The BMS could be working exactly as designed but still not be capable of balancing the cells except over a very long period of time. For example

 

Take a 36v 15AH battery pack with this commonly used bms https://vruzend.co.uk/products/10s-36v-15a-lithium-battery-bms-balance-function-18650-li-ion-uk-seller-stock it has a balancing current of just 40ma.

 

Now imagine said battery pack has been left in a damp garage for 6 months and corrosion means there's a higher self discharge rate for one of the parallel groups, or maybe through bad luck one of the cells happens to have a higher self discharge rate from the factory. So one group has only 1ah left whilst the rest has 14ah.

 

The owner then puts it on the charger and it finishes quickly because it only has most groups have only 1ah to charge. How long would the owner have to leave the battery on charge to have a fully balanced pack?

 

The BMS needs to add 13AH and can only balance at 40ma, so 13000/40 =325 hours or 13.5 days to balance the pack.

I'd be surprised if the BMS was able to balance cells that were that far out of balance.

I'm not leaving the charger connected long enough for the BMS to balance

Your pack construction is a little unconventional with uneven parallel strings of cells. I'm wondering if this is having an adverse effect on the even discharge/charge cycles. I only say this because if we assume all cells are good and you have no dead/dying cells in your pack, then inbalance is unlikely to happen. What are the voltage readings on each parallel string?

Electricaly the parallel strings are even, there are 5 cells in each parallel group but physicaly there are 10 cells in a couple of the cell holders.

 

I only have voltage measurements from taking the battery off the charger, when I balance charge the cells I charge the pack to 42.0V, then wiat for the current to drop to about 1A, that way I don't have to spend to long charging individual cell groups as this can take up a whole weekend.

 

The voltage of the 'good' cells would be 42.0V (42.2 ish if the BMS has tripped) and the 'not so good cells' 4.05 - 4.15V. When I balance charge them I try to get the charge current the same on all cells, the 'good' cells might be drawing a few mA, if I can get the other cells down to about 250mA the bms appears to be happy, by this time I've run out of weekend so can't get the charge current any lower which may be something else causing the issue.

, there are 5 cells in each parallel group

Ignore everything i said above, I was looking at your pack as 4P and that you had cable/braid linking the uneven strings.:rolleyes:It's no fun losing your w/e to battery balancing, done it a few times myself.

I know what you mean, I often spend the entire weekend maintaing / repairing my ebike just so I can get to work on a Monday morning