BMS Repair

[jerryk]

I'm in the midst of a repair to a BMS out of an LiFePO4 battery pack.  I'm attempting make a part order but am held up on the identification of one of the smd components.  It's markings appear to be BBG Z622.  There are two on the board at positions D1 and D3.  D1 is between the Drain of the power mosfet array and the positive terminal of the battery and the other D3 is between the Source of the power mosfet array and the positive terminal of the battery.  I think they are tvs or some sort of suppression diodes.  One failed via short circuit and the other tests as open circuit.  Both fail the diode test in either direction.

Any help on identifying this component would be greatly appreciated.  The board and the component are pictured below.

Thanks - Jerry

[fzabkar]

SMBJ15CA, STMicroelectronics, 15V bidirectional TVS diode, 600W, marking BBG, DO-214AA:
https://www.st.com/resource/en/datasheet/smbj.pdf

[jerryk]

That was fast and appreciated.

Jerry

[m3vuv]

ive built a few ebikes,the only issues ive had is with the battery bms units,they always seem the weakest link!

[jerryk]

I have finished the repair of the BMS but there are issues that are triggering the onboard warning led (LED1 in attached schematic) that I cannot resolve.  The manual for the battery describes various fault conditions that boil down 3 blink patterns.  The led blinks in patterns other than listed for the most part,  but will turn on solid at times also, indicating a general fault condition.  Setting that aside, the board mosfets transfer power from battery cells to main terminals and all cells seem to charge and discharge evenly, so far.

I'm hoping some of you will look at the schematic I've drawn for any ideas where such a fault might be hiding. I'm starting to worry that the microcontroller took a hit when this battery went bad.

So far I have replaced 6 of the twelve power mosfets, both gate controllers, R33 (current sense resistor?), and both tvs diodes at D1 and D3.  Three of the power mosfets were cracked and failed shorted between gate and drain.  All other components seem to test fine with diode tests.  I know that this is not much of a test but trying not to disassemble the entire board.

I have hooked up one of the lifepo4 packs and ran a small 1 amp fan load and it seemed to discharge evenly.  Prior to hooking the battery pack each cell was charged individually with a lab power supply. The cells are balanced and seem to be holding charge just fine.  The voltages at pins7,8,10, and 12 of the microcontroller show voltages almost the same.  I'm not sure how exact balancing is supposed to be.  The voltages at pins 5,6,9, and 11 of the microcontroller are always zero in both charging and discharging events(so far).

Is there any good ways to trick the cell balance function of the microcontroller without using one of the original battery packs?

Any advice you have on saving this nice battery would be greatly appreciated.  BTW - battery like this new is $450 US.

Thanks in advance - Jerry

[jerryk]

I made a less than perfect bms tester from a series of 20k pots and on the lab powers supply was able to simulate some of the conditions like over charged cells and over discharge limits.  The bms board seems to work as it was designed but there are still symptoms that point to an issue with the controller.  I still get a solid led light which according to the manual is a malfunction of the microcontroller.  If I add any load like a small fan to the main battery terminals the light turns off.  If I test the voltage at the cell vref pins 12, 10, or 8 with either a multimeter or oscilloscope probe the warning led turns off.  If I touch any of those pins at the controller led turns off.  Bottom line, any sort of impedance added to cell voltage reference pins, on the right side of the controller, causes the warning light to turn off.

I have tested all 4 cells above the over voltage limit of 3.65V and they activate the mosfets at Q1-4 for the respective cell.  Over discharge protection scheme seems in order and disconnects the ground between the lifepo4 packs below 11.5V and reconnects at around 12.5V after a cool down period.  So the microcontroller is doing what appears to be it's design function with the exception of activating a warning light that indicates a controller malfunction. Seems odd that adding impedance to right side of the controller toggles the warning light off.

I know it's a strange problem but any guesses at the what could be wrong?  Assuming the protection bit on the controller was not set is it too difficult for a noob at microcontrollers to move code to a new controller?


Jerry

[jerryk]

I hate to push this thread back up but part of me says I should at least post the end result on how this repair went.  I am still a little confused about some of the gremlin type issues in the above post.  I did replace all of the BSS192 D-mosfets since the above post.  From there I went through the  reassembly process.

Now I am putting it through several discharge and charge cycles and it seems to work without faults.  Cell condition seems be good.  There is one cell in each pack that fall a little behind in discharge but all balance on the high end of charge quite nicely.

For those that might be interested - Here is a link to photos of start to finish on the project https://photos.app.goo.gl/dUQgeV5dL2hj644Z6.  This includes the final schematic.

Although I have more high load test to perform, so far I'm calling this a successful repair in that I learned a few things along the way.

Thanks for the help.

Jerry

 

[eb4eqa]

Thanks for taking the time to come back and tell us how it went.

Roberto EB4EQA