Register control algorithm for high resolution ...

2 downloads 0 Views 2MB Size Report
bSchool of Mechanical Engineering, Konkuk University, Seoul 143-701, Korea. cFlexible Display Roll-to-roll Research Center, Konkuk University, Seoul ...
*Manuscript Click here to view linked References

Register control algorithm for high resolution multilayer printing in the roll-to-roll process Jongsu Leea, Jinwoo Seonga, Janghoon Parka, Sungsik Parka, Dongjin Leeb,c, Kee-Hyun Shinb,c,*

a

Department of Mechanical Design and Production Engineering, Konkuk University, Seoul 143-701, Korea

b

c

School of Mechanical Engineering, Konkuk University, Seoul 143-701, Korea

Flexible Display Roll-to-roll Research Center, Konkuk University, Seoul 143-701, Korea

Abstract Roll-to-roll (R2R) printing process is a key technology for achieving low-cost mass production of multilayer printed electronic devices. A register control, which maintains position accuracy between layers, is essential in the R2R multilayer printing process. In this study, a register control algorithm is developed to achieve microscale resolution in the register. The dominant factor behind machine direction register error is determined. Furthermore, an adequate register model is determined to estimate the register error according to tension disturbances. Permissible tension error is achieved for accomplishing the objective register and minimized for register control. The effect of the developed algorithm on the resolution of the register control is verified experimentally. Experimental results indicate that applying the proposed register algorithm achieved resolution below ±30 µm. This suggests that the proposed register algorithm is an essential application for a microscale resolution register. Finally, an application of the algorithm was introduced using examples considered in previous researches.

Contact information: Kee-Hyun Shin: [email protected]

1. Introduction In the technology of printed electronics, it is very challenging to manufacture next-generation electronic devices such as organic thin film transistors (OTFTs) or organic light emitting diodes (OLEDs) [1]. Flexible printed electronic devices have the advantages of light weight, flexibility, and low-cost production [2-5]; therefore, many studies have attempted to achieve these goals [6, 7] printing is an essential technology for mass producing printed electronic devices. Fig. 1 shows the R2R printing machine considered in this study. It consists of unwinding (a), infeeding (b), 1st layer printing (upstream printing) (c), drying (d), 2nd layer printing (downstream printing) (e), outfeeding (f), and rewinding (g) sections. In the unwinding section, the substrate is unwound and transferred. The transferred substrate is passed through the infeeding section, then printing is conducted by the 1st layer printing roll. During this process, various printing methods may be applied, such as direct gravure [8], gravure offset [9], or flexography [10]. The substrate where patterns are printed by the 1st layer printing roll is dried and cured in a dryer. Hot-air [11] or infrared (IR) [12] heating methods can be applied in this section. After the 2nd layer printing, the substrate on which single or multilayer patterns are printed is rewound. To accomplish the manufacturing of printed electronic devices in the R2R printing process, the ink transfer between the printing roll and substrate has to be optimized [13, 14]. In particular, a register control has to be created, which maintains position accuracy between the patterns printed in upstream and downstream printing sections [15]. Figs. 2(a) and (b) show printed thin-film transistors (TFTs) with and without register error, respectively. As shown in Fig. 2(b), register error causes disconnection between the source and the drain in the printed TFT, generating a serious defect in the device. Therefore, the register control is essential in the production of printed electronic devices that are composed of multilayered patterns; using the register, it is necessary to maintain microscale resolution below 30 µm [16]. Although study on the register control was begun in the graphical printing field [17], more researches have been conducted in the printed electronics. In printed electronics, the high resolution multilayer printing is required for high volume production of printed electronic parts [4, 15]. Brandenburg developed a linear register

model based on the mass conservation law [17]. Yoshida et al. designed a nonlinear register model using the state variable of printing press dynamics for a sectional rotogravure printing press [18]. Liu et al. developed a nonlinear register model of a multilayer gravure printer [19]. Pagilla et al. proposed a register model including variations of span length by the motion of an accumulator [20]. Kang et al. proposed a compensation method for register error using the Brandenburg model [21]. A study on the enhancement of printing location accuracy in a stop-and-go substrate feeding process was conducted by Noh et al. [22]. They designed equipment for precise overlay printing. Moreover, a compensation algorithm was proposed, considering the synchronization error between a flat gravure plate and a blanket roller. Previous studies have indicated that the dominant factors causing register error change according to the overlay printing method and the operating condition of the printing machine. Furthermore, different register models have to be applied to determine the primary reason for register error. This is because the estimation ability of the register models may change according to the primary characteristics of the register error. Therefore, to maintain register control in the R2R printing process, it is necessary to first determine the primary factor in register error and improve it below the range permissible for the register control. In this study, a register algorithm was suggested to determine the dominant factors behind machine direction register error. Furthermore, an adequate register model and its input were determined by the proposed algorithm to calculate the permissible tension error for accomplishing the objective register. The effect of the proposed algorithm was verified by comparing the resolution of register error according to its applications. The experimental results indicated that the algorithm achieved the resolution of register control below ±30 µm. This suggests that the proposed algorithm should be applied to register control for maintaining a highresolution register. Finally, an application of the algorithm was introduced using examples considered in previous researches. 2. Register control algorithm The proposed register algorithm is shown in Fig. 3. First, the most dominant tension error in the operation steps of the R2R printing process is determined by analyzing the tension disturbance using

fast Fourier transform (FFT). It is impossible to measure the register error during operation steps without printing the register marks. As an alternative, the register error is indirectly estimated by the tension error. This because the register error is a function of the strain, so it can be expressed by the elastic modulus and tension in accordance with Hooke’s law. The results determine the primary factor behind register error. Moreover, by considering the characteristics of this primary factor, an adequate register model and its inputs can be determined. Using the model, the permissible error of the factor that has to be improved to accomplish the object register is estimated. Next, improvement is conducted. Finally, the register error is compensated by the designed closed-loop proportional integral (PI) controller. 2.1 Mathematical models for register control 2.1.1 Brandenburg’s model In Brandenburg’s model, the register error was calculated by the strain of the substrate, which is a function of the velocity of the printing roll. The machine directional register error was defined by Eq (1). Considering the strain in the 1st printing section based on mass conservation law, the register error was calculated as shown in Eq (2) [17, 21].

rx (t) x w (t) x r (t) R x (t)

v 20

2

(t) v10

1

(1)

(t

)

(2)

where x w (t) is the machine directional distance between the patterns printed on the substrate in the 1st and 2nd printings; x r (t) is the intended offset with which two register markers on each layer should be spaced; R x stands for the time derivative of changes in the register error; v10 and v20 are the velocity of the 1st and 2nd printing rolls, respectively; printing sections, respectively; calculated as follows:

1

and

2

are the strain in the 1st and 2nd

is the transporting time between two printing rolls, which can be

L2 v20

(3)

where L2 is the span length of the 2nd printing section. Using this model, the register error resulting from the strain due to roll eccentricity and velocity disturbance was fully estimated [21]. However, it was not estimated accurately by the thermal strain during drying process, because a correlation has not been elucidated between the thermal strain of the substrate and the corresponding tension disturbance. 2.1.2 System modeling by system identification In the R2R printing process considered in this study, it was very difficult to estimate the register error attributable to the thermal strain due to difficulties in defining the temperature behavior of the substrate before, during, and after the drying process. Therefore, the R2R printing system was modeled experimentally by system identification (SI). SI is used to estimate the future behavior of a system by using the past inputs and outputs of the system. Of the various SI techniques, autoregressive exogenous input (ARX) is the one of the most useful prediction models [23]. The structure of the ARX model is as follows:

ˆ y k 1 A 1

yˆ k t

ˆ y k 2 A 2

t

ˆ u k 1 dˆ B 1

t

t

ˆ y k n ... A n

ˆ u k 2 dˆ B 2

t

t

ˆ u k n dˆ ... B n

(4)

t

ˆ and B ˆ (x 1, 2,...., n) are the coefficients of the measured output and input, that is, where A x x y k t and u k t , respectively; yˆ k t is the predicted output at the kth sampling;

t is the

sampling time; and dˆ is the number sampling times corresponding to the time delay. Eq (4) becomes Eq (5) through z-transform [23]:

y z

ˆ y z z A 1

1

ˆ y z z A 2

Rearranging Eq (5) yields

2

ˆ y z z ... A n

n

ˆu z z B 1

1 dˆ

ˆ u z z B 2

2 dˆ

ˆ u z z ... B n

n dˆ

(5)

ˆ z 1 A 1

1

ˆ z2 A ˆ z 3 ... A ˆ z A 2 3 n

n

y z

ˆ z 1 dˆ B ˆ z2 1 B 1 2



ˆ z3 B 3



ˆ z ... B n

n dˆ

u z (6)

ˆ ˆ A(z)y(z) B(z)u(z) As shown in Eqs (4) to (6), the ARX model is obtained by the correlation between the behaviors of input and output. Therefore, the estimation ability of the model may change according to the input of the model. It is critical to determine the input of the model. The input of the ARX model has to include the factor that dominantly affects the output. Furthermore, the stability of the derived model has to be considered to determine the order of the obtained model. 2.2. Investigation of the dominant factor of register error in R2R printing process The operation of the R2R printing process is divided into four steps: tension setting and nip roll contacting, substrate transfer with reference velocity, temperature setting, and printing. First, reference tension is applied to the substrate, and a rubber roller, designated the nip roll, is brought into contact with the driven roll under the reference pressure (step 1). This step is necessary to prevent slippage between the driven roll and the substrate. Then, the substrate is transferred with the reference velocity (step 2). The velocity is determined by drying and curing times of the printed patterns. Subsequently, the temperature in the dryer is set (step 3). During this step, drying and curing are conducted by hot air. Finally, the designed patterns are printed (step 4) and register control is performed while the 2nd layer pattern is printed. During the operating steps, tension and register error can be generated by flawed components in the R2R printing process. During steps 1 and 2, register error can be generated by velocity disturbances of the driven roll due to inadequate load cell and motor controller settings. Moreover, roll eccentricity and synchronization errors between driven rolls can cause register errors. During steps 3 and 4, temperature error in the dryer causes thermal deformation of the substrate. Additionally, the excessive pressure of the nozzle that convects hot air in the dryer may deform the substrate. To determine the most dominant factor casing machine direction register error, the tension errors due to the disturbances generated in the aforementioned steps were analyzed by FFT. During the experiment, the operating tension, velocity, and drying temperature are 2 kgf, 5 m/min,

and 80°C, respectively. Figs. 4(a) and (b) show the tension disturbances in steps 1 to 3 in the frequency domain. In this figure, the DC component of the tension signal, which means the average of all signals to which the FFT was applied, was not removed from the raw data of the tension before the FFT. This was for clearly visualizing the distribution of the tension measured in the various scenarios. The frequency ranges of the measured tension and that in the inner figure are 0 to 0.5 Hz and 0 to 0.06 Hz, respectively. This figure shows that tension errors of 0.034 (0.35 Hz) and 0.025 kgf (0.47 Hz) were generated during step 2. In particular, the tension error of 0.257 kgf occurred during step 3, which is 8 times greater than that in step 2. On the other hand, the tension errors occurring during step 2 decreased to 0.016 and 0.018 kgf, respectively. However, there was no change in the register error during this step, even though the tension error decreased. The variations in tension are attributable to the decrease in elastic modulus at high temperature, not to variations of the strain. These results indicate that the greatest register error is dominantly generated by the thermal effect in the dryer (step 3). The tension and corresponding temperature in the dryer in the frequency domain are shown in Fig. 5. The frequency range is measured from 0 to 0.03 Hz and left and right sides of the y-axis represent the tension and temperature, respectively. The red and green boxes in the figure represent the maximum values of temperature and the corresponding tension disturbance, respectively. This figure clearly shows that the frequency of temperature fluctuation corresponds to that of tension error. This means that the tension error is caused by temperature disturbances in the dryer. It is clear that the tension error from thermal deformation of the substrate is the most dominant factor that affects the register error in the R2R printing process considered in this study. 2.3. Estimation of register error by thermal strain using identified SI model To optimize the estimation ability of the model, the tension, including the thermal effect, was selected as the input of this system. Fig. 6(a) shows the pole and zero locations of the various order models. The pole and zero of the models, which are denoted by “x” and “o,” respectively, are located in the z-plane. The model is stable when the pole is placed in the unit circle of the z-plane. Furthermore, the order of the designed ARX model is expressed as a three-digit number: the first,

ˆ ˆ second, and third numbers represent the order of A(z) , B(z) , and dˆ , respectively. As shown, the first (ARX112) and second order (ARX231) models are stable. The estimation ability of the various order models is shown in Fig. 6(b). In this figure, red, blue, and green lines represent the measured distance between the register marks and the distance estimated by the first and second order models, respectively. As shown, the estimation ability of the second order model is superior to that of the first order model. Therefore, the second order ARX model is determined as the most adequate model for register estimation, which was modeled as

y(z) u(z)

z

3

0.7973 0.7536z 2 0.02053z

(5)

Using the developed model, the permissible tension disturbance by the thermal effect was determined to achieve the objective register. In this study, the desired register was determined at ±30 µm, considering previous studies [16]. The periodical tension errors from 0.01 to 0.3 kgf were applied. The period of the errors is 0.002883 Hz, considering the actual period of tension error by thermal deformation, as shown in Fig 5 (b). The applied tension errors and corresponding register error estimated by the developed model are shown in Fig. 7. As shown, the register error decreased as the tension error by the thermal strain decreased. Furthermore, it decreased below ±30 µm at a tension error below 0.01 kgf. Figs. 8 (a) and (b) show the temperature fluctuations in the dryer and the corresponding tension disturbances, respectively. The temperature fluctuation was improved by tuning P, I, and D gains, which are the parameters of the proportional integral derivative (PID) feedback controller in the dryer. As shown, the temperature fluctuations improved from ±7.5 to 0.3°C by the gain tuning. Furthermore, the tension disturbance decreased from 0.257 to 0.01 kgf as the temperature fluctuation decreased.

2.4. Process of register control Fig. 9 shows schematics of the register control process considered in this study for TFT printing. The register controller considered in this study consists of an optical sensor, a camera, two programmable logic controllers (PLC1, 2) and register control software. In this process, the register marks and corresponding patterns are printed simultaneously. The distance between the 1st and 2nd printed

patterns is derived by measuring the gap between the corresponding register marks. A trigger mark is printed in the 1st printing with the 1st layer pattern and register mark. As the trigger mark passes through the optical sensor, it is recognized by the sensor. The optical sensor sends a signal to the PLC1, and then the PLC1 transfers the signal that has a defined waiting time data to the camera. After the waiting time, the camera captures the image of the two register marks printed by the 1st and 2nd printing rolls, respectively. In the captured image, register error is calculated by register control software, and then the compensation value applied by the PI control is sent to PLC2, which manages the R2R printing process. Finally, PLC2 sends the compensation velocity to the motor driver, and the motor connected with the 2nd printing roll is rotated according to the compensation velocity. 2.5. Effect of the register algorithm on the resolution of register control The effect of the developed register algorithm was verified by examining the resolution of the register control according to the algorithm application. In the experiment, operating velocity and tension in printing section are 5 m/min and 2 kgf, respectively. The drying temperature is 80°C. Fig. 10(a) shows the changes in distance between the two register marks according to the application of the developed algorithm. In this figure, initial and desired distances are 12 and 12.5 mm, respectively. The triangle mark denotes the distance variation by the register control without the developed algorithm. Moreover, the rectangular mark stands for the register control to which the algorithm is applied. This figure shows that ±30 µm resolution of register control is achieved by applying the algorithm. On the other hand, the distance is not converged to the desired value, but fluctuates when the algorithm is not applied, even though P and I gains are the same as in the former case. The magnitude of temperature fluctuation, which is the dominant factor of the register error in this process, causes the register control range to deviate. Fig. 10(b) shows the representative images and corresponding distances of the register marks for the case in which the proposed algorithm is applied, as shown in Fig. 10(a). The bottom arrow represents the direction of the substrate transfer. The time under the corresponding image represents its captured time. The filled rectangular and diamond shapes represent the 1st and 2nd register marks, respectively. Furthermore, the shapes with white edges are the register marks in the desired distance. These results suggest that the developed algorithm should be applied for high-

resolution register control. 2.6. Application In this section, an application of the proposed algorithm is introduced. Previous researches showed some defects in the printed electronics by the register error due to the shrinkage in the substrate. It was improved by design of printed pattern, which is robust to the machine direction and cross-machine direction [4, 24, 25]. In the case, the algorithm can be useful to increase the successful printing length by decreasing disturbances that cause the register error. Using the algorithm, the disturbances including shrinkage can be established. The permissible range of amplitude of the disturbances can be calculated to achieve the desirable register performance. The desirable register performance can be achieved by decreasing the disturbances to the permissible level, which make the successful printing possible in the high volume production of printed electronics. 3. Conclusion In multilayer printing using the R2R printing process, it is essential to achieve microscale register control. In this study, a register algorithm is proposed to investigate the dominant factor affecting register error and minimize it for register control. By using the algorithm, the adequate register model and its input can be also be determined to estimate the permissible error of the dominant factor for achieving register error of ±30 µm. The effect of the algorithm is verified experimentally by comparing the resolution of register control before and after applying the algorithm. The experimental results indicate that the ±30 µm resolution of register control is achieved by applying the algorithm. On the other hand, the register is not converged but fluctuates when the algorithm is not applied. From this study, it is clear that the developed register algorithm is essential for application to highresolution register control. The proposed algorithm can be useful to establish the disturbances which generate the register error and decrease it to a permissible range for achieving a desirable register performance.

Acknowledgment This paper resulted from the Konkuk University research support program.

Reference [1] S.H. Kim, K. Hong, W. Xie, K.H. Lee, S. Zhang, T.P. Lodge, C.D. Frisbie, Electrolyte Gated Transistors for Organic and Printed Electronics, Advanced Materials, 25 (2013) 1822-1846. [2] F.C. Krebs, T. Tromholt, M. Jørgensen, Upscaling of polymer solar cell fabrication using full roll-to-roll processing, Nanoscale, 2 (2010) 873-886. [3] M. Takamiya, T. Sekitani, K. Ishida, T. Someya, T. Sakurai, Large area electronics with organic transistors, Applications of Organic and Printed Electronics, Springer, 2013, pp. 101-113. [4] F.C. Krebs, N. Espinosa, M. Hösel, R.R. Søndergaard, M. Jørgensen, 25th Anniversary Article: Rise to Power–OPV Based Solar Parks, Advanced Materials, 26 (2014) 29-39. [5] W.J. Hyun, E.B. Secor, M.C. Hersam, C.D. Frisbie, L.F. Francis, High Resolution Patterning of Graphene by Screen Printing with a Silicon Stencil for Highly Flexible Printed Electronics, Advanced Materials, 27 (2015) 109-115. [6] J. Perelaer, P.J. Smith, D. Mager, D. Soltman, S.K. Volkman, V. Subramanian, J.G. Korvink, U.S. Schubert, Printed electronics: the challenges involved in printing devices, interconnects, and contacts based on inorganic materials, Journal of Materials Chemistry, 20 (2010) 8446-8453. [7] S.H. Ko, H. Pan, C.P. Grigoropoulos, C.K. Luscombe, J.M. Fréchet, D. Poulikakos, All-inkjet-printed flexible electronics fabrication on a polymer substrate by low-temperature high-resolution selective laser sintering of metal nanoparticles, Nanotechnology, 18 (2007) 345202. [8] C. Lee, H. Kang, C. Kim, K. Shin, A novel method to guarantee the specified thickness and surface roughness of the roll-to-roll printed patterns using the tension of a moving substrate, Microelectromechanical Systems, Journal of, 19 (2010) 1243-1253. [9] T.-M. Lee, J.-H. Noh, C.H. Kim, J. Jo, D.-S. Kim, Development of a gravure offset printing system for the printing electrodes of flat panel display, Thin Solid Films, 518 (2010) 3355-3359. [10] G. Schmidt, M. Bellmann, B. Meier, M. Hambsch, K. Reuter, H. Kempa, A. Hübler, Modified mass printing technique for the realization of source/drain electrodes with high resolution, Organic Electronics, 11 (2010) 1683-1687. [11] M. Allen, C. Lee, B. Ahn, T. Kololuoma, K. Shin, S. Ko, R2R gravure and inkjet printed RF resonant tag, Microelectronic engineering, 88 (2011) 3293-3299. [12] D. Tobjörk, H. Aarnio, P. Pulkkinen, R. Bollström, A. Määttänen, P. Ihalainen, T. Mäkelä, J. Peltonen, M. Toivakka, H. Tenhu, IR-sintering of ink-jet printed metal-nanoparticles on paper, Thin Solid Films, 520 (2012) 2949-2955. [13] H.A.D. Nguyen, J. Lee, C.H. Kim, K.-H. Shin, D. Lee, An approach for controlling printed line-width in high resolution roll-to-roll gravure printing, Journal of Micromechanics and Microengineering, 23 (2013) 095010. [14] R. Kitsomboonloha, S. Morris, X. Rong, V. Subramanian, Femtoliter-scale patterning by high-speed, highly scaled inverse gravure printing, Langmuir, 28 (2012) 16711-16723. [15] J. Noh, D. Yeom, C. Lim, H. Cha, J. Han, J. Kim, Y. Park, V. Subramanian, G. Cho, Scalability of roll-toroll gravure-printed electrodes on plastic foils, Electronics Packaging Manufacturing, IEEE Transactions on, 33 (2010) 275-283. [16] H. Kipphan, Handbook of print media: technologies and production methods, Springer, Berlin, 2001. [17] G. Brandenburg, New mathematical models for web tension and register error, Proceeding of the Third International IFAC Conference on Instrumentation and Automation in the paper, Rubber and Plastics Industries, 1976, pp. 24-26. [18] T. Yoshida, S. Takagi, Y. Muto, T. Shen, Register control of rotogravure printing press. Application of nonlinear control theory to sectional drive printing press, Electronics and Communications in Japan, 94 (2011) 17-24. [19] S. Liu, X. Mei, J. Li, L.e. Ma, Machine Directional Register System Modeling for Shaft-Less Drive Gravure Printing Machines, Mathematical Problems in Engineering, 2013 (2013). [20] A. Seshadri, P.R. Pagilla, J.E. Lynch, Modeling Print Registration in Roll-to-Roll Printing Presses, Journal of Dynamic Systems, Measurement, and Control, 135 (2013) 031016-031016. [21] H. Kang, C. Lee, K. Shin, Modeling and compensation of the machine directional register in roll-to-roll printing, Control Engineering Practice, 21 (2013) 645-654. [22] J.-H. Noh, I. Kim, S.H. Park, J. Jo, D.S. Kim, T.-M. Lee, A study on the enhancement of printing location accuracy in a roll-to-roll gravure offset printing system, The International Journal of Advanced Manufacturing Technology, 68 (2013) 1147-1153. [23] S.W. Sung, J. Lee, I.-B. Lee, Process identification and PID control, John Wiley & Sons, 2009. [24] M. Hösel, R.R. Søndergaard, M. Jørgensen, F.C. Krebs, Fast Inline Roll to Roll Printing for Indium Tin Oxide Free Polymer Solar Cells Using Automatic Registration, Energy Technology, 1 (2013) 102-

107. [25] P. Sommer Larsen, M. Jørgensen, R.R. Søndergaard, M. Hösel, F.C. Krebs, It is all in the Pattern— High Efficiency Power Extraction from Polymer Solar Cells through High Voltage Serial Connection, Energy Technology, 1 (2013) 15-19.

Figure captions Fig. 1. Photo of R2R printing system considered in this study: (a) unwinding, (b) infeeding, (c) 1st printing, (d) drying, (e) 2nd printing, (f) outfeeding, and (g) rewinding sections. Fig. 2. Schematics of register error in printed TFT: (a) no register error and (b) register error. Fig. 3. Proposed register control flowchart. Fig. 4. Tension distribution of the printing section in the frequency domain: frequency ranges of (a) 0 to 0.5 Hz and (b) 0 to 0.06 Hz. Green, red, and blue lines represent the tension distribution during the steps of tension setting and nip roll contacting (step 1), substrate transfer with reference velocity (step 2), and setting temperature (step 3), respectively. Fig. 5. Temperature distribution in dryer and corresponding tension in the frequency domain. Left and right sides of the y-axis are the tension and temperature, respectively. The red and green boxes in the figure represent the maximum values of temperature and corresponding tension disturbance, respectively. Fig. 6. Pole-zero plot and estimation ability of the designed SI models: (a) pole-zero plot of the various models and (b) estimation ability of the first and second order models Fig. 7. Permissible tension error estimation: (a) applied tension errors and (b) estimated register error under the various tension disturbances. Fig. 8. Temperature fluctuation and corresponding tension disturbances: (a) temperature fluctuation and (b) tension disturbance by the temperature error. Fig. 9. Register control in R2R printing process: the register controller considered in this study consists of optical sensor, camera, two programmable logic controllers (PLC1, 2), and register control software. Fig. 10. Effect of the proposed register algorithm on the resolution of the register control: (a) changes in the distance between the two register marks according to the application of the developed algorithm and (b) representative images of the register marks controlled by the register controller to which the developed algorithm is applied. In (a), initial and desired distances are 12 and 12.5 mm, respectively.

Fig.3 Click here to download high resolution image