1. Introduction
Since the beginning of the 21st century, there has been a growing demand for proximity sensors in various fields. Due to the various demands, proximity sensors have been developing in the direction of diversification. For different applications, proximity sensors should be selected to meet the specific requirements [
1,
2]. The proximity sensor is a kind of position sensor which can operate moving objects without contact. When the object is near the sensing surface of the sensor, it does not need mechanical contact or any pressure to make the sensor act. There are many types of proximity sensors, including those that use quartz crystal oscillators, and they are very accurate [
3,
4]. Because inductive proximity sensors have the advantages of a simple structure and strong anti-interference ability, the application of inductive proximity sensors is becoming increasingly extensive. With the development of aerospace and industry, high requirements are placed on the long-distance measurement of inductive proximity sensors [
5].
In order to increase the measuring distance, the mechanical principle of a metal slope is used to realize the measurement of larger distances. The distance measuring ranges are small and strongly dependent on the sensor size. In addition, the sensors are not by themselves able to measure translatory movements of the target.To overcome this obstacle, users usually complete the system with a metal incline, which is rigidly coupled with the object. The large displacement is converted mechanically into a small change of the distance between the incline and the active face of the inductive proximity sensor. This method enables significant increases in the sensing range. The function of the incline is performed by a cone located on the device shaft. However, the advantages of inductive proximity sensors, such as linearity, accuracy and repeatability, need to find compromises with the slope conversion rate (usually 5–10). It can be seen that the method of using a mechanical slope to increase the detection distance is limited. In [
6], the displacement of a target metal object is detected by means of a permanent magnet (auxiliary target), which increases the detection distance. However, the measurement target must be magnetic, which limits the types of objects that can be measured. Therefore, in [
7], a detection method is studied which can not only eliminate the limitation of the mechanical conversion device and/or auxiliary magnet target but can also allow the inductive proximity sensor to achieve high performance. This method is realized according to the function relationship between the detection distance and transducer impedance (Z). However, it is difficult to describe the dependence of the transducer impedance on the transducer’s electrical and mechanical properties. Therefore, the relationship between the transducer impedance (Z), detection distance and operating frequency can only be reflected logically.
For long-distance measurement, a bridge differential inductance detection circuit is used in the measurement circuit of sensors in this paper [
8]. The detection circuit detects the inductance variation of the coil in the induction surface to realize the measurement of the approach distance of the target metal objects. This method can eliminate most external interference signals and, as far as possible, ensure the effectiveness of the necessary information we need to enhance the anti-interference ability [
9,
10]. In order to achieve consistent product performance, in this paper, we propose an automatic zero calibration method for the foundation of an integral–proportional-integral (I-PI) controller on the basis of a bridge differential inductance detection circuit. Because the reference inductor in this paper needs to be adjustable, this method solves the problem of the difficulty of making the reference inductor [
11].
At present, most inductive sensors adopt a measurement method based on an AC damping factor. This approach consists of an inductance coil and an external capacitance to form an oscillation detection circuit. The smaller the detection distance, the greater the damping coefficient and the smaller the amplitude of oscillation, and even the vibration is stopped.Therefore, the proximity of metal targets can be measured according to the oscillation amplitude. However, the RLC which is a series parallel circuit of resistance, inductance and capacitance of the transmission line will affect the frequency and amplitude of oscillation, which will lead to a decrease in the measurement accuracy. The method in this paper realizes the detection of all metal targets at the same distance, which has no attenuation to the measured material detection distance and does not need to be recalibrated for different materials. This method has high precision and stable performance. The inductance increment detection circuit can cancel the self-inductance of the detection coil by setting the bias inductance and then converting the equivalent inductance increment related to the proximity into an electric quantity. We can thus obtain the distance information of the corresponding metal target and meet the requirements of measurement accuracy. The capacitance range of traditional sensors such as capacitive sensors is very small, thus placing high requirements on the accuracy of capacitance detection. Especially in the process of sensor development, high-precision capacitance detection equipment is often required to test and calibrate the sensor. However, there is a lack of a special instrument for the real-time detection of micro capacitance in China and abroad. The common practice is to design and make a special capacitance detection circuit for the sensor, which undoubtedly increases the difficulty and workload of sensor design. In this paper, the I-PI automatic zero adjustment method is adopted, which is easy to use and reduces the error.
Based on the analysis of the bridge differential inductance detection circuit, a measurement method with an attenuation coefficient of 1 based on the field-programmable gate array (FPGA) is proposed, which can detect multiple metal objects at the same sensing distance.This is the attenuation coefficient of 1. When the common inductive proximity sensor detects metal objects of different materials, the detection distance will change. This is due to the different attenuation coefficients between different metals. When detecting objects made of different types of metals, different types of sensors must be used, which greatly increases the workload of staff and reduces the consistent reliability of products. A sensor with an attenuation coefficient of 1 solves these problems and unifies the switching distance of all metal types; that is, the switching distance is the same for all metal materials. This makes it unnecessary to replace new sensors when detecting the proximity of different metal objects, thus reducing the budget and costs. This paper first introduces the operating principle of inductive proximity sensors for long-distance measurement. On this basis, a small signal model from control voltage to bias inductance is established. An automatic zero calibration method based on an integral–proportional-integral (I-PI) controller is proposed, and the attenuation coefficient of 1 is achieved by searching in the table.
3. Principle of the Proposed Adaptive Zero Adjustment Technique
In the process of practical application, before the target metal enters the sensing range of the inductive proximity sensor, it is essential to adjust the
to make Lr equal Ls. From Equation (
3), we can adjust the
applied to the JFET gate source to make the reference inductance
equal to the measured inductance
. Then, the input
is zero and the output
of the instrument amplifier is also zero, which is called zero adjustment. After zeroing, the
must be kept unchanged, and the inductance value of the reference inductor remains unchanged. When the metal object enters the sensing range of the sensor, the inductance of the coil will vary and the increment
will be generated.
and the output
of the instrument amplifier will no longer be 0. The
will be sent to the FPGA of the later stage for processing so that the approach distance of the target can be obtained according to the variation of inductance. The distance measurement of the target metal object is thus realized.
If the sensor has not been zeroed—that is, the reference inductance
in the electronic measurement circuit is not equal to the value of
of the detected coil—when the target metal is far away from the sensing range of the sensor,
will not be equal to zero and the amplified signal
of the instrument amplifier will certainly not be equal to zero. This can lead to inaccurate distance measurement. In order to realize the automatic zero adjustment of the sensor, it is known from control theory that closed-loop negative feedback control must be implemented for the control quantity [
21,
22]. The control quantity here is the signal
amplified by the instrument amplifier. The error signal is compared with the given zero. After the error signal is adjusted by the controller, the control signal
of the JFET gate source is generated and
is obtained to realize the reference inductance.
As long as the negative feedback is designed and the system is stable, the reference inductance is equal to the of the detection coil. The input end of the controller is zero and the output is also zero. This realizes the automatic zero adjustment of the inductive proximity sensor. This kind of automatic zero adjustment technology is realized by the adaptive adjustment of the reference inductance. The adaptive adjustment of the reference inductance is helpful for the accurate detection of the inductance increment, remote measurement and the accurate measurement of the proximity of the inductive proximity sensor.
3.1. Derivation of Small Signal Model
The electronic measuring circuit is composed of a differential detection circuit and instrument amplifier INA. According to
Figure 2, we can set
, where
is the equivalent oscillation excitation voltage source and is connected to the operational amplifier through
. According to the characteristics of the operational amplifier, the following formula is obtained:
According to the characteristics of the instrument amplifier and the balance characteristics of the bridge, the following equations can be listed:
Therefore, we can obtain the expression of the current
flowing through
,
and
:
Then, we can obtain the expression:
By substituting Equations (5) and (6) into Equation (
7), the solution is obtained as follows:
If the amplification factor of the instrument amplifier is A, we can list the equation as Equation (
9):
Obviously, when , , , the output of the circuit is equal to zero. When the target metal is outside the measuring range of the inductive proximity sensor, we first adjust the reference inductance to make equal to zero. When the target metal objects enter the sensing range of the inductive proximity sensor, the distance measurement is carried out.
The work of the inductive proximity sensor is nonlinear. When the entire system is working,
to
is not linear, which is not conducive to the analysis and design. Imperfections are unavoidable in the production processes of real devices. Despite this, and despite the fact that real devices usually operate in regimes that are far from ideal, the sensors still work. This is related to the fact that imperfections give rise to hidden dynamics [
23]. A small signal model can reflect the dynamic performance. In this paper, the small signal model is established by the perturbation method [
24].
According to
Figure 2, the equations can be listed as follows:
where
is the voltage at steady state. By further simplification, the following results can be obtained:
Taking Laplace transformation on both sides of the equation, we obtain the following results:
Then, we obtain the following Equation (
15):
3.2. Zero Adjustment Based on I Regulator
In this paper, we choose
The JFET model is J2N4091, where
. The integral controller block diagram is shown in
Figure 5 and the adaptive zero adjustment function is realized by the closed-loop control method.
We can obtain the open-loop transfer function of the system according to Equation (
15) without considering the integral regulator. This is shown in Equation (
16):
The logarithmic frequency characteristic curve is shown in
Figure 6. It can be seen that the cut-off frequency is 3.589 khz and the phase angle is
(below
) without the regulator; thus, the system is unstable and we need to perform correction.
The integral controller is adopted and its transfer function is shown in Equation (
17):
The open-loop transfer function with an integral controller is shown in Equation (
18):
The correction is carried out according to the target with a phase angle margin of
at 100 kHz, which is calculated by the following formula:
The solution is as shown below:
By substituting Equation (
20) into Equation (
18), the corrected open-loop transfer function can be obtained as shown in Equation (
21):
The logarithmic frequency characteristic curve is shown in
Figure 7. It shows that the expected correction effect is achieved.
It can be seen from the transfer function that the system is always stable after the I-regulator is added. From Equation (
15), it can be seen that the transfer function of the system is equivalent to the first-order inertial link after adding the integral regulator. The output of the inertial link does not change in proportion to the input at the beginning. Until the end of the transition process, the output can remain proportional to the input. In order to achieve zero deviation, the system should be corrected to a type I system.Therefore, it is necessary to cascade a PI controller; the I-PI controller is used in this paper.
3.3. Zero Adjustment Based on I-PI Regulator
The control block diagram of the integral–proportional-integral controller is shown in
Figure 8.
The transfer function of the I-PI controller is shown in Equation (
22):
Then, calculation is performed according to the following conditions:
The solution is shown below:
According to the corresponding relationship between the PI parameters, resistance and capacitance, the calculation of the I-integrator is as follows:
The logarithmic frequency characteristic curve is shown in
Figure 9. The schematic diagram of the integral–proportional-integral regulating circuit is shown in
Figure 10. The simulation waveforms are shown in
Figure 11,
Figure 12 and
Figure 13. Compared with the steady-state error of an integral regulator, the introduction of an I-PI controller can reduce the error.
The simulation waveform of
Figure 11 is the positive input voltage waveform of the amplifier, and
Figure 12 is the negative input voltage waveform of the amplifier.
Figure 13 is the output voltage waveform. Simulations show that the input and output waveforms coincide. The upper and lower bridge arms reach balance, and the circuit realizes the function of zero adjustment. The circuit is stable, which verifies that the control strategy is feasible and realizes the automatic zero adjustment of the measurement circuit in this paper.
4. An Implementation Method for an Inductive Proximity Sensor with an Attenuation Coefficient of 1
It can be seen from
Figure 1 that when the target metal is close to the sensing surface of the detection coil, the inductance of the detection coil changes, resulting in an inductance increment
.
changes with the change of the approaching distance. The variable inductance increment signal is sent to the ADC converter in the FPGA circuit through the amplified electrical signal
of the conditioning circuit. In the FPGA, the algorithm program mixer is used to form the proximity distance table. The proximity distance table is a table drawn according to
. This proximity distance table is made according to the corresponding relationship between
and
. The proximity distance
can be obtained by querying the table. For different target metal objects, different proximity tables are stored in the FPGA, and the distance detection of different metal targets is realized at the same sensing distance.
The corresponding relationship between
and the proximity
is stored in the FPGA in the form of a table, and the proximity distance is obtained by looking up the table. For different metal objects, the corresponding proximity distance table can be corrected to realize the measurement of different metal objects at the same sensing distance; that is, an inductive proximity sensor with an attenuation coefficient of 1 is realized. The FPGA circuit in
Figure 1 mainly realizes the analog-to-digital conversion of
. The proximity distance table is obtained by the mixer program. The proximity distance
is queried according to
, and then the distance signal or switch signal is output through the low-pass filter. The application of the FPGA reduces the use of some hardware circuits and helps to achieve high reliability from the inductive proximity sensor.
In order to realize an inductive proximity sensor with an attenuation coefficient of 1, it is necessary to find different tables for different metals through the FPGA in order to realize the compensation output of different metals. The overall flow chart of the FPGA look-up table method is shown in
Figure 14 below.
As can be seen from
Figure 14, the signal
amplified by the instrument is converted to an analog-to-digital signal through the ADC chip. The digital signal is sent into the FPGA, and the digital signal is used as the address of the look-up table of the FPGA to be searched through. Different target metal objects switch different look-up tables for their signal output in order to realize the output compensation of metals with different attenuation coefficients and the on–off switching of the inductive proximity sensor for different metals at the same detection distance.
5. Experimental Results
The experimental results verify the efficacy of the automatic zero adjustment technology and the inductive proximity sensor with an attenuation coefficient of 1. The FPGA chip is xc7a35tfgg484-2 from Xilinx. The ADC chip AD9280 and DAC chip AD9708 were created by the Analog Device Company. The instrument amplifier was INA101HP from Texas Instruments. The waveform of the voltage source is shown in
Figure 15. The output voltage
waveform after automatic zero adjustment by the I-PI controller is shown in
Figure 16. The picture of the prototype is shown in
Figure 17. We can see that the experimental waveforms coincide with the simulation waveforms, and the circuit achieves the function of zero adjustment.
In this paper, three kinds of metal targets are tested. Firstly, the copper is tested. The look-up table of the FPGA is switched to the copper look-up table, and the target metal that is outside the detection range is close to the sensor. At this time, the voltage of the multimeter is 0. When the copper reaches the detection distance, the multimeter changes to the supply voltage of the sensor. The distance at this time, which is the reclosing distance, is then recorded. Then, the copper is slowly moved away from the inductive proximity sensor. When the multimeter jumps to zero again, the distance at this time, which is the distance when it is disconnected, is then recorded. The above process is repeated, and three sets of data are recorded. The measurement processes for A3 steel and magnesium aluminum alloy are similar. It is only necessary to switch the FPGA look-up table to continue the experiment.
Table 1 shows the test data; the unit of distance is mm. It can be seen from
Table 1 that an inductive proximity sensor with an attenuation coefficient of 1 is realized; that is, different metal objects can be measured at the same distance, and we can obtain the proximity distance of the object from
Table 1. The automatic zero calibration technology of the I-PI controller designed in this paper is helpful for high-precision measurement.