Hello ADI team,
I have experienced two incidents with a BMS implementing active balancing based on the LTC3300-1.
The design is based almost explicitly on you evaluation board, DC2100B. Specifically, we implemented 2X LTC3300 in every board accompanied by the LTC6804, monitoring and balancing 8 cells.
The board can monitor and balance 8 cells while there are connected up to two boards in series for a 16S cell configuration all based on the eval. and your guidelines as per the relevant connections, fuses, layout etc.
The design has been working in a series production for more than 4 year successfully.
Recently, in a lab setup, without knowing the exact conditions, one of the secondary side's mosfet failed in shorted condition (Drain-Source-Gate all shorted) and an abnormal behavior was noticed.
The respective primary side's mosfet started to discharge the cell until it drained it out completely. After the incident we found out that the respective fuse of the 4-cell secondary winding to the 8-cell stack was blown which makes completely sense considering the short of the secondary mosfet.
Additionally, the shorted FET justifies that the WDT overvoltage protection did not work, because in this state no overvoltage appears.
The above explanation is our assumption, but it could be that the fuse might have blown first leading to overvoltage in the secondary winding and in extension the FET could be damaged, but in this case we would expect the balancing to have never started/ or ended immediately.
My question is:
Have you ever experienced such a similar failure? If not can you estimate the ICs reaction in a similar failure?
Could the shorted FET , especially considering the short also to GATE pin bring the IC to this condition and discharge the cell continuously? (i assume that in this state even if it tried to charge it, the same discharge effect would appear)
I also assume that the primary FET should have worked in switching mode after the failure thus keeping the current below nominal, because the cell fuses (12A) did not blow.
Could such failure defeat the BMS's order to LTC3300 to stop balancing when a cell undervoltage and/or overvoltage appears?
The secondary FET chosen is the SQJ488EP, while the primary is the SIR882DP.
The only failed components found were the secondary mosfet and the 7A fuse linking the 4 cells' secondary winding with the 8-cells' stack.
Thank you in advance.
Let me know if further info is required.
