Acceptable drift between cells


After a ride, most of my 10 cells battery stay within 0.05v from each other, but 2 cells differs by almost 0.2v. I was wondering, what you guys thinks is an acceptable drift between cells before starting to worry? My battery is Lithium Ion, 10s8p and the cells are 18650 Samsung 26j.

Thanks, Ishay

what cells and how many are there total? just ten? lipo? ion? iron? if you have a couple in parallel be nice to know what is really going on in the p group and maybe one is ruining it. bummer having cells in p stuck together so you cant know which is bum

1 Like

Depends on how deeply you’re draining them I wouldn’t consider .2V drift to be that bad if you have cells at 3.4V and 3.6V, if you are at 3.7V on the top and 3.5V at the bottom would be more of a concern since at 3.7-3.8V you still have a large percentage of the battery power available but at 3.6 and lower it is less than half of the full charge in the cell. Basically you want to be able to pull the same mAh from all the cells that are in series so that no given battery in series is ever below/past the voltage cliff on discharge.

1 Like

Edited, Thanks. Yes, It’s a real pain to check the batteries in the P group especially after isolating the cells like crazy with a ton of electric tape and hot glue.

After draining the battery to 87% I had a drift of 0.14v. The highest cell was at 4.05 and the lowest one at 3.91. I get to 3.6 volts per cell only when squeezing the throttle because this battery is 21ah and the rage is crazy.

I would start to get concerned actually. Since you have 8paralell cells it sounds like you most likely have one bad cell in each of those 2 groups. You have a voltage drop of around 12-15% or 1/8. Not scientific proving but its an indication.

I would charge them one by one one and se if you se the same after som riding. Just. To make sure that it isn’t the bms. You should at least investigate the pack in my opinion. Again this is what I think, but in the end its your call :slight_smile:


I’m afraid it’s not possible to charge them one by one. What the max drift(estimation) before the battery is dead? or until it goes puff on me.

Because you do have 8p it means that the problem is not so visible. Say you have lost one cell that means 1/8 of capacity for that row. Around 12.5%. As you can see it already sounds the way your pack is behaving. I’d you have 10s2p that would be 50% and directly dangerous. This problem will grow with time since those two rows will have to work harder so it won’t gett better. I personally would not sleep well without fixing that :(. Say Tesla car has over 60 cells in parallel that is not a big problem,but in your case it could be.

It is always possible to charge the rows one by one. Takes some skilles but it is possible. Or you could check if all the cells are getting fully charged after a long charging time.

1 Like

And please dont take only my word, there are more experienced guys who might have better clue than me. And maby prove my suspocion wrong.

1 Like

I get your point. Im now charging the battery and waiting for the BMS to kick in to see if it helps.

Charing the pack to full might take time :). Deppening on the bms you have. But you would be able to see if it kicks inn.

Would be interested to know what cells you’re using. I’ve got a 10S9P pack that I’ve been using daily for several months now without balancing and they’re all still within less than 0.01V of each other

1 Like

Really? Sounds weird that they don’t drift at all without balancing. I’m using Samsung 26j.

Normally good Li-Ion cells don’t drift much. 0.2 is definitely too much and I do agree that it seems you lost a cell of your pack. And why do you use 26J. Did you harvest them from laptop batteries?


That’s what I was afraid of. Not from a laptop, but I got the cells for almost free.

I’m using Bestech BMS and I think it does not balalnce the cells anymore. Does it have a way to know one of the cells in the P group is dead?

No the BMS has no way to measure that all the cells in parallel are at the same voltage. You’d have to disassemble the p groups and test the IR (internal resistance) or esr(equivalent seried resistance) to determine which cells are bad.

Also just to note I was talking numbers for lipos for liion the whole voltage scale is slid down a bit but same idea holds true.

1 Like


Cells should all drift at a similar rate over their lifespan. Your bms is more there to monitor cells than it is to “balance charge”. I have yet to have any balancing turn on.


lol my cells are not balanced at all compare to yours. btw, what is this app calls?

Xiao xiang bms Capture%20_2018-08-04-09-36-20

It’s limited to those Chinese Bluetooth BMS’s from what I understand. The pack is fairly new still maybe a dozen full cycles so I’ll report back once I’ve got more cycles on it.

The less drift the better, when your pack is fully charged leave it at least 20minutes to balance itself out, each time