Battery screening and diagnostic tool

There was no standard way to screen an incoming battery. Diagnosing one meant using the manufacturer's own tool. Those are expensive, one per brand, each with its own procedure and its own trained-operator requirement. In a high-throughput warehouse that does not scale.

The consequence was worse than slow. In practice batteries were not tested at all unless they already showed a fault, which means the first thing a bad battery met was a vehicle. A pack with a short or an over-voltage condition on its communication lines can take the vehicle electronics, or the diagnostic equipment, with it.

So the requirement was not really a measurement instrument. It was a gate: something that assumes the battery in front of it might be dangerous, proves otherwise before connecting anything sensitive, and gives a warehouse technician a clear answer with almost no training.

Structured testing paid for itself. Functional-block testing on the V1 board caught what would have been expensive later: incorrectly configured level shifters, reversed relay coil polarity, swapped Rx and Tx on the MCU UART0 interface and the CAN transceiver, a wrong push-button footprint, and MCU reset instability. Integration testing surfaced battery communication problems and incorrect ADC addresses, fixed in firmware. Because those three stages were thorough, system acceptance testing, run against a test plan agreed with the client with the board in its enclosure and a prototype harness, turned up no major issues.

Every known V1 issue is documented with a specific fix: tie the Rx level shifter permanently to 3.3 V, reverse R20 and R21 for the programming port, add a switch on the boost converter enable pin. A V2 starts from a known position rather than rediscovering all of it.