www.as-se.org/ccse Communications in Control Science and Engineering (CCSE) Volume 4, 2016 Realization of Absolute Capacitive Rotary Encoder System Based on Capacitive Gate Technology Lu Zhang 1, Dezhi Zheng* 1, Shaobo Zhang 1 1.School of Instrumentation Science and Opto-electronics Engineering, Beihang University, Beijing, 100191, China 1 zhanglu0801@buaa.edu.cn; *2 zhengdezhi@buaa.edu.cn; 3 zhangshaobo@buaa.edu.cn Abstract Based on the structure foundation of capacitive gate technology and incremental encoder, this paper proposes an absolute capacitive rotary encoder system for measuring angular displacement and angular velocity. According to the vernier effect, the encoder is composed of a pair of incremental encoders with the electrical angle difference of a cycle. The voltage signal is conducted with digital demodulation through orthogonal demodulation technology, so that the demodulated digital signal is iterated through arc tangent digital resolving method in field programmable gate array (hereinafter referred to as FPGA) to obtain the measurement angle through accumulation. The resolving result is transmitted to the host computer through the Universal Serial Bus (hereinafter referred to as USB), and processed and displayed on the established LabVIEW test platform. The absolute capacitivie encoder has strong common mode interference resistance ability, which can avoid the null point cumulative error. Moreover, it does not need to seek for the reference null before power on. Therefore, it has an extensive prospect in industrial applications. Keywords Capacitive Encoder; Angular Position Sensor; Absolute Encoder Introduction Rotary encoders, as indispensable components in automatic control technology, are widely utilized in detecting the angular displacement and angular velocity. Moreover, it can indirectly detect the linear displacement and linear velocity. In practical applications, the rotary encoder can be utilized as not only the sensitivity detection component to constitute the automatic detection system, but also the detection feedback component to form the automatic regulation and control system of closed loop or semi-closed loop. The rotary encoder can be divided into incremental encoder and absolute encoder according to the working principle. The former converts the displacement to the periodic electrical signal, and only provides the relative change information of position. It is characterized by simple system architecture and low cost. The latter directly outputs the current location. Each position corresponds to a unique output code in a circle. It does not need to memorize or set the reference point. Moreover, it has strong anti-interference characteristics and data reliability. According to the detection principle, the rotary encoder can be divided into photoelectric encoder, electromagnetic encoder and capacitive encoder. At present, rotary encoder market is almost monopolized by photoelectric and electromagnetic encoders [1,2]. Compared with the photoelectric and electromagnetic encoders, capacitive encoder has obvious advantages [3-6]. Photoelectric encoder is sensitive to the vibration, and the electromagnetic encoder is bulky and easily affected by the electromagnetic interference. In contrast, capacitive rotary encoder can not only adapt to vibration, high temperature, dust, high humidity and other harsh environments, but also has the advantages of high accuracy, high resolution, high reliability, and good dynamic characteristics. Moreover, it is characterized by simple structure, low power consumption, and easy realization of the miniaturization. At present, there are many experimental studies on the mechanical structure, geometric size and anti interference of capacitive encoders, while few research papers have been published. There are few capacitive rotary encoders in the markets even though in the products lists of well-known encoder manufacturers [7-11]. Originally, Professor Georg Brasseur proposed the principle of proportional measurement of capacitive rotary encoders in early stage. Then Yishay Netzer classified and analyzed all kinds of capacitive rotary encoders. Based on the capacitance plate with sinusoidal envelope curve as boundary proposed by Stephen M. Fortescue, he further improved the performance of the capacitive encoder. Netzer has produced capacitive encoder products with a variety of sizes and models, and the 30
Communications in Control Science and Engineering (CCSE) Volume 4, 2016 www.as-se.org/ccse highest precision is up to 90. The capacitive rotary encoder has important research value and significance. Principle and System Design The absolute capacitive rotary encoder is mainly composed of four parts, namely the capacitive sensing element, a signal processor, a unit of signal demodulation, and USB communication module. The capacitive sensing element is utilized to convert the angle signal caused by the change of the mechanical displacement into the change of the capacitance value. The signal processor converts the change of the capacitance value of the sensing element into an electric signal that has been amplified and filtered. The signal demodulation unit is utilized to obtain electrical signal and get it processed, which will be output in an appropriate form. USB communication module is utilized for data transmission with the host computer. A. Basic Sensing Element FIG. 1 STRUCTURE OF SENSIGN ELEMENT The capacitance sensing element consists of a pair of incremental coded plates, each of which is composed of a rotor and a stator, as shown in FIG. 1. The two plates coated with copper are made of printed circuit boards. The surface of the rotor is paved with a petal shaped reflector with a boundary of a sine curve. Fixed on the rotating shaft, the rotor can rotate with the measured object. The stator is fixedly connected with the encoder shell, which is composed of a surface emitting electrode and a pair of receiving electrodes. The emitter is a grid electrode which is uniformly distributed, and the receiving electrode is the two annular electrodes distributed on the inner and outer sides of grid electrodes. In order to shield the external electromagnetic interference, weaken the influence of the edge effect and parasitic capacitance, the inner and outer edges of rotor and stator are provided with protective ring. Signal excitation and demodulation system are distributed on the other side of the stator. B. Circuit Model of the Sensor Taking one of the combined plates as an example, its working principle is shown in FIG. 2. FIG. 2 CIRCUIT MODEL The emitter electrode of the stator is composed of four groups of electrically interconnected and alternately arranged excitation electrode sets X, Y, Z, W. Each excitation electrode set includes N electrodes, which are uniformly arranged on a circle in the cycle of 2π/N (N is a positive integer). Therefore, the distance between two adjacent electrodes is 2π/4N, and the angle of each electrode is λ, λ < 2π/4N. The receiving electrode is two annular electrodes which are respectively located at the inner side and the outer side of the emitter electrode, namely, the inner ring electrode A and the outer ring electrode B. In order to reduce the capacitance boundary effect between the emitter and the two sides of the receiving electrode, a grounding shielding electric power is necessary at the junction of the two electrodes. The reflector C and D are separated by a sinusoidal curve shaped insulation with the period under polar 31
www.as-se.org/ccse Communications in Control Science and Engineering (CCSE) Volume 4, 2016 coordinate system of 2π/N. The gap of the sine curve is mapped to the emitter electrode on the stator. The base of the reflector is mapped to the receiving electrode A and B on the stator. The sinusoidal signal with the phase correlation of 90 degree is applied to the excitation electrode groups X, Y, Z and W. When the rotor is rotated, the coupling area of the emitter electrode group and the reflectors C and D is changed in a sinusoidal pattern, which modulates the electric field generated by the emitter. The reflectors C and D receive the modulated charge. The coupling area between the reflector and the rotor is not changed. The charge on the reflector is coupled to the receiving electrode, and the receiving electrodes A and B produce the same modulation signal. FIG. 3 SCHEMATIC OF ENCODER The coupling capacitances of the excitation electrode set X, Y, Z and W of the emitter electrode and the inner loop reflector electrode C are assumed to be C XC, C YC, C ZC and C WC. The coupling capacitance with the outer ring reflector D is C XD, C YD, C ZD and C WD. The coupling capacity of the inner loop reflector A and the inner loop receiving electrode C is C AC. The coupling capacitance of the outer loop reflector B and the outer loop receiving electrode D is C BD. The equivalent circuit of the capacitor encoder is shown in FIG. 3. The capacitor between the emitting and the reflecting electrode is connected in series, and then connected with the capacitor in series between the reflecting and receiving electrodes A differential charge amplifier is utilized to carry out the differential detection of the receiving electrodes A and B. The charge is converted into a voltage signal V out to obtain the output signal of the rotary encoder. C. Absolute Encoder Absolute encoder outputs a unique angle in a circle of the rotor. Incremental rotary encoder outputs of N cycles of electric angle signal in the rotation of the rotor. The output range of the electrical angle of each cycle is 0~360. If the angular displacement of the same rotor is measured through two incremental rotary encoders, and the output electric angles has a difference of one cycle, namely, one outputs the electrical angle of N cycles, and the other outputs the electrical angle of N-1 cycles. The absolute angular displacement of the rotor can be obtained by the difference between the two different cycles, as shown in Figure 4. Therefore the absolute rotary encoder can be composed of two incremental rotary encoders. A 16-cycle incremental capacitance rotary encoder and a 15-cycle incremental capacitance rotary encoder share a rotating plate. The two sides of the rotor are two reflection electrodes of the incremental rotary encoder. The rotor is a multilayer printed circuit plate. The grounding shield layer is paved in the middle of the plate to prevent mutual interference between the two incremental capacitance rotary encoders. The principle of obtaining the absolute angular displacement through the difference can not only suppress the common mode interference in the demodulation process and improve the calculation precision, but also offset the phase delay in the calculation process and enhance the dynamic response of the rotary encoder. 32
Communications in Control Science and Engineering (CCSE) Volume 4, 2016 www.as-se.org/ccse FIG. 4 SIGNAL PROCESSING ARCHITECTURE D. Signal Processing FIG. 4 shows the digital demodulation system of absolute capacitive encoder. The relative angular displacement is obtained by synchronous demodulation of two incremental capacitive rotary encoders. The absolute angular displacement is obtained by subtracting. The demodulation process of incremental rotary encoder, especially quadrature demodulation and coherent demodulation, involves the generation of excitation signal, the design of band-pass filter and low-pass filter, multiplication operations and the arc tangent operations. Therefore, the whole demodulation system is suitable to be realized in digital circuit. Two incremental rotary encoders shall be demodulated synchronously. The timing sequence of the demodulation process has strict requirements. As a result, the absolute capacitive rotary encoder shall be conducted with digital demodulation through the demodulation control system with FPGA as the control core. Digital signal generator module in FPGA generates the encoder signal. According to the selected encoder working mode, four channels of excitation signals with the phase difference of 90 are generated. The ex-citation signals are applied to the emitter of the two incremental rotary encoders after DA conversion. The output voltage are demodulated in FPGA through AD collection. The two channels of modulating signals are conducted with parallel demodulation in FPGA. Two demodulation modules are exactly the same, so that the synchronous demodulation results can be obtained. The two channels of demodulation results obtained are transmitted to the host computer via USB. The calculation results are displayed in real time on the LabVIEW, and the calculation results are conducted with error analysis. Conclusions The paper proposed a novel capacitive rotary encoder. To validate the measured result of arc tangent calculation implemented in the FPGA processor, some experiments were carried out. FIG. 5 shows the result of the angular displacement when the rotor is stationary. From the waveform in the figure, the precision of the encoder is less than 0.008. Dynamic measurement was conducted when the rotor is rotating at the speed of 500 rpm, which are shown in FIG. 6. The mechanical angle of each incremental encoder is accumulated in the form of electric cycle, and the two mechanical angles can be used to realize the absolute coding, which are presented in FIG. 7 and FIG. 8. 10.428 Angle (degree) 10.426 10.424 10.422 0 10 20 30 40 50 Time (ms) (a) OUTPUT ANGLE WHEN PERIOD IS 16 33
www.as-se.org/ccse Communications in Control Science and Engineering (CCSE) Volume 4, 2016 10.252 Angle (degree) 10.25 10.248 10.246 10.244 0 10 20 30 40 50 Time (ms) (b) OUTPUT ANGLE WHEN PERIOD IS 15 FIG. 5 STATIONARY MEASUREMENT Electronic Angle (degree) 360 300 200 100 16 period 15 period 0 0 20 40 60 80 100 120 Time(ms) FIG. 6 DYNAMIC MEASUREMENT WHEN ROTOR AT THE SPEED OF 500RPM Accumulated Electronic Angle (degree) 5760 5000 4000 3000 2000 1000 16 period 15 period 0 0 20 40 60 80 100 120 Time(ms) FIG. 7 THE CONTRAST OF MECHANICAL ANGLE 360 Absolute Angle (degree) 300 200 100 0 0 20 40 60 80 100 120 Time(ms) FIG. 8 OUTPUT ABSOLUTE ANGLE To improve the performance of the encoder, further experiments such as stationary nonlinearity and dynamic property is under progress. More evaluation will focus in the analysis of the encoder performance. 34
Communications in Control Science and Engineering (CCSE) Volume 4, 2016 www.as-se.org/ccse ACKNOWLEDGMENT This paper is supported by the The National Key Technology R&D Program (2014BAF08B01) REFERENCES [1] G. Brasseur. A Robust Capacitive Angular Position Sensor[A]. IEEE Conference on Instrumentation and Measurement Technology[C]. Brussels: Belgium, 1996: 1081-1086. [2] Yishay Netzer, Capacitive Displacement Encoder, US Patent 6492911B1, 2002. [3] Zhang Yinfang. The Wworking Principle and Characteristics of Capacitive Grating Displacement Sensor [J]. Aviation Precision Manufacturing Technology, 2005, 41(4): 58-59. [4] Xu Kejun. Research and Application of Capacitive Grating Sensor [M]. Beijing: Tsinghua University Press, 1995: 6-10,64-87. [5] Zhang Rong,Xu Mingqian. Principle and Application of Capacitive Grating Rotary Encoder [J]. Sensor World, 2006 2: 21-24. [6] Kyung-pyo Kang. Method and Apparatus for Processing the Output Signal of An Encoder. US Patent 7307392B2, 2007. [7] Yingjie Lin. Capacitive Angular Position Sensor. US Patent 6774642B2, 2004. [8] James Edward Nelson. Capacitive Rotary Position Encoder. US Patent 5736865, 1998. [9] Bruce E.Rohr. Capacitance Sensing System Using Multiple Capacitances to Sense Rotary Motion. US Patent 4864295, 1989. [10] Claus-Peter Krumholz. Capacitive Rotary Transmitter for Controlling and Positioning Displaced Objects. US Patent 4851835, 1989. [11] Tibor Fabian, G. Brasseur. A Robust Capacitive Angular Speed Sensor[J]. IEEE Trans. Instrum. Meas, 1998: V47 280-284. 1 Lu Zhang received the B.Sc. degree in measurement & control technology and instrumentation and from North China Electric Power University, Beijing, China in 2013. He is currently pursuing the M.Eng. degree with the Department of Instrumentation Science and Opto- electronics Engineering, Beihang University, Beijing, China. His current research interests include signal processing an capacitive measurement. *2 Dezhi Zheng received the B.Sc. and Ph.D. degree in instrumentation science and technology from Beihang University, Beijing, China in 2000, 2006, respectively. He is currently an associated professor with the Department of Instrumentation Science and Opto-electronics Engineering, Beihang University. His current research interests include Coriolis mass flowmeter, pressure sensor and intelligent instrumentation. 3 Shaobo Zhang received the B.Sc. degree in instrumentation science and technology from Beihang University, Beijing, China in 2012. He is currently pursuing the Ph.D. degree with the Department of Instrumentation Science and Opto- electronics Engineering, Beihang University, Beijing, China. 35