As power electronics switching techniques advance, there is an increased need to accurately measure current for feedback control and system monitoring. There are several methods to achieve this, and each method has advantages and limitations. This article focuses on current measurement methods that require relatively high accuracy and bandwidth, such as those used to measure current on the input of computer/telecom boards, inverter phase currents, and other circuits carrying currents from a few amperes up to 100 A.
In particular, this article will review specific details on how to measure currents with Hall-effect current sensors, as well as sensors which use a current mirror that has been integrated into a load switch or an electronic fuse (e-fuse) device. These methods will be compared to current-sense architecture that uses shunt resistors.
Historically, current shunts and current transformers were considered the best methods to sense currents in electric and electronic circuits. However, these methods have significant drawbacks, as current shunts require a compromise between the signal-to-noise ratio (SNR) and power losses in the sensing element. This tradeoff makes it difficult to take accurate measurements across wide current ranges. In addition, current transformers are typically large, expensive solutions, and are only well-suited to measure AC currents.
Current measurement methods
Advances in semiconductor technology have introduced Hall-effect sensors and current mirrors, which are practically lossless current-sense devices with an output that is easy to scale for optimal SNR. This article focuses on these solutions.
- Current mirror sensing method
Current mirror sensing is typically used with devices that have internal power MOSFETs, such as smart power stages, load switches and e-fuses. This approach uses a few cells of the power FET to act as a current mirror, which generates a current output that is proportional to the current flowing through the main switch.
When this current flows through an external resistor, it is simple to create a voltage proportional to the current flowing through the main FET. Half-bridge power stages are ideal devices that use this measuring method, as they provide wide current capabilities (10 A to 90 A) and a current mirror output with a 5 µA/A to 10 µA/A gain. These power stages are particularly useful for applications with synchronous buck regulators in single-phase or multi-phase configurations.
For other applications, it is advantageous to have a load switch or e-fuse circuit that can protect downstream electronics from inrush current or overload conditions. Such circuits typically deploy power MOSFETs as a switch element, so current mirrors are a cost-effective way to monitor a current through these devices. One example is MP5921, a hot-swap e-fuse device that provides several levels of protection along with current and temperature monitoring capabilities (Figure 1).
Figure 1 The e-fuse device provides several levels of protection while monitoring current. Source: Monolithic Power Systems
This e-fuse device can support sustained currents of up to 50 A in standalone mode, and can support even higher currents when placed in a parallel configuration with multiple devices (Figure 2).
Figure 2 The parallel operation in e-fuse device allows it to support currents even higher than 50 A. Source: Monolithic Power Systems
The MP5921 device can operate in conjunction with the MP5920 device, which allows the current signal to be converted into a digital format monitored via the I2C or PMBus (Figure 3). Such devices can support current ratings between 2.5 A and 50 A, and provide different functions such as output discharge, configurable current limit, and slew rate.
Figure 3 Parallel operation can be carried out in conjunction with an e-fuse device and an energy monitoring device. Source: Monolithic Power Systems
These solutions are well-suited for applications related to DC/DC converters on computing and server boards. Other applications, such as those with motor drivers and AC/DC circuits, benefit from current-sense devices that can generate a bidirectional signal consistent with the AC current of a motor driver or power inverter.
- Hall-effect current sensing method
The high voltages in applications such as motor drivers and AC/DC circuits can be dangerous to the surrounding logic-level circuitry. Hall-based current sensors can provide galvanic isolation among the high-voltage, high-current and logic-level circuitry that is used to control the device. It’s achieved by sensing the magnetic field that every current generates. These sensors then output a voltage that is directly proportional to the current.
Current sensors provide high-precision Hall-effect current sensing with isolation voltages of up to 2,200 V and working voltages of up to 280 V. These current sensors can be used in applications employing up to 50 A. One common application for Hall-effect current sensors is inverters.
This article will examine a three-phase inverter, which is ubiquitous in power applications like UPS, motor driver and solar panels. It can be easily generalized to an offline application by removing one of the phases. Typically, the sensors are used to feed back the phase currents to the controller (Figure 4). In Figure 4, the high voltage side is marked red, while the logic voltage is green. The isolation is achieved by galvanically separating the primary side (current-carrying leads) from the secondary side (output leads).
Figure 4 A typical three-phase inverter system comes with isolated phase current sensing. Source: Monolithic Power Systems
Figure 5 shows the lead frame and die of an MCS180x current sensor. In Figure 5, the primary side is copper-colored, and the current flow is represented by the red arrows. Note that in actual applications, current can flow in either direction. The flow of the current creates a magnetic field (shown in blue) according to Ampere’s Law, which states that the field magnitude is directly proportional to the current density.
Figure 5 The primary side is copper-colored while current flow is represented by the red arrows. Source: Monolithic Power Systems
It’s also important to note that the primary does not touch the silicon (shown in black); instead, it is separated by an air gap and silicon oxide insulation. The direction and amplitude of the magnetic field are detected by Hall sensors in the silicon die, and that signal is then amplified and outputted to one of the secondary pins (shown in silver). The other secondary pins serve VCC, GND, and a filter. The filter pin creates a tradeoff between bandwidth and signal noise.
The MCS180x family of current sensors have a ratiometric output, meaning that it’s centered around VCC/2. When a positive current flows, the output voltage rises in proportion to the current from VCC/2 up to VCC. When a negative current flows, the output falls from VCC/2 to 0 V (Figure 6). Here, IPMAX is indicated by the part number suffix; for instance, the MCS1802-10 has an IPMAX of 10 A. It allows designers to choose an appropriate current range for their application. Available options are 5 A, 10 A, 20 A, 30 A, 40 A, and 50 A IPMAX.
Figure 6 Ratiometric output of a Hall-effect current sensor is centered around VCC/2. Source: Monolithic Power Systems
Choose the best current-sense solution
The article highlighted available options for accurate current measurement, and clarified how to select the best current-sense solution for an application. It’s important to consider technological options early in the design process, and to note the specifics of each solution, such as power supply requirements and the achievable output signal range, to create the most efficient solution.
The MCS180x Hall-effect sensors from MPS were used to demonstrate how current-mirroring technology helps designers improve system-level performance while reducing power losses and system complexity.
Ted Smith and Oleg Volfson are senior FAEs at Monolithic Power Systems (MPS).