Today’s data-acquisition (DAQ) systems are a central element to more than just industrial applications. They’re usually used for sensor-based measurements of temperature, flow, fill level, pressure, and other physical quantities, which are then converted to high–resolution digital information and communicated further for processing via software. Such systems require increasingly more precision. As a result, developers often wind up struggling to unite properties that have negative impacts to the system, such as signal noise and drift with requirements for high conversion and transmission rates.
High input impedances are typically required to directly connect different sensor types with correspondingly different analog signal outputs. In addition, the inputs should be able to buffer, amplify, and adjust levels of input signals. Or they also must be capable of generating differential signals to cover the complete voltage range of the analog-to-digital converter (ADC) inputs and simultaneously meet their common-mode voltage requirements. However, the original measurement signal should remain as undistorted as possible.
The input stage is, thus, one of the decisive factors for determining the overall accuracy of data-acquisition systems. Programmable-gain instrumentation amplifiers (PGIAs) are typically used for this purpose, which is how gain is usually adjusted via external resistors; the outputs are directly coupled to the inputs of a downstream ADC.
PGIAs are commonly equipped with single-ended outputs and hence can’t be used to drive fully differential successive-approximation-register (SAR) ADCs directly. Therefore, an additional signal–conditioning or driver stage is needed. However, the additional driver stage affects the performance of the overall DAQ system because further error components can be introduced through it. Good performance is achievable with the right selection of components (see figure).
The figure shows a simplified circuit for a DAQ system that contains a reference voltage source and a reference buffer with integrated power supply, as well as a PGIA and an AD4020 SAR ADC. The differential outputs of the PGIA consist of discrete standard components for digitally programmable gain. It features an input impedance in the gigaohm range, a common-mode rejection ratio of over 92 dB, low output noise, and low distortion. This makes it suitable for direct control of the SAR ADC without loss of performance.
The PGIA drives the AD4020, a 20-bit, 1.8 MS/s, low-power SAR ADC. The AD4020 has a number of other functions that can be used to reduce the complexity of the complete signal chain and increase channel density without detracting from performance. Additional functions include, for example, a high-impedance mode for reducing nonlinear input currents coupled with a long detection phase for direct connection of the PGIA with a simple RC filter in between.
The high sampling rate of the AD4020 enables precise acquisition of high-frequency signals up to several hundred kilohertz. It also allows for decimation so that the dynamic range can be expanded for the precise detection of low voltage signals. Moreover, the demands on the antialiasing filter can be reduced.
The SPI interface, which is compatible with different logic levels (1.8, 2.5, 3, and 5 V), can be programmed in many ways and offers both read and write functions.
With the components shown in the figure, the circuit offers a good linearity (INL) of typically ±2 ppm, low offset and gain drifts (±3.5 ppm/°C and ±6 ppm/°C, respectively), and a good noise power of over –115 dB, all at the full conversion rate and over the entire gain range. The circuit enables both bipolar and unipolar single-ended or fully differential input signals up to ±10 V for gains of 1 to 10. An overview of the input voltage range as a function of gain is given in the table below.
The circuit also offers calibration options for larger PGIA ranges. This function offers precise ratiometric performance and simplifies the system design by already containing options for signal buffering, amplification and attenuation, common-mode level shifting, and various other functions for overcoming the challenges in analog signal processing.
With the high-impedance input and the programmable gain, a wide variety of sensors with unipolar, bipolar, differential, and single-ended outputs can be connected. In addition, drift, offset, linearity, signal-to-noise ratio, and common-mode rejection requirements can be met. In turn, it’s possible to realize a high-precision DAQ system for applications with extremely high-accuracy requirements.
Thomas Brand is a Field Applications Engineer at Analog Devices Inc.