With the rise of vehicle electrification, the number of digital power control systems and electric motors in automobiles is expanding rapidly. Correspondingly, electrical systems must be as efficient as possible to minimize battery size and maximize vehicle range. Every milliwatt of power consumption in the vehicle is more significant than ever.
Early hybrid and electric vehicles used insulated-gate bipolar transistor (IGBT) and silicon field-effect transistor (FET) switches; however, modern vehicle designs frequently take advantage of newer silicon carbide (SiC) and gallium nitride (GaN) switches that offer higher efficiency and faster switching speeds. Faster switching speeds enable greater efficiency in these systems, but electrical faults such as short-circuits and overcurrent conditions will also occur at faster rates, leading to the need for much faster fault detection.
A common requirement for automotive electrical systems operating from a 400-V or 800-V battery is to detect malfunctions resulting in overcurrent conditions extremely quickly. Most modern EVs must detect and respond to overcurrent conditions at <1 µs to prevent catastrophic and potentially dangerous system failures.
The need for speed and higher efficiency
Electrified vehicle systems such as traction inverters; electric turbochargers; heating, ventilation and air-conditioning (HVAC) compressors; positive temperature coefficient (PTC) heaters; DC/DC converters; and on-board chargers require current monitoring to detect overcurrent conditions and malfunctions. These electrified vehicle systems benefit from the higher efficiency provided by the faster switching frequencies of SiC and GaN switches. Older IGBTs have a maximum switching frequency of about 20 kHz, while silicon FETs can achieve switching frequencies of about 100 kHz.
The newer SiC and GaN switches have several advantages over older switch technologies, including lower gate and output capacitance, which enables higher switching frequencies while reducing associated switching efficiency losses. These advantages also reduce the size of required passive components and heatsinks. SiC-based switches can achieve switching frequencies >6× faster than silicon FETs, while GaN switches are capable of even faster switching, with some devices achieving >10-MHz frequencies. Figure 1 shows the relative switching frequencies of all four switch architectures.
Figure 1: Switching frequencies of IGBT, silicon FET, SiC, and GaN switches (Source: Texas Instruments Inc.)
In electrical systems, SiC and GaN switches can achieve switching rates of >100 V/ns and 1 A/ns, making it possible for voltages to change from 0 V to 800 V in just a few nanoseconds. Still, these speeds require constant monitoring for overcurrent conditions and respond quickly when a fault occurs. Simple and cost-efficient current-monitoring circuits using low-side shunt resistors (see Figure 2) are popular in electric motor-control applications for HVAC compressors, as well as PTC heaters and digital power applications such as DC/DC converters and on-board chargers.
Figure 2: Shunt-based overcurrent-detection circuit diagram (Source: Texas Instruments Inc.)
A current-monitoring circuit uses an operational amplifier (op amp) in a differential amplifier configuration across a shunt resistor placed on the low-voltage side of the system. The amplifier measures the current flowing through the shunt resistor. A very small-ohm-value resistor in this type of circuit helps minimize current consumption and maximize efficiency. Because the resistor is small, the amplifier usually needs to provide a high gain, on the order of 20 V/V to 50 V/V.
It is important that the amplifier used in this part of the circuit has low noise and low offset to minimize degradation of the low-voltage signal measured across the shunt resistor. The next portion of the circuit consists of dual comparators configured as a window comparator that detects overcurrent conditions. The dual comparators have trigger levels that correspond to an overcurrent condition. One comparator is set up in an inverted configuration to make it possible to detect current peaks in the shunt resistor in the opposite direction.
When an overcurrent or short-circuit current occurs, the responsible comparator detects it and changes its output state accordingly. Finally, a D-type flip-flop acts as a gate or a switch to disable the system in the event of an overcurrent or short-circuit condition. Selecting the op amp for this current-monitoring circuit is a vital part of the design. For systems that require a high gain, the amplifier bandwidth must be fast enough to support the necessary gain while also having a high-enough slew rate to achieve a fast response time.
The calculation used for determining the required bandwidth for the amplifier in the circuit is:
- Fault response time requirement: T = 1 µs
- Assuming an allocation of 35% of T for the amplifier propagation delay: TRISE = 0.35 µs
- The required amplifier bandwidth: BW = 0.35 / TRISE → BW = 1 MHz
- For a 50-V/V amplifier gain (Equation 1):
GBW = 50 V/V × BW → GBW (min.) = 50 MHz (1)
This calculation indicates that you will need an amplifier with a gain bandwidth of at least 50 MHz to achieve a gain of 50 V/V. Figure 3 compares the rise time of an amplifier with a 10-MHz bandwidth and an amplifier with a 50-MHz bandwidth. The slower amplifier takes approximately 800 ns longer to slew to a high value than the faster amplifier. When a total fault response time of 1 µs is required, you must use an amplifier with a high-enough bandwidth and slew rate for the system to respond quickly.
Figure 3: Amplifier rise-time comparison (Source: Texas Instruments Inc.)
When the circuit requires a high gain, consider a special amplifier known as a decompensated amplifier. Decompensated amplifiers have lower integrated capacitance in the output stage of the amplifier, making them unstable at low gains; however, at high gains, decompensated amplifiers offer several advantages, including a higher open-loop gain, higher bandwidth, and increased slew rates at lower input-referred noise and lower power consumption.
The Texas Instruments OPA607-Q1 is a good example of an AEC-Q100 automotive-qualified decompensated amplifier that provides 50 MHz of gain bandwidth. This amplifier achieves a slew rate of 24 V/µs with an input voltage noise density of 3.8 nV/√Hz and a typical quiescent current of only 900 µA. The OPA607-Q1 also has a low offset voltage of only 120 μV and is rated for ambient temperature operation from −40°C to 125°C.
In automotive motor control systems, there are two main low-side shunt-resistor-based architectures used to measure current. The first uses a single shunt resistor to measure all three legs of the motor inverter system. The second uses three shunt resistors to measure the current in each of the three legs individually. Figure 4 illustrates these two architectures.
Figure 4: Single- and multi-shunt-based current measurement systems (Source: Texas Instruments Inc.)
In the single-shunt-based system, the current measurement circuit comprises a single shunt resistor, a differential amplifier, a windowed comparator, and a D-type flip-flop to measure the entire current of the motor system. This type of measurement is simpler, uses fewer components, takes up less space, and costs less than a multi-shunt-based system. The single-shunt-based system does not provide as much information about the current flow in the system, however. This architecture is slower to respond to faults and cannot indicate in which leg of the inverter the fault occurred.
The multi-shunt-based system employs three shunt resistors, three differential amplifiers, three windowed comparators, and three D-type flip-flops to individually measure the current of each leg of the motor inverter system. While this type of architecture is larger and more costly to implement, it can pinpoint the location of the fault and respond faster in the event of a fault. The multi-shunt-based architecture also provides current information about each leg of the motor inverter system to the microprocessor running the motor control algorithm to control the speed of the motor.
With the advancement of vehicle electrification and hybrid and electric vehicles becoming more mainstream, the number of electric motors and digital power control systems in automobiles is expanding exponentially. To increase efficiency, these modern systems are migrating from slower IGBT- and silicon-FET-based switches to faster SiC- and GaN-based switches.
While faster switching speeds enable greater efficiency, electrical faults such as short-circuits and overcurrent conditions occur at faster rates, driving the need for extremely fast fault detection. To prevent damage to systems and catastrophic failures, faults in most cases must be detected in <1 µs.
One of the most effective and cost-efficient ways to accomplish high-speed current fault monitoring is to employ a low-side multi-shunt-based current-monitoring circuit comprising three shunt resistors, three op amps, three windowed comparators, and three D-type flip-flops. To achieve an overcurrent-condition detection response time of <1 µs, choose an amplifier with a bandwidth of over 50 MHz.
Additional resource
TIDA-060019 High-speed current shunt monitor reference design
The post Detecting overcurrent conditions in EV systems appeared first on Electronic Products.
