The electrification of vehicles has spurred an increased demand in accurate current measurement in new electronic drive systems, such as traction inverters, DC-DC controllers, battery managements systems and electronic power steering. Autonomous vehicles are much closer on the horizon, and new advanced driver assistance systems (ADAS) will be critical to their safe operation. These systems include light detection and ranging (LIDAR), radar and ultrasonic systems, as well as surround-view cameras. Knowing that these systems are operating within the correct operational guidelines will be vital to helping ensure vehicle safety. Just like in the electrified power-train, accurate measurement of current is one method to improve ADAS design margins.
Standard amplifiers vs. current sense amplifiers
To maximize system performance vs. system-level cost trade-offs, it’s important to have a good understanding of data-sheet parameters and their effect on the current measurement. For many current sensing applications, especially low-side measurements, designers have multiple amplifier choices. The two most popular are general-purpose operational amplifiers (op amps) and current sense amplifiers (also known as current shunt monitors).
While there are a few fundamental differences between the two types, the main difference is that current sense amplifiers integrate the gain resistor network, while op amps use external discrete resistors for their gain network (Figure 1).
Comparison between the gain network of a discrete current sensing implementation and an integrated current sense amplifier
The inclusion of the gain network affects how you should interpret data-sheet parameters and how they relate to your application requirements. The four specifications most affected by the inclusion of the gain network are:
- Gain error.
- Input offset voltage (VOS ) and the related temperature drift.
- Common-mode rejection ratio (CMRR).
- Small-signal bandwidth.
The gain accuracy of the circuit is determined by the ratio and matching of the resistor network. The gain (G) of either implementation (Figure 2) is the ratio of the feedback resistor (RF ) over the input resistor (RIN ), as shown in Equation 1:
A current sense amplifier integrates an op amp with the gain resistors in a differential configuration
To maximize the accuracy, you need to ensure that the “error” of the two resistors is identical. If RF is off by 10%, as long as RIN is off by an identical 10%, then the gain remains at an ideal value, as shown here:
In a differential configuration, it is also critical that you match each RIN and each RF to prevent an imbalance on the differential signal. The method of manufacturing gain resistors in current sense amplifiers (like the Texas Instruments [TI] INA190) creates resistors that are nearly identically matched.
The gain error in the data sheet of such a device is actually a measure of the matching of resistors. For the INA190A5, for example, the maximum gain error is +/-0.4%; typically the resistors are less than +/-0.01% mismatched (Figure 3).
Extract from the TI INA190 data sheet highlighting the gain error specification
The gain error for an op-amp circuit is derived by the external resistors. Assuming that each resistor is a +/-1% resistor, then Equation 2 derives the worst-case formula as resulting in a 2% error – twice that of any individual resistor:
In addition to the initial room-temperature error, the temperature drift (or temperature coefficient [tempco] for discrete resistors) can further exacerbate the external gain because the discrete resistors will not likely drift at the same rate, further deteriorating the gain accuracy due to resistor mismatching. A typical discrete resistor has a 100ppm/o C tempco. An increase from 25o C to 125o C results in an additional error of 1% per resistor, which means that the worst-case gain error for the 1% resistor network is now 4%.
The manufacturing of a current sense amplifier minimizes the drift differences between resistors, leading to very low gain temperature, as shown in Figure 4 for the INA190.
Extract from the TI INA190 data sheet highlighting the gain error drift specification
The INA190A5 will have a worst-case gain error over temperature of 0.45% (5ppm/o C × 100o C = 0.05%).
VOS and CMRR
Understanding the differences between the two implementations shown in Figure 2 in VOS and CMRR is not as straightforward as the gain comparison. The key is to understand where the data-sheet parameter is specified in the circuit.
For an op-amp implementation, both VOS and CMRR are specified at the white + and – circles (Figure 2). For a current sense amplifier, VOS and CMRR are specified at the gray + and – circles (Figure 2). In an op-amp implementation, the measurement does not include variations in the discrete gain resistors.
Because the tolerance and temperature drift discussed previously will negatively impact the VOS and CMRR achieved by the op-amp implementation, you must factor in both as additional error. For a current sense amplifier, the variation and drift of the resistive elements are already included in the specification, and you can use the data-sheet specifications directly in your error calculations.
The bandwidth is normally specified at unity gain in an op amp. As you increase the gain, you must typically decrease the gain by a corresponding factor. In other words, the bandwidth specified is really a gain-bandwidth product specification, as defined by Equation 3:
For example, if you have an op amp with a unity gain of 2MHz and you put it in a circuit with a gain of 20, then the small-signal bandwidth will be 100kHz.
Since current sense amplifiers include a gain network, the bandwidth specified in the data sheet is at the specific gain; in many cases, the bandwidth will change with gain. Figure 5 shows how the TI INA2180 specifies small-signal bandwidth.
Extract from the TI INA2180 data sheet highlighting the small-signal bandwidth specification
It is critical to understand how gain affects the small-signal bandwidth in order to ensure that the circuit can support the necessary frequency.
The electrification and autonomy of vehicles is creating a demand for precision current sensing in systems such as battery management systems, motor current systems and ADAS and so on. Matching performance requirements to application needs can be challenging.
About the author
Dan Harmon is the automotive marketing manager for the Current and Magnetic Sensing product line at TI. In his 30-plus-year career, he has supported a wide variety of technologies and products including interface products, imaging analog front ends and charge-coupled device sensors. He has also served as TI’s USB Implementers Forum representative and TI’s USB 3.0 Promoter’s Group chairman. Dan has a bachelor’s degree in electrical engineering from the University of Dayton and a master’s degree in electrical engineering from the University of Texas at Arlington. If you have questions about this article, you can ask a new question in the TI E2E Community Amplifiers forum.