In the case of hybrid electric vehicles (HEVs) and electric vehicles (EVs), the use of lithium-ion batteries provides the best balance between power, energy density, efficiency and environmental impact. But at the same time, lithium-ion batteries are fragile and dangerous, and the automotive environment is tricky and difficult to deal with. The challenge for electronics in hybrid and electric vehicles is to bridge the gap between the demanding automotive environment and battery sensitivity. The harsh environment of the car and the sensitivity of the battery are the perfect match in hell.
Given the energy, power and environmental requirements of a vehicle, the safe and reliable use of large lithium-ion battery packs is by no means an easy task. When a Li-Ion battery is operated in a fully charged or fully discharged state, the capacity decreases. The capacity of each cell degrades and deviates over time, taking into account cyclic charging, bank-to-bank differences, and varying environmental conditions. Therefore, for the battery pack to achieve the 15-year, 5000 charge cycle target, each cell must remain operating within a limited operating range. By controlling the state of charge (SOC) of each lithium-ion cell, the capacity of the battery pack can be maximized while minimizing capacity degradation. Ensuring efficient and safe use of automotive battery packs is the responsibility of the battery management system (BMS).
The task of the battery management system is to carefully track and control the state of charge of each battery. The measurement accuracy of the battery management system is critical because it determines how close each cell can operate to the edge of its reliable state-of-charge range. The ability to maximize the available capacity determines the number of batteries required, and the number of batteries has a large impact on cost and weight. Accurately measuring the voltage of each cell is difficult because the cells in a battery pack are susceptible to high common-mode voltages and high-frequency noise. To understand this, consider the fact that EV/HEV battery packs are typically very high voltage and consist of 100 to 200 cells connected in series. Such battery packs must provide rapid charge and discharge currents that may exceed 200A, and voltage transients may exceed 100V at the top of the battery pack.
The focus on cost and reliability has led to a move toward higher integration and lower component counts in automotive electronics. This trend is especially evident in highly complex battery management systems where we have seen battery monitoring ICs such as Linear Technology’s LTC6802 emerge. These highly integrated devices are key data acquisition components in new battery management systems, reducing cost, space and component count compared to previous discrete solutions. The primary function of the battery monitor is to directly measure the voltage of batteries connected in series, typically 12 channels per IC. Also included in this class of ICs are cell capacity balance control and additional measurement inputs (eg for temperature). To cope with high voltage battery packs, such devices are typically designed to communicate with each other via a daisy-chained serial interface. One component of a battery management system that typically cannot be successfully integrated into a battery monitoring IC is embedded software. State-of-charge algorithms are closely guarded technologies that are specific to chemical composition, size, form factor, operating conditions, and application. For new high voltage, high power battery packs, off-the-shelf algorithms are unlikely to be useful, and embedded software complicates Failure Mechanism Effects Analysis (FMEA), which system designers cannot do with embedded software direct control. Figure 1 illustrates the basic configuration of a battery module consisting of any cell, where the algorithm of the battery pack management system is software coded and controlled exclusively by the developer.
Figure 1: Basic topology of an electric/hybrid vehicle battery module
A key consideration for a battery monitoring IC is how to handle the noise it will encounter in the car. For example, many battery monitors use fast SAR converters to digitize batteries, which seems advantageous in data acquisition systems with more than 100 channels. However, the automotive environment is noisy and requires a lot of filtering, and this filtering determines the effective throughput, not the sample rate. For this reason, delta-sigma (DS) ADCs have advantages over SAR converters. For a given amount of noise rejection at 10kHz, a DS ADC at 1000 samples per second provides the same throughput as a SAR ADC at 1 million samples per second. For example, the LTC6802 employs a 1000 samples per second DS ADC that sequentially samples 10 input channels in 10ms. The built-in linear phase digital filter provides 36dB rejection of 10kHz switching noise. To achieve the same noise rejection at 10kHz, a SAR converter with 1 million samples per second requires a unipolar RC filter with a corner frequency of 160Hz on each cell (see Figure 2). The 12-bit settling time of the RC filter is 8.4ms, and even if the SAR ADC can sequentially sample 10 channels within 10us, more than 1 scan every 8.4ms is meaningless due to the filter’s response.
Figure 2: Comparison of delta-sigma converters and SAR converters with RC circuits delta-sigma converters provide the same effective throughput with better filtering
In the case of a long string of battery monitoring ICs, the serial interface is also an important consideration, and Linear Technology offers two distinct options. One option (and supported by most battery monitoring ICs) is a daisy-chain interface. With a daisy-chain interface, each IC in the chain can communicate with adjacent ICs without the need for optocouplers or isolators, leaving only the bottom device connected to a single microprocessor or control unit. In addition, Linear Technology offers a second option, an individually addressable serial interface. With this interface, a single microcontroller communicates with multiple parallel devices through isolation. This topology provides an inherently more reliable “star configuration” because loss of communication with one device does not interrupt communication with any other device. Addressable devices can also be used in a modified daisy-chain topology where relatively expensive isolators are a thing of the past and replaced by less expensive “transistorized” SPI bus configurations. The result is a serial interface with an extremely wide range of compatibility.
Figure 3: Gen 2 Daisy Chain Resists Strong Noise
The LTC6803 has a separate power input that can be disconnected while other connections remain intact (Figure 4a). In this hardware shutdown condition, the LTC6803 draws only a few nA. This is important for long-term storage of battery packs, as the current drawn by the integrated battery management system has the potential to unbalance the capacity of the cells in the pack. The LTC6803 can also operate from a separate power supply, allowing supply current to be drawn from a separate power supply rather than the battery pack, as shown in Figure 4b. The device also allows the use of a simple power-down function. In addition, with a separate power supply, the LTC6803 can continue to monitor a large number of cells even if the voltages of all cells have dropped dramatically (as can occur with supercapacitors and fuel cells). Figure 4c illustrates the advantage of having independent power inputs.
Figure 4a: Hardware shutdown
Figure 4b: Independent power supply
Figure 4c: Supply current versus operating mode
The growing number of electronics in cars has led to new standards for the quality and reliability of automotive electronics. As a result, automotive electronics standards such as AEC Q100, ISO 26262 and others have emerged. These standards translate into extensive qualifications and internal functions to ensure that system safety requirements are met. For example, the LTC6803 is an ISO 26262 compliant system. ISO 26262 is a practical safety standard that defines safety requirements for automotive systems and impacts system-level design issues such as redundancy, network configuration, data collection, sensors, and more. The LTC6803 has built-in open wire detection, digital filter checks, multiplexer decoder checks, as well as watchdog timers and an alternate voltage reference for comprehensive self-test capabilities.
Figure 5: Internal self-test function of the LTC6803
Many other improvements are included in the LTC6803 to meet needs beyond standard automotive designs. For example, the LTC6803 offers an extended -300mV to 5V measurement range that supports supercapacitors and NiMH batteries. Fully specified over the -40°C to 125°C temperature range, the LTC6803 is also designed to withstand supply voltages up to 75V to provide over 20% overvoltage margin.
The automotive environment is harsh for electronics, but the increasing electrification of cars is also an indisputable fact. Li-ion battery systems in electric and hybrid vehicles will soon become mainstream, and cutting-edge measurement devices such as the LTC6803 are essential for the success of lithium-ion battery systems. Not only are such devices required for accurate measurements, but these devices must perform reliably over long periods of time under very difficult conditions. Today, Linear Technology’s LTC6802 has proven this to be true in cars on the road. There is no doubt that the LTC6803 will continue the brilliance of the LTC6802 in the automotive market of tomorrow.
The Links: LMG9980ZWCC-01 EP2S60F484I4N