Designing sensor signal conditioning electronics and related calibration algorithms is no easy task and is time-consuming. Using a hardware sensor simulator can significantly shorten development time. For example, a major reason for the time-consuming temperature calibration is the long temperature cycling time in the environmental chamber, which can take several hours to stabilize; a typical three-point calibration (-55°C, 25°C, and 85°C) can take eight hours. Sensor simulators already have models of sensor temperature drift set up, allowing control engineers to complete simulation calibration in just a few minutes. This allows for many calibrations to be performed in a single day, saving time so that attention can be focused on debugging the sensor hardware or developing calibration algorithms. Another issue with sensors is that some require expensive equipment to provide excitation (e.g., humidity, acceleration, pH, pressure, and tension). In many cases, access to this equipment is not readily available. Using a sensor simulator allows for debugging of the hardware and software before using such equipment, improving the efficiency of laboratory research. Furthermore, the equipment used to generate excitation is sometimes developed for a specific sensor and may not be validated. For example, Contrivance Engineering developed a custom mechanical calibration system to apply torque to the sensor. Since sensors are developed concurrently with calibration systems, it is difficult to debug them without a sensor simulator. Furthermore, if mechanical oscillations exist in the calibration series from the outset, it is difficult to determine whether the oscillations originate from the mechanical or electrical systems without a simulator. Sensors may also exhibit non-repeatable issues, making it difficult to determine whether the problem lies with the sensor itself or its performance limitations. For example, pressure sensors, in addition to delays, may exhibit pressure hysteresis or temperature hysteresis. Because sensors and electronics are typically housed in sealed enclosures, it is difficult to determine whether the fault originates from the sensor or the electronics. Sensor simulators do not exhibit non-repeatability, thus allowing for the evaluation of the accuracy of the sensor electronics. Before detailing how to implement a sensor simulator, it can be considered a black box. The sensor simulator presented in this paper is used to simulate a four-element Wheatstone bridge sensor. Since most bridge sensors provide associated temperature sensors, it can emulate the effects of two different types of temperature sensors (diode and series resistor). Figures 1 (a and b) show examples of bridge sensors to be simulated. The example bridge sensor, PGA309, is a typical signal conditioning system-on-chip that can be used to compensate for temperature drift and nonlinearity in resistive bridge sensors. Figure 1a shows an example of a bridge sensor, whose temperature coefficient is used to determine the resistance voltage (Rt). The voltage across Rt is the temperature signal. The simulator can simulate the Rt temperature sensing method, providing a programmable temperature signal that modulates the bridge voltage. The sensor simulator can also simulate the circuit in Figure 1b, where diodes are used to measure the bridge temperature. Figure 2 illustrates the sensor's temperature drift. Figure 3 shows the relationship between the sensor output signal and the excitation. Figure 2 depicts the sensor's temperature drift. The sensor simulator models the drift at three different temperatures (room temperature, hot, and cold). Figure 3 shows the relationship between the sensor output signal and the excitation. Typically, this signal only includes second-order nonlinearity. The sensor simulator simulates the sensor response with five discrete points at room temperature and three discrete points at both hot and cold temperatures. Sensor temperature drift, output signal, and excitation relationship. Sensor signal conditioning systems typically modulate the excitation voltage to correct for sensor nonlinearity. Since Vexc adjustment can be used to correct for nonlinearity of the applied pressure, it can be used to model real-world sensors; that is, changes in the sensor simulator input Vexc directly affect the simulated bridge output. Figure 4 illustrates a sensor simulator viewed as a black box. Note the Vexc input, sensor output, and temperature signal output, as well as the control of the temperature and sensor outputs. The temperature output control has three different temperatures (room temperature, high temperature, and low temperature). The sensor output has the following outputs: 0, 50%, and 100% low temperature; 0, 25, 50, 75, and 100% room temperature; and 0, 50, and 100% high temperature. The sensor simulator can be considered a black box. Some engineers might ask why increasing accuracy doesn't solve the problem, for example, by using a precision power supply (such as a millivolt calibrator) to simulate the sensor output. The main problem with using a voltage source to simulate the sensor output is that it cannot be modulated by the sensor's excitation voltage. The sensor's electronics typically correct for nonlinearity by changing the sensor's excitation voltage, which also changes when the sensor is used for radiometric titration. In this mode, the sensor and electronics share a single power supply. Testing radiometric titration power supply suppression is difficult to achieve with a precise power supply. When simulating the sensor, three precise power supplies are needed: one for the common-mode signal, one for the differential signal, and another for the temperature signal (see Figure 5). This setup is more expensive than the simulator approach suggested in this paper, and it requires reconfiguring the output configuration for each sensor. The simulator output configuration, however, only needs to be set once and is selected via a rotary switch. Figure 5 shows one of the three precision voltage sources required for the temperature signal power supply simulation sensor. A simple implementation of the sensor simulator: There are several ways to implement a Wheatstone bridge sensor simulator. The approach presented here is very intuitive, using a shaping potentiometer and a rotary switch. More complex methods can be used with a D/A converter, microcontroller, PC interface, and related software. Both methods have their advantages, and the intuitive approach avoids the need for software. Figure 5 shows a single channel of the sensor simulator. The complete design uses eleven channels and a rotary switch to produce eleven unique output states. These eleven different output states are typically used to simulate the sensor output at three different temperatures under different excitations. This configuration is used because most commonly used sensor calibration algorithms require three different temperatures and three different levels of excitation. Differential signals are generated by adjusting R8 (R9 for fine-tuning). Using the components shown in the figure, the circuit output range is ±25mV with an excitation voltage of 5V. The measurement stability over several hours is approximately 0.03%. The output range of this circuit can be adjusted by changing R7 and R10. For example, with an excitation voltage of 5V, using a 1kΩ resistor can change the range to ±250mV. This allows the circuit to simulate sensor outputs of different ranges, improving accuracy and repeatability. The sensor simulator can also simulate temperature output signals. Most sensors incorporate simple temperature sensors to monitor bridge sensor temperature. As mentioned earlier, sensors typically use diodes to generate temperature signals or utilize the temperature coefficient of the bridge resistor (the Rt method). Figure 6 illustrates how the Rt temperature signal is generated. R2 and R3 are used to simulate the temperature coefficient of the bridge resistor, and R4 is used for the temperature sensor resistor Rt. The same resistive voltage divider can be used to generate room temperature, high temperature, and low temperature signals. Another advantage of this circuit is that the Rt temperature signal can be modulated by the sensor's excitation voltage. U3 and U4 buffer the temperature output signal, which can be used to adjust the sensor output signal, thus making the sensor output signal equivalent to a series bridge. The diode method for temperature sensing can be achieved simply by replacing the diode with a resistive voltage divider. Furthermore, the same circuit can be used with a rotary switch to generate room temperature, high temperature, and low temperature signals.