets look at some theoretical numbers. Ohms law says E = IR, or I = E/R. The difference in voltage was 2 volts and the total resistance should have been 0.1108 ohms. This predicts that our initial surge current pulse should have been around 18 amps.
Either my meter is incapable of measuring the fast transient surge pulse, or there are other things involved that we are not taking into consideration. If I have another 0.1969 ohms of resistance in the system, the numbers balance. It is quite feasible that the total internal resistance under load is greater than the individual 1kH impedance measurements. On top of that, the voltage differential is quickly diminishing. If the resistance is correct, that would mean that the actual voltage difference was only 0.72 volts.
If we look at the worst case with a 0.5 volt difference, using the same set up, we end up with a theoretical initial surge current pulse of about 4.5 amps, however the actual measured value would be closer to 1.6 amps.
So, what does all this mean?
If you charge Li-Ion cells in series, there can be a problem if your cells become unbalanced. However, if you charge in series and utilized a balancer, the problem goes away.
If you donít have a balancer, you can charge your cells individually or in parallel. Individual charging has no special precautions, however when parallel charging, keep an eye on the voltages of the cells you are paralleling together to charge and it is best if the voltage difference is less than 0.5 volts. If you have a greater imbalance, start charging the lowest voltage cell first, then as it comes up to the value of the next cell, or cells, add them in. Parallel charging is very safe, but you need to be aware of the initial balancing currents involved when you first hook your cells up in parallel.
If you use a charging cradle, or jig, you can even parallel charge different capacity cells at the same time. You have to be aware of the initial surge pulse and the voltage differential, and also you need to make sure the charge current does not exceed 1C for the lowest capacity cell. This means that you could hook up 2 R-CR123 850 mAh cells, along with a 650 mAh 14500 cell, and throw in a 2200 mAh 18650 in your 4 slot charging cradle. Make sure that all of the voltages are within 0.5 volts, parallel all the cells, then hook up your charger. With those cells your charger would see 1 cell of 4550 mAh of capacity. Since 1C for your smallest cell (14500) is 650 mA, you would set your charger to charge at that rate. Assuming all of your cells were empty to start with, you would come back in around 4 hours and all of your cells would be fully charged, and equally balanced. If you group your cells with similar capacities, you can cut the charging time down to around 2 hours or less.
When parallel charging, set you charger cell count to 1 cell and adjust the charging current to a 1C rate for the lowest capacity cell.