...
For MS15, some modules had banks with some amount of voltage difference. the highest difference was in somewhere in the range of 0.5V-0.6V. After taking apart one of the modules with a 0.5V imbalance, we found that there was a single cell in the outlier bank that was dead. We want to know what killed this cell. This Confluence page will outline our findings.
...
What caused the cell to die? Was it because of:
Spot welding
Connecting a fully charged cell to a fully discharged cell in parallel
Over-discharging?
Are there any other reasons?
Why is there a 0.2V imbalance between banks in some modules?
Is this also caused by the dead cell, a capacity issue, or something else?
Existing Information:
The dead cell likely did not short, as there are 15ohms of internal resistance and the dead cell is randomly located in the module (i.e., it is not in the same/similar location between the affected modules).
The IC’s transistors have high internal resistance
Source: analysis by hardware/electrical sub-teams
There are no visible signs of damage to the dead cells.
Note: there is some corrosion on random cells in an over-discharged pack, unclear if it is also present in normal-charged packs. It is present on the 2 dead cells and on cells with a functional voltage, indicating that corrosion is likely a separate issue and unrelated to the random dead cell in each module.
The dead cells are severely undervoltaged and act like capacitors.
...
The following modules are ordered from the positive terminal to the negative terminal, while the module numbers are as indicated on the orange wire bundles. The goal of checking this is to see if the parallel connections have any imbalances.
Module #9: 12.813V
3.2696V
3.2103V
3.1969V
3.1861V
Module #8: 13.129V
3.2562V
3.2650V
3.2684V
3.3429V
Module #7: 12.661V
3.4155V
3.4123V
3.4132V
2.4244V
Module #6: 12.998V
3.2720V
3.2380V
3.2438V
3.2453V
Module #5: 13.049V
3.2630V
3.2647V
3.2598V
3.2650V
Module #4: 12.991V
3.2314V
3.2448V
3.2314V
3.2875V
Module #3: 13.100V
3.2926V
3.2695V
3.2691V
3.2724V
Module #2: 13.096V
3.2756V
3.2751V
3.2703V
3.2793V
Module #1: 13.074V
3.2546V
3.2339V
3.2906V
3.3161V
Research:
If the heat is too intense or is applied for too long during spot welding, the cell can be damaged.
If too much pressure is applied during spot welding, the cell can be deformed, especially if its casing is thin. This can compromise the integrity of the cell.
If a fully charged cell were connected to a fully discharged cell, then the more likely scenario would be that both cells are damaged.
When such a connection is made in parallel, there will be a large current flow from the charged cell to the discharged cell in an attempt to equalize their voltages. This surge can generate excessive heat, potentially damaging both cells.
It is possible that the high current can cause only the discharged cell to overheat, which might lead to swelling or leakage (neither of which were noticed in the dead cells).
Maybe the dead cell was initially degraded and died later on when discharged?
Would be worthwhile to check if cells that are adjacent to the dead cell are performing worse than cells that are further away from the dead cell
The stress from high current flow may degrade the internal chemistry of the fully charged and fully discharged cells, reducing their lifespan.
Discharging the modules without a battery management system (BMS) could have resulted in cell voltage imbalances, with certain cells dying after accidentally being over-discharged [Source].
The BMS prevents imbalances and acts as a safeguard. Removing this safeguard, whether for testing or by accident, could have led to imbalances, with some cells being discharged below safe levels while other cells remain at a safe capacity.
This is the most likely cause of the dead cell, as testing was not consistently performed using a BMS. It would also make sense that only 1 cell was dead in each pack, as discharging was always carefully monitored to ensure that the modules would not become over-discharged. Differing internal resistances for each cell may have resulted in the dead cell, as one cell with a significantly lower internal resistance may have discharged more quickly and died, while the others remained fine.