SiC-Based BMS Architecture for High-Voltage EV Packs: From Cell Monitoring to Thermal Safety

  • Home
  • Blog
  • SiC-Based BMS Architecture for High-Voltage EV Packs: From Cell Monitoring to Thermal Safety

Modern electric vehicles push battery pack voltages past 800 V while demanding sub-millisecond fault response and cell-level monitoring precision. Silicon carbide (SiC) switching technology is at the center of that evolution.

Key TakeawayThis article covers a complete 800 V EV BMS — daisy-chained cell monitors, Kalman filter SOC estimation, SiC-based contactor circuits, and multi-zone thermal management.

Why SiC Changes the BMS Game

SiC devices offer 10x lower switching losses vs silicon IGBTs, handle junction temperatures up to 175 degC, and support switching frequencies exceeding 100 kHz. Their low RDS(on) under 10 mohm minimizes conduction losses in the main power path.

BMS architecture
Fig 1. High-level BMS system architecture for an 800 V NMC pack.

Cell-Level Monitoring Architecture

For a 96S2P NMC pack you need continuous measurement of 96 cell voltages with +/-1 mV accuracy. The two dominant IC families are the TI BQ76952 (16-cell, stackable via SPI daisy-chain) and the ADI LTC6813-1 (18-cell, isolated SPI).

ParameterTI BQ76952ADI LTC6813-1
Cells per device1618
Voltage accuracy+/-1 mV+/-0.8 mV
CommunicationSPI daisy-chainIsolated SPI
Cell balancingActive + PassivePassive only
/* Trigger simultaneous ADC across all BQ76952 devices */ BQ76952_SendCommand(CMD_ADC_CVT_ALL); HAL_Delay(2); HAL_SPI_Receive_DMA(&hspi1, cell_data_buf, 192); void HAL_SPI_RxCpltCallback(SPI_HandleTypeDef *hspi) { BMS_ParseCellVoltages(cell_data_buf, cell_voltage_mV); BMS_CheckOVUV(); }

Extended Kalman Filter for SOC Estimation

The EKF achieves +/-1.5% SOC accuracy over a full drive cycle by fusing coulomb counting with an OCV model, treating the battery as a second-order RC equivalent circuit. Compare to +/-8-12% for coulomb counting alone.

/* State vector: x = [SOC, V_RC1, V_RC2] */ float A[3][3] = { {1, 0, 0 }, {0, exp(-dt/tau1), 0 }, {0, 0, exp(-dt/tau2)} };
Engineering NoteOCV-SOC curve is characterized offline at 0.05 C, 25 degC. Temperature compensation of -0.5% SOC per 10 degC applies in real-time.

Active Cell Balancing

Inductive active balancing transfers charge between cells at ~88% efficiency vs 0% for passive dissipation. Trigger threshold: delta-V greater than 10 mV between adjacent cells.

  • Balancing current: 2-5 A per channel
  • FreeRTOS task at 1 Hz dispatches balancing commands
  • No heat dissipation inside cell module

Multi-Zone Thermal Management

32 NTC sensors across 8 zones report at 10 Hz. Charge current derates on T_max:

  1. T_max below 35 degC: Full 1C charge rate
  2. 35-45 degC: Derate 2% per degC above 35
  3. 45-55 degC: Activate liquid cooling; 50% current limit
  4. 55 degC+: Open contactors immediately
  5. 65 degC+: Thermal runaway protocol

ISO 26262 Safety Architecture

ASIL-B compliance requires redundant monitoring and full diagnostic coverage:

  • Dual-path voltage: STM32H7 + TMS570 safety monitor compare results
  • Contactor feedback confirms state within 5 ms
  • IMD checks pack-to-chassis isolation greater than 500 ohm/V continuously
  • External watchdog: safety co-processor pets independently of main MCU

Conclusion

An 800 V EV BMS is a multi-disciplinary challenge at the intersection of analog electronics, power semiconductors, signal processing, and functional safety. At Firmness Technologies, this is the kind of engineering we deliver every day.

Social Share:

Leave A Comment