1. Introduction Sensors and their related circuits are used to measure various physical characteristics, such as temperature, force, pressure, flow rate, position, and light intensity. These characteristics excite the sensors. The sensor output is conditioned and processed to provide corresponding measurements of the physical characteristics. Digital signal processing utilizes computers or dedicated processing equipment to numerically process signals through acquisition, transformation, estimation, and identification, thereby extracting information and facilitating applications. Instrumentation amplifiers possess superior characteristics, enabling them to amplify very weak sensor signals without distortion for signal acquisition. This paper describes how to use an instrumentation amplifier to condition sensor signals to conform to the operating range of analog-to-digital converters in a smart vibration isolation system where the sensor data acquisition system has numerous sensors with significantly different signal types. 2. Application of Instrumentation Amplifiers in Sensor Signal Conditioning Circuits An instrumentation amplifier is a high-gain, DC-coupled amplifier characterized by differential input, single-ended output, high input impedance, and high common-mode rejection ratio. Differential amplifiers and instrumentation amplifiers use essentially the same basic components (operational amplifiers), but their performance differs significantly from standard operational amplifiers. Standard operational amplifiers are single-ended devices, and their transfer function is primarily determined by the feedback network. Differential amplifiers and instrumentation amplifiers, however, can amplify very weak differential signals even with common-mode signals, thus exhibiting a high common-mode rejection ratio (CMR). They typically do not require an external feedback network. An instrumentation amplifier is a closed-loop gain unit with differential inputs and a single-ended output relative to a reference. Its input impedance is symmetrical and has a large value (typically 10⁹ or greater). Unlike operational amplifiers whose closed-loop gain is determined by an external resistor connected between the inverting input and output, instrumentation amplifiers use an internal feedback resistor network isolated from their signal inputs. The gain is either preset internally or set by the user using the input signals applied to the two differential inputs, either directly or via internally or externally isolated gain resistors. Typical instrumentation amplifier gain settings range from 1 to 1000. Characteristics of instrumentation amplifiers: (1) High common-mode rejection ratio (CMRR): The common-mode rejection ratio (CMRR) is the ratio of differential-mode gain (Ad) to common-mode gain (Ac), i.e., CMRR = 20lg(Ad/Ac) dB. Instrumentation amplifiers have a very high common-mode rejection ratio, with typical CMRR values ​​of 70 to 100 dB or more. (2) High input impedance: Instrumentation amplifiers must have extremely high input impedance. The impedances of the non-inverting and inverting input terminals of the instrumentation amplifier are both very high and very balanced, with typical values ​​of 10⁹ to 10¹² Ω. Low noise: Since instrumentation amplifiers must be able to handle very low input voltages, they cannot add their own noise to the signal. Under 1 kHz conditions, the input noise referred to the input terminal should be less than 10 nV/Hz. (3) Low linearity error: Input offset and proportional coefficient error can be corrected by external adjustment, but linearity error is an inherent defect of the device and cannot be eliminated by external adjustment. A typical linearity error of a high-quality instrumentation amplifier is 0.01%, and some are even lower than 0.0001%. (4) Low offset voltage and offset voltage drift: The offset drift of an instrumentation amplifier also consists of two parts: input and output. The typical values ​​of the input and output offset voltages are 100 uV and 2 mV, respectively. (5) Low input bias current and offset current error: The base current of a bipolar input operational amplifier and the gate current of a FET input operational amplifier will generate an offset error when the bias current flows through the unbalanced signal source resistor. The typical bias current of a bipolar input instrumentation amplifier is 1 nA to 50 pA, while the typical bias current of a FET input instrumentation amplifier at room temperature is 50 pA. (6) Ample bandwidth: Instrumentation amplifiers provide sufficient bandwidth for specific applications. The typical unity-gain small-signal bandwidth is between 500 kHz and 4 MHz. The unique feature of an instrumentation amplifier with a "detect" and a "reference" terminal is that it has a "detect" and a "reference" terminal, which allows for remote detection of the output voltage while minimizing the effects of internal resistor voltage drop and ground voltage drop (IR). For effective operation, the instrumentation amplifier must not only amplify microvolt-level signals but also suppress common-mode signals at its input. This requires the instrumentation amplifier to have a high common-mode rejection ratio (CMR): a typical CMR value is 70–100 dB. CMR typically improves further as gain increases. 3. Current-Type Sensor Data Acquisition System Structure Diagram Figure 1 shows how the signal from a 4–20 mA current-type sensor is connected to the 16-bit Simultaneous ADC AD7656. The 4–20 mA sensor signal is single-ended. This initially necessitates a simple shunt resistor to convert the current into a voltage applied to the high-impedance analog input of the ADC. However, any line resistance in the loop (to the sensor) will increase the current-related offset error. Therefore, the current must be detected differentially. In this system, a single 24.9 Ω shunt resistor generates a maximum differential input voltage between 100 mV (corresponding to 4 mA input) and 500 mV (corresponding to 20 mA input) at the AD627 input. Without a gain resistor, the AD627 amplifies this 500 mV input voltage by a factor of 5 to reach 2.5 V, which is the full-scale input voltage of the ADC. The 4 mA zero-point current corresponds to code 819, and 1 LSB corresponds to 0.61 mV. The entire system logic is controlled by a CPLD and exchanges data with the DSP. 4. Low-Power Instrumentation Amplifier AD627 Features and Performance The AD627 is a low-power instrumentation amplifier. It can be powered by single or dual power supplies and can achieve rail-to-rail output. The AD627 operates normally at 85 uA and has excellent AC and DC characteristics. The AD627 uses an industry-standard 8-pin package, and the pinout is shown in Figure 2. The most significant feature of the AD627 is that it allows the user to set the gain using an external resistor. The AD627 exhibits low offset voltage, offset drift, gain error, and gain drift, thus minimizing DC errors in user systems. Due to its excellent high-frequency common-mode rejection ratio (CMRR), the AD627 maintains minimal high-frequency errors. Its high CMRR (up to 200 Hz) also effectively eliminates transmission line interference and harmonics. The AD627 employs a true instrumentation amplifier architecture with two feedback loops. Its basic structure is similar to a typical dual op-amp instrumentation amplifier, with only minor differences in details. Furthermore, the AD627's current feedback structure contributes to its superior CMRR. The basic circuit of the AD627 is shown in Figure 3. A1, V1, and R5 form the first feedback loop, through which a stable collector current is obtained on Q1 (assuming the gain setting resistor is not present). The feedback loop formed by resistors R1 and R2 ensures that the output voltage of A1 is equal to the reverse voltage. A2 forms another almost identical feedback loop, making the current of Q2 equal to that of Q1, while also providing the output voltage. When the two loops are balanced, the gain from the non-inverting input to VOUT is 5, the gain from A1 output to VOUT is -4, and the gain at the inverting input of A1 is 1.25 times the gain of A2. The gain of the AD627 in differential mode is 1 + R4/R3, with a rated value of 5. The AD627's gain is set via resistor RG. The gain G can be determined by the following formula: G = 5 + (200 kΩ/RG). It can be seen that the minimum gain of the AD627 is 5 (when RG = ∞). When its gain accuracy is between 0.05% and 0.7%, a 0.1% external gain setting resistor should be used to avoid significant attenuation of the full gain error. Alternatively, the gain setting resistor RG can be selected from the standard setting resistor table, choosing the closest value. The AD627 instrumentation amplifier, with its dual-supply operation and parallel detection, offers better DC/AC performance than discrete components and allows for easy gain setting using external resistors, making it a superior choice for sensor signal detection. 5. Instrumentation Amplifier RFI Suppression Circuit Design The low-power instrumentation amplifier AD627 is susceptible to RF rectification, requiring a more robust filter. The AD627 features low input stage operating current. Simply increasing the values ​​of the two input resistors R1a and R1b, or the capacitor C2, would provide further RF attenuation at the cost of reduced signal bandwidth. Because the AD627 instrumentation amplifier has higher noise (38 nV/Hz) than general-purpose ICs (e.g., the AD620 series), higher input resistors can be used without significantly degrading the circuit's noise performance. To utilize higher input resistor values, an RC RFI circuit was designed, as shown in Figure 4. The filter bandwidth is approximately 200 Hz. With a gain of 100, applying a 1 Vp-p input signal within the 1 Hz to 20 MHz input range, the maximum DC offset drift of the RTI is approximately 400 uV. Under the same gain conditions, the circuit's RF signal suppression capability (RF amplitude at output / RF amplitude applied to input) is better than 61 dB. As shown in Figure 4: 6. Introduction to the ADG707 Differential Analog Multiplexer The ADG707 is an 8-to-1 differential input analog multiplexer with low on-resistance as low as 2.5 Ω, a 40 ns switching time, and a low-voltage supply of +1.8 to +5.5 V. It has a wide range of applications in video and audio switching, data holding systems, and communication systems. This system uses a 3.3 V supply to meet the overall system power distribution requirements. Since the sensor signals used in this system are small signals, they meet the operating requirements of the ADG707. 7 Circuit Configuration of AD7656 The signal of the current-type sensor is converted into a voltage signal through the instrumentation amplifier conditioning circuit mentioned above. The voltage-type sensor signal can be directly input to AD7656 through an operational amplifier (e.g., AD8021). This system uses a 16-bit ADC AD7656, which can meet the high precision requirements of the system. At the same time, the update frequency of the sensor signals used in the system is relatively low, with a maximum of no more than 20 kHz. The sampling frequency of AD7656 is 250 kb/s, which obviously meets the requirements. AD7656 can perform 6-channel synchronous sampling, which provides a lot of leeway for expanding the number of sensors. The circuit configuration of AD7656 is shown in Figure 5: 8 Conclusion Design Considerations In the circuit design of the instrumentation amplifier, the following practical issues need to be considered: (1) The gain of AD627 is achieved by changing the programming resistor RG. In order to make the output voltage gain of AD627 accurate, a resistor with an error of less than 0.1% to 1% should be used; at the same time, in order to maintain the high stability of the gain and avoid high gain drift, a resistor with a low temperature coefficient should be selected. (2) Since the output voltage of AD627 is relative to the reference terminal, the REF pin should be connected to a low impedance point to obtain a high common-mode rejection ratio. (3) All instrumentation amplifiers can rectify high-frequency signals outside the passband; after rectification, these signals manifest as DC offset errors in the output. A low-pass filter can be designed to prevent unwanted noise from reaching the differential input. In many applications, shielded cables are used to reduce noise; to obtain the best common-mode rejection ratio across the entire frequency range, the shielding layer must be properly connected. In this paper, based on my actual work, I have detailed the design of a sensor signal conditioning circuit based on an instrumentation amplifier, analyzed the problems that are easy to encounter, and provided an effective solution from an engineering perspective.