High-Speed Tracking Control System for Broadcast-Use Optical Disk Drive

Transcription

1 Japanese Journal of Applied Physics Vol. 45, No. 2B, 26, pp #26 The Japan Society of Applied Physics High-Speed Tracking Control System for Broadcast-Use Optical Disk Drive Daiichi KOIDE, Haruki TOKUMARU, Kiyoshi OHISHI 1, Kazuyoshi KURAMOCHI 1 and Toshimasa MIYAZAKI 2 Science and Technical Research Laboratories, NHK (Japan Broadcasting Corporation), Kinuta, Setagaya-ku, Tokyo , Japan 1 Nagaoka University of Technology, Kamitomioka-cho, Nagaoka, Niigata , Japan 2 Nagaoka National College of Technology, 888 Nishikatakai, Nagaoka, Niigata , Japan (Received August 21, 25; accepted November 4, 25; published online February 24, 26) A high-speed tracking control system using feed-forward control with the zero phase error tracking method (ZPET-FF control) has been applied to high-data-transfer-rate and high-density optical disk drives for broadcast use. To realize high-datatransfer-rate optical disk drives, the optical head must be controlled for track-following with an accuracy of less than 9 nm during high-speed disk rotation. We compared the characteristics of the tracking accuracy of ZPET-FF control with those of conventional feedback control at 15 rpm and analyzed their frequency responses in a simulation. The results show that the ZPET-FF control is accurate and robust toward periodic disturbances at high speeds. Experimental results show that ZPET-FF control is more precise than feedback control, and the residual tracking error is suppressed to 8.2 nm at a linear velocity of 18.3 m/s in an optical disk drive having a blue-violet laser (wavelength: 45 nm), an object lens with a numerical aperture (NA) of.85, and media with a track pitch of.32 mm. [DOI: /JJAP ] KEYWORDS: optical disk drive, high-speed, track-following control, ZPET-FF control, zero phase error tracking method, residual tracking error, HDTV recording, broadcast use, Blu-ray disc, NA of Introduction A high-data-transfer-rate optical disk is required to record and play back the signals of high-definition television (HDTV) for professional use. To achieve this, we are now developing an optical disk system with a high data transfer rate of more than 2 Mbps and sufficient capacity to store HDTV programs. 1) Figure 1 outlines the target performance for HDTV recording on an optical disk for broadcast use, which is compared with the specifications of current digital HDTV recording equipment such as the Blu-ray disc recorder using a blue-violet laser with an objective lens of numerical aperture ðnaþ ¼:85, and high-density media with a track pitch of.32 mm. HDCAM is the VTR format of professional HDTV camera recorders for field recording, which have a total data transfer rate of 185 Mbps. 2) HD-D5 is the format of the professional HDTV VTR for studio recording and archiving, which has a total data transfer rate of 31 Mbps. 3) Moreover, high-data-transfer-rate playback is required for high-speed copying, high-speed transfer, and special types of playback such as nonlinear editing in prompt news gathering and program production. To record and play back a signal at rates of 2 Mbps or higher, track-following control must be performed precisely on a high-density optical disk rotating at high speed. In general, tracking error is increasing when the disk rotates at a high speed. The error contains the periodic disturbances caused mainly by disk eccentricity. However, conventional servo methods cannot realize precise control; therefore, new servo technology is required in order to suppress the servo error extensively. Recently, various servo technologies for precise control of optical disk drives have been reported, 4 6) but they have never significantly reduced servo error at high speed. We have proposed a high-speed tracking control system that combines a feedback (FB) control system with a feedforward (FF) control system. 7 1) The FF control system is based on the zero phase error tracking (ZPET) method 11) (referred to as the ZPET-FF method). This system has been shown to reduce residual tracking error far better than Data transfer rate (bps) G 16G 1G 3M 2M 1M Super Hi-vision HDTV for studio recorder Ultra high-definition TV scanning 4 lines (Program production, archiving) HDTV for camera-recorder (News gathering) for consumer (e.g. Blu-ray disc, HD-DVD) Fig. 1. Target performance for professional HDTV recording on optical disk for broadcast use. conventional FB control or repetitive control, 12) and has demonstrated the capability to reduce residual tracking error on an optical disk rotated at high speed with an optical system with a NA of.65. 7,8) In this study, to achieve high-speed recording on highdensity media, we applied the ZPET-FF method to an optical system having a blue-violet laser (wavelength: 45 nm), an object lens of NA ¼ :85, and media of Blu-ray disc format (track pitch:.32 mm). First, we designed a precise trackfollowing controller using ZPET-FF control in the system and showed the results of precise control at a high speed of 15 rpm in a simulation. Next, the experiment showed the ZPET-FF control method successfully reduced tracking error in comparison with conventional FB control during highspeed rotation. 2. Conventional Tracking Servo System: Feedback Control System The configuration of the conventional feedback control system generally used in tracking control for phase-change optical disk drives is illustrated in Fig. 2. In this figure, x d ðtþ represents the target track position on the optical disk, xðtþ represents the position of the optical beam spot, and eðtþ represents the tracking error signal. The tracking actuator PðsÞ moves so that the light beam spot xðtþ approaches the target track position on the disk x d ðtþ. Here, feedback control High-speed transfer, copy

2 Jpn. J. Appl. Phys., Vol. 45, No. 2B (26) x d (t) - e(t) Sampler Fig. 2. Feedback (FB) Controller C 1 (z -1 ) Zero-order holder Feedback control system. Tracking Actuator P(s) is accomplished by a feedback controller C 1 ðz 1 Þ, which is designed so that the control system has a large error margin and a wide servo range. The tracking error signal eðtþ is used as a parameter indicating high-accuracy tracking performance. The sampler and zero-order holder are, respectively, A/D converter and D/A converter models. Here, the feedback controller C 1 ðz 1 Þ is a generally used phase compensator or a proportional, integral and derivative (PID) controller. The feedback controller is tuned to have a larger DC gain and a wider servo bandwidth at the zero cross point, with a gain margin and a phase margin in the openloop characteristics of the system. 3. High-Speed Tracking Servo System: ZPET-FF Control Figure 3 shows the configuration of the high-speed tracking servo system we have proposed. FB control is performed by the tracking actuator PðsÞ based on an FB controller with the transfer function C 1 ðz 1 Þ. In addition, the tracking error (TE) is inserted into the FB system after being operated by a ZPET-FF control system consisting of a precompensator, memory z n, a signal estimator (a block that has the transfer function Cðz 1 ÞPðz 1 Þ=f1 þ Cðz 1 ÞPðz 1 ÞgÞ, and low-pass filter (LPF) Fðz 1 Þ. The pre-compensator has the transfer function G f ðz 1 Þ employing the ZPET method. 7) G f ðz 1 Þ is equal to the inverse transfer function of the closed-loop system of the FB control system, which is considered the effect of unstable zeros. 11) When the transfer function of the closed-loop system of the FB control system is G closed ðz 1 Þ, G closed ðz 1 Þ may be defined by G closed ðz 1 Þ¼ C 1ðz 1 ÞPðz 1 Þ 1 þ C 1 ðz 1 ÞPðz 1 Þ ¼ z d B c ðz 1 Þ : ð1þ A c ðz 1 Þ B c ðz 1 Þ¼b c þ b c1 z 1 þþb cm z m ; b c 6¼ ð2þ A c ðz 1 Þ¼1þa c1 z 1 þþa cn z n ; ð3þ x(t) where z is a complex number. B c ðz 1 Þ and A c ðz 1 Þ are polynomials of z 1. The transfer function of the precompensator of the FF control system, G f ðz 1 Þ, is defined by G f ðz 1 1 Þ¼ G closed ðz 1 Þ ¼ zd A c ðz 1 Þ B c ðz 1 Þ ; ð4þ where the term z 1 is equivalent to a one-step lead, and z d is a d-step lead. Here, considering of the effect of unstable zeros, FF control output e ff ðkþ is introduced to eq. (5) using the estimated error input signal ^eðkþ. e ff ðkþ ¼G f ðz 1 Þ^eðkÞ ¼ A cðz 1 Þ B c ðz 1 Þ ^eðk þ dþ: ð5þ Here, the pre-compensator requires a future input signal, which is two samples ahead (i.e., d ¼ 2). 7,13) To acquire the future value, the tracking error signal at a time before the previous disk rotation period is regarded as the future value, because the eccentric component caused by removable optical disk media typically appears in the tracking error signal periodically as the disk rotates. Therefore, the tracking error signal before the previous rotation period and two samples ahead can be applied to the future value using the memory z n. Here, n is the number of memory size fittings of a disk rotation period and two samples ahead in eq. (6), where N denotes the number of disk revolutions per minute and T denotes the sampling time in seconds. n ¼ 6 2: N T ð6þ The signal estimator produces a stable estimated tracking error signal ^eðkþ for calculating G f ðz 1 Þ. For the control system shown in Fig. 3, the following equations hold. e ¼ x d x x ¼ðe þ e ff ÞC 1 ðz 1 ÞPðz 1 Þ From those equations, eq. (9) is introduced. x d 1 þ C 1 ðz 1 ÞPðz 1 Þ ¼ e þ C 1 ðz 1 ÞPðz 1 Þ e 1 þ C 1 ðz 1 ÞPðz 1 ff ¼ ^eðkþ Þ ð9þ Here, the left-hand side is the error of the feedback control system itself. When the input signal of the precompensator of the FF control system is equal to ^eðkþ, the ð7þ ð8þ ê(k) C 1 (z -1 ) P(z -1 ) 1C 1 (z -1 ) P(z -1 ) ZPET- FF Controller Memory LPF Z -n F(z -1 ) ê(k2) Signal estimator G f (z -1 ) e ff (k) Precompensator x d (t) e(t) Sampler Feedback (FB) Controller C 1 (z -1 ) Zero-order holder Tracking Actuator P(s) x(t) Fig. 3. High-speed tracking control system (ZPET-FF control system). 1188

3 Jpn. J. Appl. Phys., Vol. 45, No. 2B (26) Gain (db) 4-6 Phase Gain -18 Phase (deg) Gain (db) 1 (75.) k 1k k 1k -18 Phase (deg) -1 Fig. 4. Frequency response of tracking actuator (measurement) k 1k transfer function of the signal estimator is introduced to Cðz 1 ÞPðz 1 Þ=f1 þ Cðz 1 ÞPðz 1 Þg, and ^eðkþ is calculated by the block diagram of the ZPET-FF controller shown in Fig. 3. LPF reshapes the waveform and reduces some noise to acquire a clear estimated tracking error signal ^eðkþ as the input signal for the pre-compensator. 7) A high-speed and precise track-following control system using ZPET-FF control is composed in this way. 4. Design and Simulation 4.1 Design of high-speed, accurate controller We designed a more accurate and faster control model and investigated it by comparing the tracking characteristics of the ZPET-FF control model and conventional control models in a simulation. Figure 4 shows the measured frequency responses of the tracking actuator, which has an objective lens of NA ¼ :85. We have modeled the actuator to the second-order transfer function. Since the component of the second resonant frequencies is more than the Nyquist frequency, it is ignored. The model of the tracking actuator is shown in eqs. (1) and (11), where K a is the coefficient of the driving force of the actuator, M is the mass of the moving part, D is the coefficient of viscosity, and K s is the coefficient of elasticity. PðsÞ ¼ ¼ K a =M s 2 þðd=mþs þðk s =MÞ 9: s 2 þð4: Þs þð6: Þ ðm/aþ ð1þ ð11þ Using the tracking actuator model PðsÞ, we designed the FB control system. Figure 5 shows the open-loop characteristics of the FB control system, which was designed to be highspeed and accurate. The FB controller is composed of phaselead controllers. Table I shows the parameters of the FB control system. As shown in the table, this control system was designed as a high-dc-gain, stable controller using the parameters of gain margin and phase margin and a wider servo bandwidth. On the basis of the design of the FB control system, we designed the ZPET-FF control system using more accurate high-speed control models according to the algorisms of the ZPET method. Fig. 5. Table I. Open-loop characteristics of feedback control system. Parameters of the designed feedback control system. DC gain (db) 75. Gain margin (db) 7.6 Phase margin (deg) 27.9 Gain cross frequency (khz) 5.6 Table II. Simulation conditions. Radial run out of optical disk (mm p-p ) 39. Behavior of disturbance of disk eccentricity Sinusoidal during disk rotation period Sampling frequency (khz) Comparison of tracking errors of tracking control methods (simulation) We carried out the simulations of high-speed tracking control with conventional FB control and ZPEET-FF control. Table II shows the conditions of the simulation. The radial run out of optical disks was estimated from the measured value. The sampling frequency corresponded to the experimental conditions. We calculated and monitored the tracking error of each control model. Figure 6 shows the rotation speed residual tracking error of the control models for disk rotation speeds from 3 to 16 rpm. As shown in Fig. 6, the tracking error levels of both controls were gradually increased as much as the disk Residual tracking error (nm) Fig. 6. FB Control ZPET-FF Control Disk rotation speed (rpm) Tolerance Rotation speed residual tracking error characteristics (simulation).

4 Jpn. J. Appl. Phys., Vol. 45, No. 2B (26) 15 1 Red: FB control Green: Repetitive control Blue: ZPET-FF control Gain (db) Phase (deg) k 1k Fig. 7. Characteristics of various tracking control methods (open-loop characteristics). rotation speed was increased. The tracking error of FB control exceeds 9 nm of the tolerance for a Blu-ray disc format beyond about 39 rpm, whereas the ZPET-FF method controls the tracking error to within the tolerance at 15 rpm. The simulation indicates that the residual tracking error of ZPET-FF control is less than that of FB control. Therefore, ZPET-FF control is more precise than conventional FB at high speed. 4.3 Comparison of frequency responses of the tracking control methods (simulation) We carried out a calculation of frequency responses of the following control methods.. FB control. Repetitive control 8,12). ZPET-FF control Figure 7 shows the comparison of the open-loop characteristics of the control methods. Memory sizes of both repetitive control and ZPET-FF control were optimized to a disk rotation speed of 72 rpm in the figure. The red line shows the characteristics of FB control. The green line shows the characteristics of repetitive control. The phase characteristics of repetitive control sometimes show a delay over the value of 18, indicating instability at high frequency. The blue line shows the characteristics of ZPET- FF control. Compared with FB control, the DC gain of ZPET-FF control increased to þ25 db, and reaches about 1 db at 1 Hz. Furthermore, as the size of memory was optimized for the disk rotation speed of 72 rpm, the gain of the frequency corresponding to disk rotation speed increased approximately an additional þ4 db at 12 Hz. Generally, when the size of memory was optimized for a disk rotation speed of x, the gain of the frequency of x=6 was increased in the same way in the ZPET-FF control. (Here, x is the optional number of disk rotation speed.) These analysis and results indicate the effect of extreme error suppression at the disk rotation speed to which the memory size was optimized. Therefore, the figure indicates that the proposed control method performance well, and is fast and robust toward periodic disturbances. 119 Spindle index pulse Fig. 8. Digital servo controller Interface circuits DSP Optical disk 16-bit ADC Optical disk tester 16-bit DAC Configuration of experimental setup. Control input Control output 5. Experiments and Results 5.1 Experiments of high-speed tracking On the basis of simulation results, we evaluated the highspeed tracking characteristics of FB control and ZPET-FF control. The configuration of the experimental setup is illustrated in Fig. 8, consisting of an optical disk tester and an external digital servo controller. The digital servo controller takes the control input signal from the optical disk tester and emits the signal after executing the control algorithm. ADC is an analog-to-digital converter, and DAC is a digital-to-analog converter, both of which have 16-bit resolution. The interface circuits in the digital servo controller takes the spindle index pulse from the optical disk tester and control the optimized size of memory corresponding to the disk rotation speed for ZPET-FF control. Table III shows the specifications of the experimental

5 Jpn. J. Appl. Phys., Vol. 45, No. 2B (26) Table III. Experimental conditions. Laser wavelength (nm) 45 NA.85 Thickness of cover layer (mm) 1 Working distance (mm) 1 Track pitch (mm).32 Data capacity (GB) 23.3 Detection of tracking error Push pull DSP 32-bit floating point, 1GFLOPS, 167 MHz Sampling frequency (khz) 5 DSP bandwidth of FB control (khz) 5.6 setup using a blue-violet laser optical head and phase-change high-density optical disk. The system was externally controlled by the digital servo controller that consists of a digital signal processor (DSP). The DSP was operated at 1 giga floating-point operations per second (GFLOPS), and the matrix calculation representing the state space parameters of the pre-compensator, which has a high-order transfer function, was quickly performed. The sampling period of the digital controller was 2 ms. The tracking error was measured while the disk was rotating at high speed (linear velocity: 18.3 m/s). Figure 9 shows the tracking error signal waveforms evaluated at a linear velocity of 18.3 m/s for conventional FB control and ZPET-FF control. This tracking error signal was measured after 1.8 khz LPF. It was observed at a radius point of 43.7 mm on the disk. The experimental results clearly show the effect of ZPET- FF control. As the disk rotation speed was increased, the tracking error signals of both controls were increased. The Fig. 1. Eye pattern. tracking error level of the FB control reached nm at 18.3 m/s, whereas that of the ZPET-FF control remained 8.2 nm at 18.3 m/s. Therefore, at this speed, ZPET-FF control achieved the necessary level of tolerance for this format of optical disks using a blue-violet laser. 5.2 Test of writing and reading data We also investigated a test of writing and reading data using the ZPET-FF control. Figure 1 shows the eye pattern obtained with 1 7 modulations which were given a random recording pattern with 2T 8T marks. When ZPET-FF control was applied to the optical disk recording system, the recording and playback of data were carried out normally, and it was confirmed that ZPET-FF control was effective for the high-speed servo control of the optical disk drive of the system with NA ¼ :85. Residual tracking error (nm) Residual tracking error (nm) Time (ms) (a) Time (ms) Fig. 9. Tracking error characteristics (experiment, 18.3 m/s): (a) FB control and (b) FB and ZPET-FF control. (b) Conclusions The proposed ZPET-FF control method was shown by simulation and experiment to be superior to the conventional FB control method at high speed in an optical recording system with NA ¼ :85. The high-speed tracking system using ZPET-FF control suppressed the tracking error below 9 nm tolerance level at 15 rpm as compared with the large tracking error of the conventional FB control system in the simulation. By the analysis of frequency response, the ZPET-FF control method was also proved to be robust toward the disturbance of disk eccentricity compared with the conventional control methods. We have experimentally achieved precise track-following control at a high disk rotation speed of 18.3 m/s in the optical recording system with NA ¼ :85. The ZPET-FF control method is effective for high-data-transfer-rate optical recording and playback for broadcast-use optical disk drives and high-speed Blu-ray disc drives. In the future, we will experimentally establish the high accuracy of the ZPET-FF control method at higher disk rotation speed. 1) H. Tokumaru: Proc. 15th Symp. Phase Change Optical Information Storage (PCOS) 23, Atami, 23, p ) SMPTE Standard, 367M-22, SMPTE (22). 3) SMPTE Standard, 342M-2, SMPTE (2). 4) Y. Urakawa and T. Watanabe: Jpn. J. Appl. Phys. 44 (25) 3427.

6 Jpn. J. Appl. Phys., Vol. 45, No. 2B (26) 5) K. Takahashi, K. Tateishi, Y. Tomita and S. Ohsawa: Jpn. J. Appl. Phys. 43 (24) ) I. Watanabe, T. Kawabe and J. Ichimura: Tech. Dig. ISOM, 2, p ) D. Koide, H. Yanagisawa, H. Tokumaru, H. Okuda, K. Ohishi and Y. Hayakawa: Jpn. J. Appl. Phys. 42 (23) ) D. Koide, H. Yanagisawa, H. Tokumaru, S. Nakamura, K. Ohishi, K. Inomata and T. Miyazaki: Jpn. J. Appl. Phys. 43 (24) ) K. Ohishi, Y. Hayakawa, H. Yanagisawa, D. Koide and H. Tokumaru: Proc. American Control Conf. (ACC) 23, Denver, 23, p ) T. Miyazaki, K. Ohishi, K. Inomata, K. Kuramochi, D. Koide and H. Tokumaru: Proc. IEEE/IES Advanced Motion Control (AMC) 24, Kawasaki, 24, p ) M. Tomizuka: ASME Trans. J. Dyn. Syst. Meas. Control 19 (1987) ) T. Katayama, M. Ogawa and M. Nagasawa: IEEE Trans. Consum. Electron. 41 (1995) ) K. Arai, H. Okumura, H. Tokumaru and K. Ohishi: Jpn. J. Appl. Phys. 39 (2)