The Battery Management System is the hardware and software control unit of the battery pack. It measures every cell voltage and temperature, estimates the state of charge and state of health, controls the contactors, manages cell balancing, protects against abuse conditions, and communicates pack status to the wider system.
No other single component in a battery system touches as many engineering disciplines simultaneously. The BMS sits at the intersection of electrochemistry (it must model cell behaviour), embedded software (it must run real-time algorithms on constrained hardware), functional safety (it must be designed to ISO 26262 at ASIL C or D for critical functions), and systems engineering (it must interface correctly with the vehicle, charger, and thermal management system).
BMS functional safety is more demanding than most automotive ECU development. ISO 26262 at ASIL C or D requires systematic hazard analysis, safety goals, safety requirements, design verification, and documented evidence of compliance. If you’re coming from body electronics or infotainment development, the rigor is substantially higher.
Getting the BMS right determines whether the pack delivers on its energy and power targets, whether it lasts its intended lifetime, and whether it is safe. No amount of good cell selection or pack design compensates for a poorly implemented BMS.
Where to Start
New to battery management systems? This is a suggested reading order through the core material. For the complete list of every page in this section, see the Full Index below.
- Hardware – the whole picture: what a BMS physically is and does.
- Sensors – what a BMS measures in order to do its job.
- State of Charge – the most fundamental value a BMS calculates.
- State of Health – tracking how a cell degrades over its life.
- Battery Management System Algorithms – how raw sensor data becomes those estimates.
- Functional Safety – how a BMS protects the pack and guards against faults.
- Harness – the wiring that connects it all together.
Full Index
A complete, always up to date index of every page in the BMS section, organised by hierarchy. Click any heading to expand it.
What the BMS Does
The BMS has two broad categories of function: measurement and estimation, and control and protection. These are deeply interdependent — the quality of the control decisions depends entirely on the accuracy of the measurements and estimates.
Measurement
The BMS directly measures:
- Cell voltages — every cell in the pack, continuously. Typically via an analogue front-end (AFE) IC that measures voltage across each cell in a series string and reports via SPI or I²C to the main processor
- Pack current — total current in and out of the pack via a shunt resistor or Hall-effect sensor in the HV circuit
- Cell temperatures — via NTC thermistors placed on cell surfaces or between cells; typically one sensor per 4–8 cells, positioned at the expected thermal hotspots
- Pack voltage — total HV bus voltage, independently of cell voltage sum, for isolation monitoring
- Isolation resistance — the resistance between the HV bus and chassis ground, monitored continuously
Estimation (State Functions)
Quantities that cannot be measured directly must be estimated using models fed by the measured inputs:
- State of Charge (SoC) — the amount of charge remaining in the pack as a percentage of usable capacity. Cannot be measured directly; must be estimated from voltage, current, temperature, and cell model
- State of Health (SoH) — the current usable capacity as a percentage of the original capacity when new. Tracks degradation over the pack’s lifetime
- State of Power (SoP) — the maximum power the pack can deliver or accept at the current instant, accounting for SoC, temperature, and SoH
- State of Energy (SoE) — the remaining energy available for the application
- Remaining Useful Life (RUL) — a prediction of the number of cycles or calendar time before the pack reaches end-of-life
Eatron Technologies Our intelligent software platform approach significantly reduces cost, risks and time to market. Eatron offers embedded applications for High and Low Voltage Battery Management that are not only automotive grade, safe and robust but also integrated with AI & cloud layers with analytics, offering OTA updates and continuous software improvements.
Control and Protection
Based on its measurements and estimates, the BMS:
- Controls the contactors — switching the HV circuit on and off, executing the pre-charge sequence on power-up, and opening contactors on fault conditions
- Manages cell balancing — dissipating or redistributing energy between cells to keep all cells at equal SoC
- Controls the thermal management system — switching cooling or heating on and off to keep cells within their operating temperature window
- Enforces operational limits — derate charging or discharging current as the pack approaches voltage, temperature, or SoC limits; cut off entirely if limits are exceeded
- Runs safety diagnostics — monitors for cell voltage deviations, temperature anomalies, isolation faults, and other indicators of developing faults
- Communicates — reports SoC, SoP, SoH, fault status, and operational limits to the vehicle control system or energy management system via CAN, LIN, or Ethernet
State of Charge Estimation
SoC is the battery’s equivalent of a fuel gauge — except it cannot be measured directly. A battery has no dipstick. The SoC must be inferred continuously from observable quantities using a model of the cell’s behaviour.
Three fundamental approaches are used in production BMS systems, typically in combination:
Coulomb Counting
Integrates current over time: SoC = SoC_initial + (∫I dt) / Q_nominal
Simple to implement and accurate when the current sensor is accurate and the initial SoC is known. The limitation is cumulative error: any current sensor offset, measurement noise, or uncertainty in Q_nominal accumulates over time, causing SoC drift. Coulomb counting alone is reliable over short periods but requires periodic correction from another source.
OCV-Based Estimation
The open circuit voltage of a cell is a deterministic function of SoC — for a given chemistry, the OCV-SoC curve is well-characterised from cell testing. By measuring the cell voltage during a rest period (no current flow), the BMS can read SoC directly from the OCV lookup table.
The limitation: this only works when the cell is at rest. Under load, the terminal voltage differs from OCV by I×R_int. In most applications — EVs in use, BESS cycling — the cell is rarely at rest for long enough to get a stable OCV measurement.
Model-Based Estimation (Kalman Filter, ECM)
The most accurate production approach combines an Equivalent Circuit Model (ECM) of the cell with a Kalman filter or similar state estimator. The ECM predicts the expected terminal voltage given the current SoC and current demand. The Kalman filter continuously compares predicted voltage against measured voltage and corrects the SoC estimate. This approach handles dynamic operating conditions, compensates for sensor noise, and tracks SoC accurately under load.
The accuracy of the model depends on how well the ECM parameters (OCV curve, internal resistance, RC time constants) are characterised from cell testing — and how those parameters change with temperature and age. A well-parameterised model-based estimator achieves SoC accuracy of ±2–3% across the operating envelope.
→ SoC estimation techniques — full detail → State of Charge algorithms
State of Health Estimation
SoH tracks the long-term degradation of the battery pack — how much of its original capacity and power capability remains. It is the parameter that tells the user, service engineer, and recycler what the pack is actually worth at any given point in its life.
Capacity-based SoH — SoH = Q_current / Q_nominal × 100%. In most automotive applications, 80% SoH is defined as end of life for the primary application (though the pack may be suitable for second-life use at lower SoH).
Resistance-based SoH — tracks the growth in internal resistance as the pack ages. Resistance growth reduces available power before capacity fade becomes significant in many high-power applications.
Accurately estimating SoH in service is harder than SoC estimation. Capacity can only be directly measured during a full charge-discharge cycle — which may happen rarely or never in normal use. Production BMS systems use a combination of incremental capacity analysis, resistance tracking, and long-term trend models to estimate SoH from partial cycle data.
SoH tracking has become significantly more important with the EU Battery Regulation (2023/1542), which mandates that EV batteries report a State of Certified Energy (SoCE) — a standardised SoH metric — from 2026 onwards. This drives OEMs towards more rigorous on-board SoH estimation than was previously required.
Cell Balancing
Cells in a series string will inevitably develop differences in SoC over time, due to manufacturing variation in capacity and self-discharge rate. As these differences accumulate, the cell with the lowest SoC limits charging (it reaches maximum voltage first) and the cell with the lowest capacity limits discharging (it reaches minimum voltage first). The pack delivers less energy than the sum of its cells would theoretically allow.
Cell balancing corrects this by equalising the SoC of all cells in the string. There are two approaches:
Passive (dissipative) balancing — a resistor is switched across each cell that is above the target SoC, dissipating the excess energy as heat. Simple, low-cost, and universally used. The limitation is that energy is wasted rather than recovered, and high balancing currents generate significant heat in a confined space.
Active (redistributive) balancing — energy is transferred from high-SoC cells to low-SoC cells using DC-DC converter circuits. More energy-efficient but significantly more complex and expensive. Used in applications where balancing currents are large enough to justify the cost — some BESS applications and high-performance automotive systems.
In practice, passive balancing is the dominant approach in automotive and most BESS applications. Balancing currents are typically 50–200mA — small enough that the energy loss is acceptable, and the heat generated is manageable.
→ Cell balancing in depth → Passive balancing → Cell balance and end of life
BMS Architecture
The physical architecture of the BMS — how the measurement, processing, and control functions are distributed across hardware — depends on the size and complexity of the pack.
Centralised BMS
All functions — cell voltage and temperature measurement, balancing, current sensing, contactor control, and communication — are implemented in a single control board. Long wiring harnesses run from the board to every cell measurement point.
Centralised BMS is most suitable for smaller packs with a modest cell count. It is simpler in terms of software architecture (one processor, one firmware image) but harder to scale — the harness becomes impractical above a few hundred cells, and a single board failure takes down the entire system.
Master and Slave (Distributed) BMS
The measurement functions are distributed across multiple slave boards — typically one per module — each monitoring the cells within its module and communicating back to a central master controller. The master aggregates all slave data, runs the state estimation algorithms, controls the contactors, and manages external communication.
Distributed architecture reduces harness complexity dramatically — each slave board connects only to the cells within its module, and a digital bus (typically isoSPI or CAN) runs between slave and master. It is also more scalable and more fault-tolerant — a failed slave module isolates to that module rather than the whole pack.
The dominant architecture for large automotive and BESS packs.
Wireless BMS
A more recent development, used in production by GM (Ultium platform) and others, replaces the wired communication between cell sensor circuits and the master controller with a wireless link. This eliminates the low-voltage wiring harness within the battery pack — a significant weight and cost saving in large packs — and improves modularity.
The challenges are latency management, electromagnetic interference in an HV environment, power supply for the wireless nodes, and functional safety certification of a safety-critical wireless link.
Sensors
The BMS is only as good as the measurements it receives. Sensor selection, placement, and calibration directly determine the accuracy of every state estimate and the reliability of every protection function.
Cell voltage measurement — the most critical measurement. Accuracy of ±5mV or better is typically required; errors directly propagate into SoC and SoP estimates. AFE ICs from suppliers including Texas Instruments (BQ series), Analog Devices (LTC series), NXP, and Renesas provide multi-cell voltage measurement at the required accuracy.
Temperature sensing — NTC thermistors are the dominant technology; PT100/PT1000 RTDs are used where higher accuracy is required. Sensor placement is critical — the BMS can only protect what it can see. Sensors placed between cells at the expected thermal hotspot (typically the centre of the module, or the cell adjacent to the inlet of the cooling circuit) are more informative than sensors on the module housing.
Current sensing — shunt resistors or Hall-effect sensors. Shunts offer high accuracy and low offset but generate I²R heat. Hall sensors are non-invasive and generate no heat but require careful calibration for temperature drift and magnetic offset. Current measurement error accumulates directly into SoC error via coulomb counting.
Isolation monitoring — continuous measurement of the resistance between the HV bus and chassis. A dedicated isolation monitoring IC (from suppliers including Bender, ISOMETER) injects a low-level signal onto the HV bus and measures the response. ISO 6469-3 requires the BMS to detect isolation faults and open the contactors.
Additional sensors — humidity (condensation detection), gas sensing (early thermal runaway detection from VOC or CO₂ release), pressure, and coolant flow are used in more sophisticated designs.
→ BMS sensors in depth → Cell temperature sensing → Current sensor → Isolation resistance monitoring
Harness Design
The BMS harness is the low-voltage nervous system of the battery pack — carrying cell voltage measurements, temperature sensor signals, balancing current, and communication signals between the cells, slave boards, and master controller.
In a large automotive pack, the harness can include hundreds of individual connections. It must route through a mechanically stressed, thermally cycling, chemically challenging environment and maintain reliable contact for ten or more years. Harness design failures — broken connections, chafed insulation, corroded contacts — are a significant source of field BMS failures and warranty costs.
Key design considerations:
- Connector selection — automotive-grade connectors with positive locking and vibration resistance (typically Molex, TE Connectivity, or Amphenol)
- Wire gauge — voltage sensing leads carry microamps to milliamps; balancing leads carry up to 200mA; sized accordingly
- Routing — away from sharp edges, hot surfaces, and areas of high vibration; secured at regular intervals
- Shielding — voltage sensing leads in high-EMI environments may require shielding to prevent measurement noise
- Redundancy — critical signal paths (isolation monitoring, contactor control) may be duplicated for safety
Functional Safety
The BMS implements safety-critical functions — isolation fault detection, overvoltage and overtemperature protection, contactor control — that directly protect the user and the vehicle. In automotive applications, these functions must be developed to ISO 26262, with the safety integrity level (ASIL) determined by a Hazard Analysis and Risk Assessment (HARA).
Most main BMS safety functions — contactors, isolation monitoring, overtemperature protection — are classified at ASIL C or ASIL D. This requires:
- Systematic hazard and risk assessment establishing safety goals
- Safety requirements derived from safety goals and allocated to hardware and software
- Hardware metrics (SPFM, LFM, PMHF) demonstrating adequate fault coverage
- Software developed to ISO 26262 Part 6 (ASIL C or D as appropriate)
- Validation evidence demonstrating that safety goals are met
For BMS suppliers without an automotive background, this is often the most significant challenge in entering the EV market. The functional safety process is not a documentation exercise — it fundamentally shapes hardware architecture (independent monitoring paths, watchdog circuits, safe-state design) and software architecture (AUTOSAR, separation of safety and non-safety code).
The IEC 61508 and IEC 62619 standards cover functional safety in industrial and stationary applications respectively, with analogous (though less stringent) requirements to ISO 26262.
→ Functional safety in depth → Software development and functional safety → Emergency shutdown
Cloud Data and Fleet Analytics
The BMS collects a continuous stream of cell voltage, temperature, current, and state data throughout the pack’s operating life. In a connected vehicle or BESS, this data can be transmitted to the cloud for analysis at fleet scale.
Cloud-based BMS analytics enables:
- Improved SoH estimation — fleet-wide ageing data improves the accuracy of individual pack SoH models beyond what any on-board model alone can achieve
- Predictive maintenance — early detection of developing faults from subtle trends in cell voltage deviation, internal resistance growth, or balancing current demand
- OTA calibration updates — BMS algorithm parameters and limits can be updated over-the-air as the understanding of pack behaviour improves
- Second-life assessment — accurate SoH data supports the economic case for second-life repurposing rather than premature recycling
- Regulatory compliance — EU Battery Regulation Battery Passport requirements include lifetime data logging that cloud connectivity can satisfy
The shift from standalone BMS to cloud-connected BMS is a defining feature of the fourth generation of BMS development, and represents a significant part of the value-add that BMS software suppliers are competing on.
→ Cloud data → In the cloud → BMS history and roadmap
BMS Roadmap
BMS development is following a clear trajectory across four generations:
Generation 1 — basic protection: overvoltage, undervoltage, overtemperature cutoff. Simple threshold-based logic.
Generation 2 — state estimation: coulomb counting SoC, basic SoH tracking, passive balancing. The BMS becomes an estimator as well as a protector.
Generation 3 — model-based: Kalman filter SoC, ECM parameterisation, sophisticated SoH tracking, functional safety compliance, CAN communication to vehicle network.
Generation 4 (current frontier) — cloud-connected, AI-enhanced: fleet-scale learning, machine learning SoH models, OTA updates, predictive maintenance, battery passport compliance, wireless architecture.
The next frontier is the integration of physics-based degradation models directly into the on-board BMS — moving from empirical capacity tracking to mechanistic understanding of SEI growth, lithium plating risk, and particle cracking, enabling truly predictive lifetime management.
Related Sections
The BMS sits at the centre of the battery system, interfacing with every other section:
- Battery Cell — the BMS models, measures, and protects the cell; cell parameters are the inputs to every BMS algorithm
- High Voltage System — the BMS controls the contactors, monitors isolation, and commands the pre-charge sequence
- Thermal — the BMS commands the thermal management system and uses temperature measurements to adjust operational limits
- Safety — thermal runaway detection, emergency shutdown, and isolation fault response are BMS functions
- Electrical — current sensing, isolation monitoring, and HV system interfacing are shared responsibilities
- Lifetime — the BMS operating strategy (SoC window, charge rate, temperature management) is the primary lever for controlling cell degradation rate
- Modelling — ECM and physics-based models are the computational core of BMS state estimation algorithms
- Legislation, Rules and Regulations — ISO 26262, IEC 61508, and EU Battery Regulation compliance requirements drive BMS architecture and data logging
- System — the BMS communicates pack status to the vehicle control system, energy management system, or grid controller
See also: GB 17761-2024, Battery Charging, BMS History and Roadmap