bypassing faulty cells

I've got a 36v/16ah battery pack with 40 individual cells and a 10 lead bms.
i found 1 group of 4 cells is faulty. (discharges in seconds under load) so was thinking if i could somehow bypass those 4 cells i would still have a perfect "32v" battery right?
Well i was thinking about making a switch to bypass when in use and switch the faulty cells back together when charging.
But then i was thinking about the lvc and bms.. What would happen if i cut one of the balance leads and put load on the pack?

Simply putting a short around the dead group could damage the BMS itself, depending on how it's designed, so unless you're sure it's ok I wouldn't do it until you can be sure.

Cutting the balance lead completely, but leaving the cell in place and unbypassed, will probably not work for one or more reasons. The first is that the BMS may respond to a cut wire by shutting down, if it has cell-level LVC. If it doesn't, then the pack LVC (if it has one) may trigger, as the group dies quickly. If neither of those happens, it's possible for the bad group to actually reverse in voltage, as current flows thru it under load in reverse of charging, and the results of that are unknown--dramatic failure is unlikely but possible, especiallly with cylindrical cells if their vents are blocked (or if they don't have any vents).


Some BMSs can be used with various numbers of cells, simply by not hooking anything up to the top cell connection(s) on the BMS board itself. Some them have jumpers or pads on the board to set it to differnet numbers of cells; you'd just move that jumper to the correct point as marked, if so.

To do this you'd need to disconnect all the wires from the cell groups "above" (more positive than) the dead group. Then reconnect them one "down" from where they were, so that you now have all the cells except the dead group connected to the BMS, and one unconnected wire from the "top" (most positive) balance wire of teh BMS.

This may or may not work. It should not damage anything, though.


Another way you can fix it (but reducing capacity and C-rate, by 25%) is to take a cell from each of three other groups, and use them to replace three of the bad cells, leaving the fourth spot unoccupied, and completely removing all of the bad ones.

FWIW, it is quite likely that only one of the dead group is really dead, and the others may be mostly ok (though reduced in capacity from their excursion and stay at low voltage). You can disconnect them from each other and single-cell charge them to full, then load them and see what capacity/voltage sag they have. Any that are really bad, just leave out. It's probalby only one.

Unfortunatley if you only have 4P, then even one removed takes out 1/4 of your pack capacity, and stresses the other cells a lot more than they would be otherwise, under loads.


So removing the whole group and shifting around the BMS wires (if this will work on yours) is your best bet, other than replacing the bad cell(s).

I wasn't planning on just short around the group but first cutting the connections and then put a switch between them for charging as i only have an 42v charger.
Shifting the balance leads isnt a bad idea but then i need to somehow put the faulty group at the 10th balance lead to able to switch it back on when charging.
I just pulled out the balance leads and the bms reads 36v while the end of the battery itself measures 38.7. Does this mean that the lvc tripped?
I cant test it under load until tomorrow.

Actually you don't want to put the bad group in when charging, either. You'd want instead to adjust the charger's output voltage down to account for the loss of that group.


Otherwise it may never properly charge all of the rest of the battery, *or* you may end up overcharging the bad group, eventually causing unpredictable behavior or failure, *or* it might take a really long time to charge up the pack, becuase it will stop every time it thinks the bad group is fully charged. Depends on how you work around it.

amberwolf said:

Actually you don't want to put the bad group in when charging, either. You'd want instead to adjust the charger's output voltage down to account for the loss of that group.


Otherwise it may never properly charge all of the rest of the battery, *or* you may end up overcharging the bad group, eventually causing unpredictable behavior or failure, *or* it might take a really long time to charge up the pack, becuase it will stop every time it thinks the bad group is fully charged. Depends on how you work around it.

Click to expand...

ballance charging means it ballances the charging right?
So if one group is full it stops charging that group right?
Or are thous balance lead for measuring purpose only?
And i got one of those "brick" chargers so their isn't much to adjust.

Since you really dont know how to work around those dead cells, or what the result may be when you do your .. "modification"..
..why dont you do the logical/sensible thing and simply replace the dead cells ??

Enpro said:

ballance charging means it ballances the charging right?

Click to expand...

Balance charging means that (usually) it slightly overcharges a cell or cells, up to the HVC limit it's built for, and then it cuts off voltage from the charger, so current stops flowing. Then it uses (usually) resistors to drain off the cell(s) that are over the actual full charge voltage, until they get down to that voltage. Then it restarts the charging process once there aren't any cells over that limit, if the total pack voltage is still not "full" voltage.

So if one group is full it stops charging that group right?

Click to expand...

Usually it stops charging *all* groups, as they are charged in series. So if you have one group that is already fully charged (or appears to be because it has high resistance, so it's voltage climbs while current is flowing thru it), then the BMS will stop the charger very early, and wait for that group to drain, then reengage the charger, and repeat this process until the rest of the groups catch up. This can take a VERY long time, hours, days or even weeks, if the imbalance is very bad and the BMS uses a very small balancing current, which is usually the case, like with my Vpower/CammyCC pack's BMS, as it has very tiny SMT resistors that can't handle much in the way of drainage/balancing currents.

If the one group is left "full", and the rest are drained pretty far during a ride, then reconnecting that full one and then charging will take quite a while to fully balance the pack, and may even take that long just to do most of the charging, not even getting them to the real balancing stage. That's because it "thinks" the bad group is already fully charged, and it doesn't want to overcharge it by pouring charging current thru it, but that is what is needed to charge the rest of the groups that are in series with it.

Or are thous balance lead for measuring purpose only?

Click to expand...

I'm not quite sure what you mean, but the balancing leads on a BMS are typically used both for the BMS to keep track of individual group voltages, monitoring LVC and HVC, and also to bleed off voltage once a group does reach HVC.

And i got one of those "brick" chargers so their isn't much to adjust.

Click to expand...

It might not have adjustment pots in it, but it does have a way of altering the voltage output, if you can find the right point in there to change resistance on, most likely. If you look thru the Fechter Mini Meanwell Current LImiter thread, you'll see discussion about finding the 2.5V reference point to hook the limiter to. That same point can be found in most of these SMPS type chargers, which yours probably has, too. If you can find it, you might be able to change the voltage at that point by one method or another (changing the resistance of the voltage divider is probably easiest), and thus adjust the final output voltage. It'll take experimentation and probing around in there with a DMM, but it's possible. Take a look at that thread first.

Hillhater said:

Since you really dont know how to work around those dead cells, or what the result may be when you do your .. "modification"..
..why dont you do the logical/sensible thing and simply replace the dead cells ??

Click to expand...

Because their arnt any cells like these available in the Netherlands + i have no idea what kind of cells these are. They could be lifepo4 or linmo2.
Its eighter work around them or trough the pack away.

And amberwolf, thx for the info so far.
I have a look into that Fechter Mini Meanwell Current LImiter thread.
If that doesnt work i just drain the faulty group everytime i need to charge the pack..

It could be an option to make a 24v pack from the 36v. You'd need a new 24v bms and charger. Especiallly appealing if you have a lawn to mow, where some electric mowers can use 24v.

Given the state of bike batteries now, maybe time to just get a new one? Or figure out what chemistry you have now, and get something you can replace the cells with.

If you have the budget, a new pack would sure be simpler. Especially if you got your moneys worth out of this one a long time ago.

I'm saving money now to get 72v/16ah headway system so this only need to work for a couple of months..
The label says lifepo4 but everyone that i showed this pack at on the golden motor forum tells me they are linmo2 cells...
The cells themselves dont have any label or markings whatsoever...
The state of the battery is almost new as they only been use like 10 times because the cells were bad right from start.
My retailer doesn't want to give me a new one because he thinks i broke it by putting it in series.. (the other pack still works fine)
i asked him for some replacement cells but he didn't had any...
So its i dont really care about the lifespan of this pack i just want it to work for like maybe 6 months or so or lets say another 100 cycles.

Enpro said:

The label says lifepo4 but everyone that i showed this pack at on the golden motor forum tells me they are linmo2 cells...

Click to expand...

Curious...Assuming "linmo2" means LiMn-type chemistry, those two chemistries have (AFAIK) different HVC and LVC voltages, with the nominal voltage of the former being notably lower (3.2V) than the latter chemistry (3.7-3.8V).

If indeed it isn't what it says it is, then if it has a BMS and charger that only charges them to LiFePO4 levels, and allows them to discharge down to LiFePO4 levels, perhaps it has damaged the cells in the process. Or alternately, maybe the cells that are bad are just bad, but you would get more capacity from them (and longer life) if they were charged up to their correct levels, and prevented from discharging down to too low a level?

10 cells in series, 42V charger implies it is limno4 like the konions and it may not even have a BMS. just buy some more limno4 cells on ebay to match the capacity and use them and then you can use the original charger like you wanted to.

I dissembled the 42v charger and it does have a little pot meter covered with glue.
How big are my chances that this is indeed a voltage regulator?
And can i just dail the thing a little bit without anything exploding?

Edit: Probably a bad idea after all so i just stick with the original bypass plan.
The bms still works when i pulled out the the last balance lead.