If it’s widespread then it might be a software regression, where the most recent release was built with some back level BMS integration code.
Or the newest release may have updated BMS firmware but not included updated integration code.
Complete speculation here, but this could be a loose end as some advancements from the Model 3 update are deployed across other lines.
These can be subtle issues. The “fail-safe” mechanism - the car’s computer falls back to default values when communicating with the unfamiliar BMS interface - can be either safe, where the default values work, or a failure with the car’s computer missing or misunderstanding key BMS data.
Years ago I dealt with a similar situation when several truckloads of older servers had network problems after moving from a site data center to a regional data center. The older servers worked flawlessly with the older Cisco network switches at their original site.
Unfortunately, the servers’ back-level network interface firmware could not integrate with the newer Cisco switches in the regional data center. Both the server and Cisco switch were thus forced to silently fail back to their default configuration for the network connection. The defaults were different on each end.
The mismatch was not evident at low traffic volumes while the servers were powered on and checked out. Once the applications started and network traffic volume ramped up the server was useless because of tremendous packet loss.
