Faulty BMS diagnosed; now what?

[harrisonpatm]

I was assisting a client in fixing his ebike. Long story short, the new 48v nominal Li-ion battery he purchased has something wrong with it. It has a switch on the side to turn the voltage on and off. When connecting the + and - to the controller, and turning the battery switch on, a small spark can be seen on the positive connection, a click can be heard from within the battery case (not the controller), and the voltage is disabled. Turning the battery off, rechecking connections, turn it back on, same thing. At one point a 10A fuse was put in line to the controller, which blew. This all happened without the motor phase, hall, or throttle connected. This happened using 3 different controllers. When the battery was swapped with a different 48v nominal battery, everything worked as expected: controller powered on, motor spun up, ect.

So, it seems like the "new" battery has something wrong with it. Likely the internal BMS. The fix is already done: the client will order a different battery, the ebike works as it should. My question is, why is this happening in the first place, to a newly purchased battery? It seems as though when simply connecting to the + and - of a motor controller (two different controllers), the battery immediately attempts to deliver a huge rush of current, which pops a fuse and causes the BMS to trip an internal overcurrent protection. Why? I don't believe it to be the controller's fault, as the same behavior happened with two different controllers, and nothing else was connected to the controllers except the battery. A different battery powers everything appropriately, with no sparks. What would cause this battery to exhibit this behavior?

I asked because the client will attempt to return the battery from where he got it. If he can't, I wonder if it's something that I can fix for him. Apologies: I don't have the battery with me currently, so I don't actually know the brand or model to look up. I was simply asking if this was a generic issue that someone has seen before.

Thanks!

 

[amberwolf]

It sounds like the BMS is still able to turn it's output off under overcurrent conditions, and back on once cleared.

If the discharge FETs had failed, allowing it to discharge even though it is "off", it wouldn't be able to turn it's output port off at all, so that eliminates the FETs as a failure under the above conditions.

If the BMS has a self-resetting polyfuse (which shouldn't make a noise) or circuit breaker (which might make a click) it could trip on overcurrent and shut off output, then self-reset once cleared, even if the FETs were failed stuck on...but the BMS couldn't be turned off anymore by the switch.

Actual cause of failure: most likely positive and negative are reversed in internal wiring between BMS and whatever output connector the battery has. Only thing I can think of that could cause what you see, without there being some other symptoms or problems.


This phrase:

When connecting the + and - to the controller, and turning the battery switch on, a small spark can be seen on the positive connection,

is a bit unclear. Does this mean it was connected *first*, then turned on, and then a spark was seen? If that is the order of events, it means that there is a very poor connection at the positive connection, because it could not have a spark across it if it was correctly / fully connected.

 

[harrisonpatm]

Actual cause of failure: most likely positive and negative are reversed in internal wiring between BMS and whatever output connector the battery has. Only thing I can think of that could cause what you see, without there being some other symptoms or problems.

See, this was my thought too. But when checked with a multimeter, the output wires from the battery terminals were correct. Positive 50v coming from the red + cable.

This phrase:

When connecting the + and - to the controller, and turning the battery switch on, a small spark can be seen on the positive connection,

is a bit unclear. Does this mean it was connected *first*, then turned on, and then a spark was seen? If that is the order of events, it means that there is a very poor connection at the positive connection, because it could not have a spark across it if it was correctly / fully connected.

This is correct, as I was performing these tests using alligator jumper cables. So you're spot on, the spark occurred at the alligator's teeth. Using more solid connections, no sparks were visible, and yet the rest of it still happened, noise in the battery, ect. In fact, using the weak alligator connections helped me diagnose that for some reason, significant current was being drawn from the battery, as indicated by the spark.

Edit: I failed to mention that one "test" i performed was to leave the battery switch on the "on" position, and then connected the + and - to the + and - of the controller. I was using an xt60 connector at the time, and the result was an exploded xt60 connector, positive pole. After that i would make connections, then turn the battery on, resulting in the small spark and audible battery pop, followed by voltage cutoff.

 

[lnanek]

Maybe one of the balance wires is disconnected? That wouldn't interfere with discharge voltage immediately, but it would convince the BMS a cell is dead, so it might disable discharge as soon as it starts.

If you take the battery apart, worth testing every connection into the BMS and checking the BMS for burn marks. I have one BMS where the board is burned where the charger connects, but it functions other than charging, which I do slowly over the balance wires instead.

 

[amberwolf]

Except for huge-capacitance systems (very very large controllers, for instance, not something you'd find on a typical ebike) I've never encountered a situation that could cause a current large enough to blow a fuse like that, other than failures eliminated by stuff you've already tested:

--Internal controller short (failed fets)

--reversed wiring between battery and controller (or inside one or the other)

--ohter system wiring short between main + and -.

There's nothing in the battery itself that could "force" current out in a way that could cause this, that wouldn't also probably damage the electronics being connected to it, and those have also been eliminated by your tests.

--reversed wiring

--pack wired internally as 26s instead of 13s

Current is determined by load, so if the load isn't high enough to blow the fuse using a different 13s (or higher) pack it can't be high enough for this one to do so.


You could certainly get sparks upon initial connection of battery to controller as it charges the caps (which is the reason for that precharge stuff you already know about ) but you wouldn't get what you're seeing from that. Sometimes a BMS with too sensitive an overcurrent could trip upon connection, but you're not seeing that (and it wouldn't blow the fuse).


Even if the BMS had failed in "interesting" ways there's still no way for it to force current out in a way that would blow fuses like you see, without one of those other things being the case.

 

[From-A-To-B]

... I have one BMS where the board is burned where the charger connects, but it functions other than charging, which I do slowly over the balance wires instead.

I worry for you, my ES friend! Some folks on here don't trust BMSs at all, and you've established a partnership with a burnt board.

@Harrisonpatm - I've got nothing helpful to add about why the BMS isn't behaving well, or what's up with your sparking pack behavior. Given that you've established there's something wrong with the battery (because other packs behave well), I'd skip the step of trying to troubleshoot the BMS and just install a new one. For $30 and a little time, you've got a functioning battery. While you're installing the new kit, be sure to peek at the rest of the battery to make sure everything looks good inside! (I'm mostly thinking about the main + / - connections to the pack, where manufacturers skip corners and solder 12awg wires into a single cell at the end).

 

[harrisonpatm]

I may do this. Ultimately it's not up to me, as the client hopes to simply return it for a refund. I would absolutely open it up if it were mine, I'm all for the learning experience.

 

[amberwolf]

Just note that even if there is a faulty BMS, it couldn't force current thru to blow a fuse, etc. Some other problem has to cause that....