Reliability of electric/hybrid electric vehicle battery management systems

There are a number of ways to consider when designing a reliable battery management system for electric vehicles (EVs) and hybrid electric vehicles (HEVs). One way to achieve reliability is to use a fully redundant circuit, of course, assuming that the cost is not a problem. This type of system uses the exact same circuit to perform the same function in parallel, and some form of voting on the result to always get the most secure effect. In such systems, if a faulty circuit is detected, the faulty circuit is automatically stopped and replaced by an identical backup circuit. However, in some systems, the consequences of failure cannot be borne by redundant circuits with high cost. Then, such systems may rely solely on the built-in reliability of each component used. Although a design with this feature can be provided at the lowest cost, this design poses a great risk. However, there is an optimal solution to this problem.

This article refers to the address: http://

The golden solution to high-reliability systems is fault monitoring, in which the circuit monitors various system components and reports any anomalies. Since the fault can occur anywhere in the circuit at any time, the more components that are monitored, the better. The response to the detected fault can be either a complete system shutdown or a simple service alert like an alarm light in a car.

Achieving longer battery life challenges

Lithium-ion batteries are popular because of their high energy density, and batteries made up of lithium-ion batteries are smaller and lighter than batteries of other chemical compositions with the same energy. In high power applications such as electric vehicles, hundreds of cells are stacked together to form a high voltage power supply that allows less current to pass through the thinner and lighter wiring. In this type of automotive application, the safety of the driver is the most important, followed by the satisfaction of the owner. Therefore, there is a strong drive to drive safe and reliable long-term operation. To achieve this, the amount of electricity on each cell must be continuously monitored to maintain an optimal level for years of use.

In its simplest form, the circuit is required to measure the voltage across each cell in the battery pack. This measurement is typically performed by an analog-to-digital converter (ADC) that passes the information to a microcontroller. The controller carefully manages the charging and discharging of all batteries so that they do not exceed a tight range, as exceeding this range can significantly shorten battery life. An integrated measurement circuit can greatly reduce the number of components when there are hundreds of batteries in the system. Linear Technology's LTC6802 is an integrated feature that measures the voltage of up to 12 cells through a built-in 12-bit ADC. Any number of batteries can be stacked on top of each other, and each set of measured voltages is serially transmitted to a main microcontroller in groups of 12 batteries. These components form the core of the battery management system.

In terms of extending the usable life of the battery, it is important to carefully control the state of charge of each battery, but this may not be enough to meet the demands of increasingly demanding automotive customers. To make customers satisfied and carefree for a long time, it is necessary to conduct a “what-and-result” analysis of the system. A few key questions to consider are: What if a wire connected to the battery is disconnected? What happens if the voltage measurement accuracy is shifted? What if the measurement IC is somehow damaged by the system voltage transient?

Most hidden faults cause the controller to assume that the battery or battery pack is in perfect condition, and in fact, the battery is not accurately measured. These batteries may then be fully discharged or dangerously overcharged without providing an alarm to the system. Need something to "monitor the monitor" to achieve a higher level of reliable operation.

BMS fault monitoring

As an alternative to the fully redundant measurement method, the fault monitoring circuit is connected in parallel with the measuring device and functions to review the basic functions of the system. The circuit in Figure 1 shows the circuit implementation of this method using the LTC6802 measurement device for a battery pack consisting of a 12-cell Li-Ion battery and an accompanying LTC6801 fault monitor.

The LTC6802-1 is the primary electronics in the system, measuring and reporting the voltage of each cell in accordance with the instructions, and adding a discharge current to the battery to distribute the charge on each cell. At the same time, the LTC6801 also monitors each battery on the battery pack. Without the intervention of the system controller, the LTC6801 periodically samples the voltage of each cell and performs simple undervoltage and overvoltage comparisons. If everything is ok, the LTC6801 provides a differential clock signal on the Status Output line. If something goes wrong, the clock stops. It does not provide information about the substance of the problem, it only indicates that something is not normal. Once this clock is stopped, the controller can perform diagnostics to determine the specific nature of the problem.

The LTC6801 has been carefully designed with many potential system failures in mind and is also easy to use. An important design requirement is to allow the device to work automatically without any software. The only external requirement is the power provided by the battery pack itself, as well as an enable clock signal. If the clock input is not enabled, the LTC6801 stops in a static low power state. All settings for the device's operating characteristics are done by device pin bonding, eliminating the need for external components.

Any number of LTC6801s can be stacked on top of each other to monitor hundreds of cells in very high voltage systems (see Figure 2). The enable clock is buffered and is output on two signal lines that are connected to the enable input of the previous device in the set of devices. The enable clock goes in and out of each device up to the top of the set. Similarly, the most important state output clock from each device is passed down to the Status Input pin of the next device in the group. If any fault is detected on any battery anywhere in the battery pack at any time, the state clock of the device monitoring the problem battery stops triggering. This quiescent state will be passed along the set of devices to the bottommost device. After the clock transition is stopped, the device can issue a signal to troubleshoot the fault.

Providing a continuous clock for the status line is an important feature. A static logic level that marks any system fault can indicate an erroneous logic state when indicating whether everything is normal in the system. This will make the fault monitoring solution useless. With the clock mechanism, the monitor component must work continuously to keep the clock running, and everything in the system must be normal, otherwise it stops. The fault signal cannot be fixed in all normal conditions.

Monitor component problem

There is no doubt that the reliability of the system is improved through redundant monitoring, but how can we ensure that the monitor itself works properly? It is very important to prevent undetectable failure modes. To solve this problem, the LTC6801 provides a built-in automatic self-test feature. This self-test feature is automatically executed after every 1024 test cycles. This self-test examines 4 main functions.

Checking that the ADC, voltage reference, and comparator are working properly is one of the tests. Check the internal reference voltage to make sure it is within a strict window. In addition, undervoltage and overvoltage conditions are also generated internally and the comparator must be responsive. This will ensure that the analog portion of the ADC is working properly, and it is equally certain that the comparison threshold can be changed and is accurate.

The digital portion of the ADC is also tested. Two test signals force the 12-bit output code to be 0xAAA or 0x555, alternating between 1 and 0. This confirms that no ADC bits are fixed.

A multiplexer switch failure can cause one or more cells to be skipped while other cells are measured repeatedly. Skipping the battery means that a bad battery may not be detected. Self-tests ensure that each cell is measured, or that each error is flagged.

The fourth very important self-test feature determines if a battery connection is open. For this test, each cell is measured with a small 100uA pull-down current, which is an error in the case of an open battery connection. This periodic self-test allows the system to operate more reliably. Checking the device that is performing the check function is more certain that everything is working properly in the system.

Keeping all batteries in the system at the right level of charge will extend the service life of expensive battery packs for years. The LTC6801 is an affordable way to improve the long-term reliability of Li-Ion battery management systems with redundant fault monitoring. Operating in parallel with a more accurate battery measurement system, the LTC6801 provides a review function to check that all system components are functioning properly, enhancing safety and reliability.