Sensor excitation and measurement solution composed of AD5933/AD5934

Many of today’s industrial and instrumentation applications involve sensor measurements. The function of a sensor is to monitor changes in a system and then feed this data back to a host controller. Sensors used for simple voltage or current measurements may be resistive. However, some sensor systems may be inductive or capacitive, meaning that the impedance changes are nonlinear over the sensor’s frequency range.

Typical examples of this type of complex impedance sensor are proximity sensors, which are used to detect the relative distance of a moving object; in addition, capacitive sensors or inductive sensors are used in medical equipment to measure blood flow or analyze blood pressure or blood quality.

In order to make measurements with these "complex impedance sensors", an alternating current (AC) excitation frequency source must be provided to sweep across the frequency range of the sensor. This article attempts to show how such frequency sweeps up to 10 MHz can be easily achieved using a single-chip digital waveform generator. A single-chip sensor solution with integrated excitation, response, and digital signal processor (DSP) functionality is also presented for applications requiring excitation frequencies up to nearly 50 kHz.

Sensor: Working Principle
Figure 1 shows a complex impedance sensor model consisting of inductive and capacitive reactance characteristics.

The excitation frequency signal passing through the sensor will show a change in amplitude, frequency or phase according to the instantaneous value of the sensor's L or C. For example, an ultrasonic flow meter will show a phase shift, while a proximity sensor will cause an amplitude change.

Figure 1. Sensor model with complex impedance characteristics

The most common method of tracking this changing impedance is to monitor the resonant frequency of the circuit. The resonant frequency is the frequency at which the capacitance is equal to the inductance. This is also the frequency at which the maximum impedance value occurs on the frequency curve. For example, consider the proximity sensor shown in Figure 2. Under normal circumstances, such as static conditions, the sensor's L, R, and C all have a unique value, with a maximum impedance value at the resonant frequency Fo. When a moving object approaches the sensor, the sensor's L and C values ​​change and a new resonant frequency is created. By monitoring the change in resonant frequency (and thus the change in impedance), it is possible to infer how far the moving object has moved relative to the sensor.

Figure 2. The resonant frequency of a proximity sensor changes with distance traveled.

Calculating the resonant frequency
Calculating the resonant frequency of a circuit requires measuring the relationship between frequency and impedance (as shown in Figure 2), and specifically requires a waveform generator that has the ability to sweep over a range of frequencies. A simple, low-cost approach to this is to use the AD5930 waveform generator. The AD5930 has the ability to provide a linear sweep over a set of preset frequency ranges. Once the conditions are set, no further control is required, except for a trigger to initiate the frequency sweep.

AD5930 has many advantages: The output frequency resolution is 28 bits, so users can control the output frequency with less than 0.1 Hz. Its output frequency range is 0 to 10 MHz, which provides great flexibility in selecting sensors. For example, some sensors have a narrow frequency range, but require high resolution within this frequency range. Other sensors may require a wide frequency modulation range, but require lower resolution.

Using this method it is easy to calculate the resonant frequency of the sensor.

System Block Diagram
A typical block diagram of such a system is shown in Figure 3. The AD5930 digital waveform generator is set up by the BF-535 DSP processor. The sine wave output voltage waveform generated from the AD5930 needs to be low-pass filtered and amplified to eliminate feedthrough caused by the master clock (MCLK), image frequencies, and high-frequency noise. The filtered signal can be used as the excitation frequency source for the sensor. Depending on the impedance response of the sensor, the signal may need to be amplified to bring it into the dynamic range of the analog-to-digital converter (ADC). The output of the sensor and the excitation frequency source are input to the AD7266, a 12-bit, 2 MSPS simultaneous sampling dual ADC. The data output from the ADC is saved in memory for further analysis to calculate the phase and amplitude offset of the sensor.

Figure 3. System block diagram

Complete Integrated Sensor Solution The
discrete solution described above is a commonly used sensor impedance measurement solution. This solution may require many discrete components, making it a high-cost sensor analysis solution. These individual components also add their own error sources. The active components in the design also add phase errors, which also need to be corrected. In addition, the DSP is required to handle some complex mathematical calculations, which may require external memory to store the raw ADC data, further increasing the cost.

The solution to the above low-frequency sensor analysis problem is the AD5933/4 device, which integrates the above major processing blocks into a single chip. The core of the chip includes three main units: a direct digital frequency synthesizer (DDS) waveform generator for providing frequency sweeps; a 12-bit, 1 MSPS ADC for measuring the sensor's response; and finally a DSP engine capable of performing a 1024-point discrete Fourier transform (DFT) operation on the ADC measurement data.

The DFT operation result provides a real part (R) and an imaginary part (I) data, which can be easily calculated from the impedance. The magnitude and phase of the impedance can be easily calculated using the following formula:

In order to determine the actual real impedance value Z(ω), a frequency sweep is usually required. The impedance at each frequency point can be calculated, so that a frequency-amplitude curve can be obtained. This makes it easy to measure impedance in the range of 100 Ω to 20 MΩ. The system allows the user to set a 2 V peak-to-peak (PK-PK) sinusoidal signal as the excitation frequency source for the external load. The output range can also be set to 1V, 500 mV and 200 mV. The frequency resolution can reach 27 bits (< 0.1 Hz).

Implementing frequency scanning:
In order to implement frequency scanning, the user must first set the conditions required for frequency scanning: a starting frequency, frequency interval and number of sweep points are required. Then a start command is required to start the scan. At each sweep frequency point, the ADC first completes 1024 samples, and then performs DFT calculations to provide the real and imaginary data of the waveform. This real and imaginary data is provided to the user in the form of two 16-bit words through the I2C interface. The advantage of the on-chip DSP processing unit is that the user does not have to perform complex mathematical calculations or store the ADC raw data, but only needs to provide two 16-bit data. Therefore, it also allows the selection of a cheaper DSP solution because the requirements for the final processing power are greatly reduced.