Computer-to-plate (CTP) production systems transfer digital data, consisting of text, images, and graphics, directly from computers onto printing plates. This type of data delivery eliminates the need for and high cost of film products and film exposure, stripping, and processing, resulting in inexpensive and faster printing operations.
Removing the film-handling steps from the production cycle also reduces prepress production time and increases product quality. Indeed, CTP systems produce sharper printed materials because plate registration is performed by the setting equipment. This automated production step eliminates the film-stripping errors introduced by human operators.
After plates are produced, they must be customized by locating and punching registration marks on them for use on a designated printing press This registration can be done manually or by a semi-automated punching machine. In the manual process, an operator aligns the registration marks on the plates through a series of optical devices. Once the plate is aligned with the registration marks, the plate is manually punched.
With the semi-automated method, the plate is mounted on a CTP punch machine, such as the one developed by Billows Protocol Ltd. (Wolverton, Milton Keynes, England). The company's CTP Vision Punch, for example, can accurately punch color separations produced without a pin-register system (see Fig. 1).
Modern printing methods use a laser-based image setter to set a desktop publishing image, complete with registration marks, directly onto a printing plate. The plate is later attached to a printing-press cylinder using punched registration holes and hooks to properly align each colored plate. "A common problem for printers has been how to accurately align each plate on the cylinder and make it ready for printing," says John Wickson, electronics engineering manager at Billows Protocol.
FIGURE 2. Before a plate is punched, locations of fiducial marks must be measured and the distance between them calculated to within specific tolerances. To do this, the images captured by the two cameras are displayed on RS-170 CCTV monitors (top). Using an RS-232 controller and a target-positioning system, the operator moves the fiducial marks captured from one camera within a specified bounding box on its CCTV screen (bottom). Then the second camera digitally registers the other fiducial mark.
"In the past, the edges of the plate were used as a datum for punching the registration holes, but this was unreliable and did not give the same results for each plate," he says. The sizes of the plates can vary slightly and handling the plates proves challenging and subject to error. The CTP Vision Punch system uses fiducial marks on the plates to align the registration marks with those of an image of the plate using an RS-170 monitor. After proper registration, the system punches alignment holes onto an area of the plate image that is not used for printing.
In the semi-automatic model of the Vision Punch, an operator manipulates the plate in the punch, and a machine-vision system informs the operator about which way to align the plate. "In the development of the vision system," says Wickson, "two VC-13 machine-vision cameras from Vision Components (VC; Ettlingen, Germany) were mounted on a linear slide." Depending on the size of the computer plate, each camera is manually positioned over the location of the fiducial marks.
If the computer plate is precisely manufactured, these fiducial marks fall exactly within the center field of view of both cameras. In practice, however, these fiducial marks are not precise and may fall to the right or the left of the field of view within a specific tolerance. In other words, the locations of the fiducial marks must be measured and the distance between them calculated to within specific tolerances.
To calculate the tolerances, the two cameras acquire images of the plate and display them on two RS-170 CCTV monitors (see Fig. 2). Using a VC RS-232 communications interface and a VC target positioning system, the operator initially moves the fiducial marks captured by the first camera within a specified bounding box on its associated CCTV monitor. "Each pixel on the monitor screen represents 0.0007 in.," says Wickson, "and the tolerance of the right-hand camera is set tightly to ±0.5 pixel element in the x and y axes." After this digital registration is accomplished, the second camera (with a tolerance of ±5 pixel elements in the x-axis only) digitally registers the other fiducial mark.
Although previous versions of the company's Post Etch punches for printed-circuit boards used off-the-shelf XC-75 cameras from Sony Electronics (Park Ridge, NJ) coupled to PC-based frame grabbers, such as the Bandit from Coreco Imaging (Billerica, MA), the company chose the VC-13 DSP-based camera because of its lower cost. "When we designed the CTP Vision Punch," says Wickson, "we wanted to produce a semi-automated machine at a competitive cost."
According to Wickson, the Sony cameras were programmed in C with the aid of a cross-development system on a commercial PC. To develop the embedded 160-MIPS ADSP2181 processor from Analog Devices (Norwood, MA) used in the 500 × 585-pixel, 25-frame/s cameras, VC supplied a real-time operating system and an image-processing library for the cameras.
Digital inputs and outputs from the camera are monitored and controlled with an Alpha logic controller from Mitsubishi Electric Automation (Vernon Hills, IL). The cameras are automatically armed and activated from the Alpha, and, when alignment is achieved from each camera, the Alpha implements and controls the punch operation.
"The logic controller uses an icon-based, graphical function-block language," says Wickson, "with an instruction set that includes logic gates, such as AND, OR, and NOT." Using the Visual Logic software supplied with the controller, the programmer can "wire" these blocks together to produce control algorithms. The programming software also includes a simulation mode that allows off-line program testing without any hardware connected. "This is useful for pretesting software or verifying proper operation prior to download," adds Wickson.
After a hole in the plate has been punched, the alignment system waits five seconds for the plate to be moved before automatically rearming. Then the cameras repeat the imaging process. "We would like to use an RS-232 link, via a serial port, to enable the two cameras to talk to one another. Currently, each camera does not know that another one is used in the system. If they were linked, we could remove the bias toward one camera and center the control," says Wickson.
Billows Protocol is planning to make a completely automatic punch system. The operator would place the plate on the machine, and the integrated vision and motion controllers would automatically position the plate to be punched. Wickson also intends to use the VC-13 cameras tied to a KV40-DT intelligent programmable logic controller (PLC) from Keyence of America (Woodcliff Lake, NJ).
"The intelligent PLC includes a built-in Access Window, an ac power supply, and an operator interface panel. No PC or hand-held programmer is required to monitor operation or make minor changes," explains Wickson. Instead, the PLC's CPU and built-in display allow connections to the PLC to be checked upon start-up and during modification or changeover. "Using such a distributed computing architecture tied to x/y/q positioning tables will allow the automated CTP punch system to be built at a competitive price," Wickson adds.
Andrew Wilson, Editor
Billows Protocol Ltd.
Wolverton, Milton Keynes, UK
Keyence of America
Woodcliff Lake, NJ
Mitsubishi Electric Automation
Vernon Hills, IL
Park Ridge, NJ