High-Voltage Battery Management System

Passive balancing dissipates excess energy from higher-voltage cells as heat through resistors. It's simpler and lower cost but wastes energy and generates heat. Active balancing transfers energy from higher-voltage cells to lower-voltage cells using inductors or capacitors. It's more efficient (80-90% energy transfer) but more complex and expensive. For most applications, passive balancing is sufficient and preferred for cost reasons. Active balancing is justified for large battery packs (>10kWh) where energy efficiency is critical, or where heat generation must be minimized. The ADBMS6815 supports passive balancing natively

active balancing requires external circuits like the LT8584.

Use passive balancing for cost-sensitive applications; active balancing for large packs where efficiency is critical.

ASIL-D is the highest automotive safety integrity level requiring <1% probability of dangerous failures. Key requirements: 1) Redundant measurement paths for critical parameters (voltage, temperature)

2) Comprehensive fault detection with diagnostic coverage >99%

3) Safe state transition on detected faults (contactors open, pre-charge active)

4) Independent watchdog and safety monitor

5) FMEDA (Failure Modes Effects and Diagnostics Analysis) documenting all failure modes. The ADBMS6815 includes built-in self-test, redundant ADCs, and CRC protection on communication. Use the provided safety firmware as a starting point, but you'll need to perform your own safety analysis and testing. Work with a functional safety consultant for first-time ASIL-D designs.

Use redundant measurement paths; implement >99% diagnostic coverage; plan for extensive safety validation testing.

High-voltage BMS requires reinforced isolation (2500Vrms) between high-voltage battery and low-voltage control circuits. Key isolation barriers: 1) Cell monitoring IC to communication interface (integrated in ADBMS6815)

2) Communication lines between battery modules and master controller (use isolated CAN or SPI)

3) Power supply isolation (use isolated DC-DC converters). Isolation must be maintained under single-fault conditions. Requirements per IEC 60664: clearance distances (air gap) and creepage distances (along surface) depend on working voltage and pollution degree. For 800V systems, typical clearance is 8mm and creepage is 16mm. Use optocouplers or digital isolators (ADuM series) for signal isolation. Test isolation resistance periodically as required by safety standards.

Use reinforced isolation rated for 2500Vrms; maintain 8mm clearance and 16mm creepage for 800V systems.

SOC (State of Charge) estimation accuracy depends on algorithm quality and sensor accuracy. With the ADBMS6815's 1mV voltage accuracy and ±1°C temperature accuracy, combined with Kalman filter-based algorithms, typical SOC accuracy is 2-3% across the full temperature range (-20°C to +60°C). Key factors affecting accuracy: 1) Current sensor accuracy and calibration

2) Battery model parameters (capacity, impedance vs temperature)

3) Algorithm tuning for specific cell chemistry

4) Initial calibration and learning period. Coulomb counting alone drifts over time

voltage-based correction at rest periods is essential. The provided firmware includes adaptive algorithms that learn battery characteristics over multiple cycles. For best accuracy, implement both voltage-based and current-based estimation with Kalman filtering.

Expect 2-3% SOC accuracy with proper calibration; use Kalman filtering combining voltage and current measurements.

The ADBMS6815 supports daisy-chain SPI communication between battery modules. The master controller communicates with the first slave via isolated SPI

each slave passes data to the next in the chain. This reduces wiring complexity compared to individual SPI connections. For external communication, isolated CAN 2.0B is the automotive standard for BMS. The system can also support Ethernet for grid-scale storage applications. Data rates: Daisy-chain SPI up to 1Mbps

CAN up to 1Mbps. The daisy-chain architecture supports up to 16 slaves (256 cells) with <10ms update time for all cell voltages. For systems requiring higher data rates or longer distances, consider using multiple daisy chains with an FPGA aggregator.

Use daisy-chain SPI between modules; isolated CAN for vehicle communication; supports up to 256 cells with <10ms update time.