Industrial motion control encompasses a range of applications, including inverter-based fan or pump control, factory automation with more complex AC drive control, and advanced automation applications such as robots with advanced servo control. These systems require the detection and feedback of multiple variables, such as motor winding current or voltage, DC link current or voltage, rotor position, and speed. Among considerations such as value-added features (e.g., condition monitoring), end-application requirements, system architecture, target system cost, or system complexity will determine the selection of variables and the required measurement accuracy. With motors reportedly accounting for 40% of global energy consumption and international regulations increasingly emphasizing system efficiency across industrial motion applications, these variables, particularly current and voltage, are becoming increasingly important.
This article will focus on current and voltage detection in various motor control signal chain topologies, based on motor rated power, system performance requirements, and end applications. In this context, the implementation of the motor control signal chain will vary depending on sensor selection, current isolation requirements, analog-to-digital converter (ADC) selection, system integration, system power consumption, and grounding configuration.
Industrial drive application map
From simple inverters to complex servo drives, motor control applications encompass a range of motor types, but all motors include motor control systems at specific power levels, and processors with varying levels of sensing and feedback that drive pulse-width modulator (PWM) modules. Figure 1 is a simplified diagram of the application landscape, showing systems with increasing complexity from left to right. It begins with simple control systems, such as pumps, fans, and compressors that can be implemented using only a simple microprocessor without requiring precise feedback. As system complexity increases (i.e., moving to the higher end of the landscape), complex control systems require precise feedback and high-speed communication interfaces. Examples include sensor-controlled induction motors or permanent magnet motors, with or without sensors, and high-power industrial drives (such as large pumps, fans, and compressors) designed for the efficiencies shown in Figure 1. At the very top of the landscape are complex servo drives used in applications such as robotics, machine tools, and surface mount technology (SMT) machines. As system complexity increases, the sensing and feedback of variables become increasingly critical.
Driver architecture system partitioning
We may encounter various problems when designing systems that meet the needs of various industrial motion control applications. A general motor control signal chain is shown in Figure 2.
Isolation requirements are critical and often have a significant impact on the resulting circuit topology and architecture. Two key factors need to be considered: the reason for isolation and its location.
The isolation classification requirements depend on the former. It might require safety high voltage isolation (SELV) to prevent electric shock, functional isolation for level transitions between non-lethal voltages, or isolation for data integrity and noise elimination. The isolation location is typically determined by the system's expected performance. Motor control often operates in harsh environments filled with electrical noise, and designs typically withstand hundreds of volts of common-mode voltage, may switch at frequencies exceeding 20 kHz, and have extremely high transient dv/dt rise times. Therefore, high-performance systems and inherently noisy, high-power systems are often designed with power stages isolated from the control stage. Whether a single-processor or dual-processor design is used also affects the isolation location. In lower-performance, low-power systems, isolation is usually at the digital communication interface, meaning the power stage and control stage are at the same potential. Low-end systems require isolated communication interfaces with lower bandwidth. Isolating communication ports in high-end systems is often difficult due to the higher bandwidth requirements of high-end systems and the limitations of traditional isolation techniques. However, this is changing with the advent of magnetically isolated CAN and RS-485 transceivers.
In high-performance closed-loop motor control design, two key components constitute the PWM modulator output and the motor phase current feedback. Figures 3a and 3b illustrate the locations where safety isolation is required, depending on whether the control stage shares the same potential with the power stage or is ground-based. Regardless of the location, the high-side gate driver and current sensing node require isolation; however, the isolation levels differ in Figure 3a, where only functional isolation is required, while in Figure 3b, personnel safety isolation (i.e., current isolation) is crucial for these nodes.
Measurement techniques and topologies for current and voltage sensing
In addition to the system power and grounding considerations mentioned above, the signal chain implemented for current and voltage detection will vary depending on sensor selection, current isolation requirements, ADC selection, and system integration. Signal conditioning for high-fidelity measurements is not easy. For example, recovering small signals or transmitting digital signals in such noisy environments is very challenging, while isolating analog signals is an even greater challenge. In many cases, signal isolation circuitry can cause phase delays that limit system dynamics. Phase current detection is particularly difficult because the circuit node connected to this node is the same as the node at the gate driver output in the power stage (inverter module) core, and therefore the requirements for power supply isolation and switching transients are also the same. The measurement signal chain (technology, signal conditioning, and ADC) to be implemented in a motor control system is typically determined based on the following three key factors:
1. Determine the points or nodes in the system that require measurement.
2. Motor power level and the final selected sensor (whether it has isolation capabilities). Sensor selection greatly influences the choice of ADC, including converter architecture, functionality, and analog input range.
3. End Applications. This can drive the need for higher resolution, accuracy, or speed in the sensing signal chain. For example, achieving sensorless control over a wider speed range requires more frequent and precise measurements. End applications also influence the requirements for ADC functionality. For instance, multi-axis control may require an ADC with a higher channel count.
Current and voltage sensors
The most commonly used current sensors in motor control are shunt resistors, Hall effect (HE) sensors, and current transformers (CTs). While shunt resistors offer isolation and exhibit losses at higher currents, they are the most linear, lowest-cost, and suitable for both AC and DC measurements. Signal level attenuation due to shunt power loss typically limits shunt applications to 50A or lower. CT and HE sensors offer inherent isolation, enabling their use in higher current systems. However, they are more expensive, and solutions using these sensors are less accurate than those using shunt resistors due to their inherently lower initial accuracy or temperature-dependent accuracy.
Besides sensor types, there are many other options for motor current measurement nodes. Average DC link current is sufficient for control requirements, but in more advanced drivers, motor winding current is used as the primary feedback variable. Direct-phase winding current measurement is ideal for high-performance systems. However, winding current can be measured indirectly by using shunts on each low-side inverter pin or a single shunt in the DC link. The advantage of these methods is that the shunt signals are all referenced to a common power supply, but extracting winding current from the DC link requires sampling synchronized with the PWM switching. Direct-phase winding current measurement can be performed using any of the above current sensing techniques, but the shunt resistor signal must be isolated. High common-mode amplifiers provide functional isolation, but personnel safety isolation must be provided by isolated amplifiers or isolated modulators.
Figure 4 illustrates the various current feedback options described above. While only one option is needed for control feedback, the DC link current signal can also be used as a backup signal for protection.
As mentioned earlier, system power and grounding configuration will determine the required isolation classification and thus the appropriate feedback. The system's target performance will also influence sensor selection or measurement techniques. Looking at the overall performance profile, many configurations are also possible.
Low-performance example: Power stage and control stage on a common potential, detection option A or B.
Using pin shunt is the most cost-effective technique for measuring motor current. In this example, the power stage and control stage share the same potential, there is no common-mode to be processed, and the output of option A or option B can be directly connected to the signal conditioning circuit and ADC. This type of topology is common in low-power and low-performance systems where the ADC is embedded in the microprocessor.
High-performance example: Control stage grounding, detection options C, D, or E
In this example, personnel safety isolation is required. Detection options C, D, and E are all possible. Of all three options, option E offers the best current feedback and, as a high-performance system, may include an FPGA or other form of processing to provide digital filters suitable for the isolated modulator signal. For the ADC selection in option C, a discrete isolated sensor (most likely a closed-loop HE) is typically used to achieve higher performance than current embedded ADC products. Option D in this configuration is an isolated amplifier compared to a common-mode amplifier because safety isolation is required. Isolated amplifiers limit performance, so an embedded ADC solution is necessary. This option offers the lowest fidelity current feedback compared to options C or E. Furthermore, while an embedded ADC can be considered "free" and an isolated amplifier "cheap," implementation typically requires additional components for offset compensation and level shifting for ADC input range matching, increasing the overall cost of the signal chain.
In motor control design, many topologies can be used to sense motor current, and various factors such as cost, power level, and performance level must be considered. A key objective for most system designers is to improve current sensing feedback to increase efficiency within their cost targets. For higher-end applications, current feedback is crucial not only for efficiency but also for other system performance measurements such as dynamic response, noise, or torque ripple. Clearly, among the various available topologies, there exists a continuum of performance from low to high; Figure 5 is a rough mapping showing low-power and high-power options.
Goals, requirements, and development trends for motor control system designers: From HE sensors to shunt resistors
A shunt resistor coupled to an isolated Σ-Δ modulator provides the best current feedback, where the current level is low enough to fully meet shunt requirements. Currently, system designers are clearly inclined to switch from HE sensors to shunt resistors, and also prefer isolated modulator solutions over isolated amplifier solutions. Simply replacing the sensor itself can reduce bill of materials (BOM) and PCB assembly costs and improve sensor accuracy. Shunt resistors are insensitive to magnetic fields or mechanical vibrations. System designers replacing HE sensors with shunt resistors often choose isolated amplifiers and continue using the ADC previously used in HE sensor-based designs to limit level variations in the signal chain. However, as mentioned earlier, regardless of ADC performance, this performance will be limited by the performance of the isolated amplifier.
Furthermore, replacing the isolated amplifier and ADC with an isolated Σ-Δ modulator eliminates performance bottlenecks and significantly improves the design, typically boosting its superior feedback from 9 to 10 bits to 12 bits. Additionally, digital filters required to process the Σ-Δ modulator output can be configured to enable a fast OCP loop, thus eliminating analog overcurrent protection (OCP) circuitry. Therefore, any BOM analysis should include not only the isolated amplifier, the original ADC, and signal conditioning between them, but also the OCP-eliminating device. The AD701A isolated Σ-Δ modulator, based on Analog Devices' iCoupler® technology, features a differential input range of ±250mV (typically ±320mV full scale for OCP), making it particularly suitable for resistive shunt measurements and an ideal choice for expanding this trend. The analog modulator continuously samples the analog input, while the input information is contained within the digital output stream in the form of a data stream density, with data rates up to 20MHz. The original information can be reconstructed using appropriate digital filters, typically Sinc3 filters suitable for precision current measurements. Because a trade-off can be made between conversion performance and bandwidth or filter group delay, simpler and faster filters can provide fast OCP response on the order of 2μs, making them ideal for IGBT protection.
The need to reduce the size of shunt resistors
From a signal measurement perspective, some of the main challenges currently involve selecting shunt resistors, as a balance needs to be struck between sensitivity and power consumption. Larger resistance values will ensure maximum dynamic range across the entire or as large a range of analog inputs using a Σ-Δ modulator. However, due to I²×R losses in the resistor, large resistance values also lead to voltage drops and reduced efficiency. The nonlinearity caused by the resistor's own heating effect is another challenge when using larger resistors. Therefore, system designers face trade-offs and potential consequences, often needing to select an appropriately sized shunt resistor to meet the needs of various models and motors at different current levels. Maintaining dynamic range is also challenging when dealing with peak currents several times the motor's rated current and reliably capturing both values. The ability of a control system to handle peak start-up current varies considerably depending on the design, ranging from strict controls such as 30% above the rated current to coefficients as high as 10 times the rated current. Acceleration and changes in load or torque also generate peak currents. However, peak currents in a system are typically within four times the driver's design rated current.
Faced with these challenges, system designers are seeking high-performance Σ-Δ modulators with wider dynamic range or higher signal-to-noise ratio and signal-to-speech ratio (SINAD). The latest isolated Σ-Δ modulator products offer 16-bit resolution and ensure performance up to 12 effective bits (ENOB).
SINAD = (6.02N + 1.76) dB, where N = ENOB
Following the trend of using shunt resistors in low-power drivers, motor driver manufacturers are also trying to increase the power rating of drivers that can utilize this topology, for both performance and cost reasons. The only feasible way is to use shunt resistors with smaller resistance values, which requires the introduction of higher-performance modulator cores to identify weakened signal amplitudes.
System designers (especially servo designers) are constantly exploring ways to improve system response, such as reducing analog-to-digital conversion time or decreasing group delay by employing digital filters with isolated Σ-Δ modulators and shunt resistor topologies. As mentioned earlier, trade-offs can be made between conversion performance and bandwidth or filter group delay. Simpler, faster filters can provide a faster response but at the cost of reduced performance. System designers analyze the effects of filter wavelengths or decimation ratios and then weigh the trade-offs based on their end application requirements. Increasing the modulator's clock rate can be helpful, but many designers have already implemented operation at the maximum clock rate of 20 MHz supported by the AD7401A. One drawback of increasing the clock rate is radiated potential and interference (EMI) effects. At the same clock rate, a higher-performance modulator can improve the trade-off between group delay and performance, resulting in a faster response time with less impact on performance.
Industry-leading isolated Σ-Δ modulator
Clearly, by reducing the size of the shunt resistor, improving the sensorless control scheme, and enabling control of high-efficiency internal permanent magnet motors (IPMs), higher-performance isolated Σ-Δ modulators can meet various needs and development requirements in industrial motor design and improve the efficiency of motor drivers. Analog Devices' AD7403, a new generation of the AD7401A, offers a wider dynamic range at the same 20MHz external clock rate. This allows designers greater flexibility in selecting the shunt resistor size, optimizing driver-motor matching, improving the accuracy of rated and peak current measurements, reducing the impact of a single shunt resistor size applicable to a range of motor models, and enabling the replacement of HE sensors with shunt resistors at higher current levels. Furthermore, it improves dynamic response by reducing measurement latency. Compared to the previous generation AD7400A and AD7401A, the AD7403's isolation scheme also allows for the use of a higher continuous operating voltage (VIORM), thereby improving system efficiency by using a higher DC bus voltage and lower motor current.
A wider range of system solutions, including the ADSP-CM40x mixed-signal control processor.
As mentioned earlier, implementing a Σ-Δ modulator requires a digital filter in the system. This is typically achieved using an FPGA or digital ASIC. The advent of the ADSP-CM408F mixed-signal control processor (which includes Sinc3 filter hardware and can be directly connected to isolated Σ-Δ modulators from the AD740x series) has the potential to accelerate the adoption of resistive shunt current sensing technology coupled to isolated Σ-Δ modulators. As discussed in this article, resistive shunt current sensing technology has historically been considered expensive by designers due to its increased complexity in the digital domain and associated (FPGA) costs. The ADSP-CM408F offers a cost-effective solution that allows many designers previously constrained by cost to consider this technology.
