Camera-based determination of the work offset
Quicker series production
To ascertain the zero point of a workpiece with a high degree of accuracy, touch-trigger probes combined with Sinumerik measuring cycles are the first choice. For smooth, efficient series production, cameras offer a quicker, albeit slightly less precise, alternative.
The digital processing of camera data has been successfully applied to many areas of automation technology. In machine tools, too, cameras can have a multitude of uses. They can monitor and analyze manufacturing processes in areas otherwise difficult to see, or digitally document the production process for training or quality assurance purposes.
Also takes over additional tasks
Another advantage of camera-based zero point determination is that a camera installed in the machine room can take on other tasks. For example, it can be used for blank recognition for NC program selection, for documentation and error analysis of the machining process, for tracking and tracing workpieces based on serial numbers and codes, and last but not least for training and quality assurance.
Object and pattern recognition
A typical example of the new generation of cameras is the Simatic MV440 1-D/2-D code-reading system. Specially developed for industry, it can be used in even difficult machine spaces thanks to its compact design, its flexible lighting options, and its many communication and connectivity alternatives. Combined with a range of software options, the code-reading system can also be used for text and object recognition. To determine a zero point, the camera records as an image the contours of the blank or workpiece before analyzing the information. By comparing the state of the blank with the predefined size of the reference object, the offset and rotation can be ascertained in two dimensions. The automatic correction of the zero point is carried out using ATRANS and AROT in the NC program.
Precalibrating the camera
For the automated measuring process, the camera must be calibrated to a reference part in advance.
- The camera must be mounted axially parallel above the machine table. In the example, the camera is connected to a SINUMERIK 828D via PROFINET.
- The scaling factor of the camera must be known (travel of the axis per pixel, distance between camera and workpiece surface and, if necessary, focal length of the lens must be taken into account).
- The sample workpiece is clamped into the machine.
- A defined Z distance from the camera to the reference part is approached. The position must be within the focus range of the camera.
- The X and Y positions or the zero offset are saved as user data (GUD).
- Creating images of the reference part
- Using the supplied camera software, mark workpiece identification features (e.g. corners, holes) for the measurement.
The coordinates of the recognition features are recorded as pixel values in X and Y and transferred to the SINUMERIK control system. These pixel values can then be converted into travel paths using the previously determined scaling factor.
User-defined dialogs created with "Run MyScreens" facilitate calibration.
Measurements in automatic mode
In automatic mode the camera moves into the precalibrated reference position and creates an image of the blank. The analysis software compares this image with the reference image and determines the offset and rotation of the blank compared to the reference object using the distinctive features. The zero offset is then automatically corrected in the program via ATRANS and AROT.
Very fast or very accurate?
The camera-based approach saves time compared to measuring the position with a touch-trigger probe. Deviations must be taken into account in determining repeat accuracy. This is dependent on lighting, contrast and surface quality. The accuracy of position measurement is +/-0.02 mm, and that of rotation measurement +/-0.2°. For many applications in series production, however, the time gained is sufficient compensation for this loss of accuracy.