How to test a Lithium ion BMC

[radar_macgyver]

I have two identical Lithium ion packs for a bicycle light, they're potted so it's impossible to reach the BMC board within without damaging the packs. Lately, the bike light shuts off about 15-20 seconds after switch-on. Measuring the pack shows zero volts. If I attempt to charge (even for just a second) the pack voltage goes back up to ~16V (it's a 4S pack). Same behavior across both packs.

Is there some means of testing the BMC to figure out why it's cutting out, without damaging the potting? The bike light itself runs happily for hours off a bench supply (current draw is 0.5 A); it uses a Luxdrive BuckPuck 3021 and some Cree XP-E LEDs.

Thanks!

[sleemanj]

If removing the load and re-applying it does not bring it back to life, but attaching a charger does, then it's probably shutdown due to under voltage rather than over-current. 

Typical protection circuits will re-enable from under voltage only when you apply reverse current (charge).

Assuming you have actually charged the battery packs, then most likely one of the cells in it has gone dead.

If you do want to test it under load, then use an incandescent light bulb (12v car brake/indicator one, say 5 or 10w) will be the easiest way.

But if it's potted, there's not much you can do about it either way.

[radar_macgyver]

Well looks like that might be the case. After charging overnight, I tried to discharge it with an RC Lipo pack charger/discharger, and it cut out pretty quick. Did it again with a voltmeter on the output, and it fell to about 14.5 V before the BMC cut off. Rats - need a new pack Strange that both behaved in such a similar manner.

Thanks for the pointer. I forgot that BMCs can cut off even if a single cell falls below threshold.

[jbb]

Sounds like your BMS isn't balancing the cells. If one cell is near empty, the under voltage protection will trip after a few seconds of discharge. If another  cell is near full, then over voltage protection will trip after a few seconds of charge.

The pack is unfortunately done for.

However, one question remains: does the BMS try to balance the cells, or does it just provide protection? To find out you'll need to do a tear down... (or maybe find someone else's). Maybe the BMS never balanced, or maybe the balancing broke...

[radar_macgyver]

So I took apart the pack, and got that familiar smell of the Lithium-ion electrolyte. I discharged the pack as before, watching each of the cells with a voltmeter. Sure enough, one cell's voltage fell much quicker than the rest, and when it reached about 2.8V, the BMS cut off the output. I desoldered the BMS board, and found it uses an S8254A controller, which does individual cell monitoring, but does not balance across cells. Cells are Samsung.

Guess I can make a new pack from the three good cells (they held up 4.1V, even with a 1 A discharge current) plus others I have lying around. Maybe get a better BMS board - any recommendations?