Radiant energy – Photocells; circuits and apparatus – Optical or pre-photocell system
Reexamination Certificate
2001-10-29
2004-07-27
Luu, Thanh X. (Department: 2878)
Radiant energy
Photocells; circuits and apparatus
Optical or pre-photocell system
C250S230000
Reexamination Certificate
active
06768100
ABSTRACT:
TECHNICAL FIELD OF THE INVENTION
The present invention relates to positional calibration of servo-motor driven systems; and more particularly to an integral, continuous calibration system of the position sensor circuit of a galvanometer.
BACKGROUND OF THE INVENTION
Galvanometer scanners are often used, either singly or in multiples, to point a light beam with high resolution, linearity, and repeatability. As an illustrative example of a demanding application, a pair of galvanometers arranged in Cartesian coordinates cooperate to point a laser beam over a solid angle of 30 degrees to a precision of 1 micro-radian or less anywhere in that field of view. The accomplishment of this task requires that the system be carefully calibrated in advance, to remove all the geometrical errors in mounting of the parts, and correct all of the residual non-linearity in the position detectors used to provide the feedback error signals to the servo system.
This calibrating is often done by commanding a series of positions in the field of view, recording the actual positions achieved, measuring the actual positions, and generating a set of correction factors by desired field position which are then combined with the command signal in such a way that the final position corresponds to the intended command. This set of correction factors is often stored in a look-up table. Such a table is constructed with rows and columns of cells, each representing a solid angle position. In each cell is stored one or a pair of correction values to be utilized when a measured position coincides with the cell's field position.
Due to the complexity of the overall system of which the galvanometer scanners are but a part, and because the task for which the system is designed is usually a repetitive production task, such as drilling 1000 via holes per second in pre-wired boards or printed circuit boards, it is desirable for operating cost considerations that the system, once calibrated, operate continuously around the clock for extended periods without the need for periodic time consuming maintenance or adjustment.
Since a two dimensional, 30 degree solid angle field of view contains approximately 2.5×10{circumflex over ( )}11 resolvable points, the calibration process is complex and very time consuming. The calibration is more accurate when a larger number of points are utilized, but the increased number of points equates to more time and expense. Although by necessity carried out with the aid of high-speed data processing equipment, the calibration process usually takes several hours to complete. The calibration process is lengthy and tedious, but provides an accurate means of ensuring that the actual position is the same as the commanded position—at least as of the time the angular position was calibrated. For angles between the calibrated points simple interpolation is used.
Unfortunately, the galvanometer scanners are inherently incapable of maintaining the linearity and precision of their position detectors over long periods of time. Nor are they immune entirely to the influences of change in temperature and relative humidity in their operating environment. As a result, the galvanometer or galvanometers begin to drift away from their calibrated condition immediately after calibration, and eventually again produce errors in pointing that offends the limits of accuracy required of their operation. Because of the high-speed production of parts that is the purpose of the system, it is often the case that a considerable quantity of scrap has been produced before the out of tolerance condition is detected.
A number of calibration techniques have been used in the past to re-establish the angular relationship to account for the drifts. One such method, termed in-field fiducials, uses detectors positioned in the field of view with special locations defined by X and Y coordinates. The differences are translated into factors that are stored in the machine circuitry and used to recalibrate the system from time to time. Others use a sample product every hour to define error values by physical measurement, and plot the deterioration of performance. It is also possible to employ fences as thresholds to determine when to recalibrate.
But, in a production run on a system with a capital cost that may approach one million dollars, stopping production in order to calibrate a relatively cheap component is not cost-effective or desired. The production machines need to run continuously, day and night for seven days per week, in order to be efficient.
As stated, there are a number of ways to recalibrate the system. The in-field fiducials are not part of a galvanometer head but are part of the overall machine. The in-field fiducials are targets, light detectors that signal when illuminated. About once a minute the system makes a measurement of the target sensors. There is processing required to compute the error amount, which requires some computational time as well as computer resources. Direct position error of load is achieved by this method.
Although in-field fiducials can be designed and manufactured as part of a new system, it is very difficult to upgrade or convert an existing system after the fact, because of the high degree of precision in positioning the target sensors remotely from the galvanometer head. Even if the targets/fiducials are placed inside the galvanometer head and look at the back of mirrors, an alternate configuration that has been tried, this is an intermediate step and still requires stopping and running a separate procedure and taking processing time to compute the calibration factors. Finally, all this does is calibrate the load with respect to the head.
Besides the calibration to resolve individual galvanometer characteristics, there are latency issues. The latency issues arise because the acceleration and maximum speed of galvanometers are limited. For example, the time to go from point A to point B is a time T. But, the time to go from point A to point
2
B, is not
2
T. It is necessary to calibrate these motions so the time intervals of a large number of points are measured and interpolation is used for points in-between the measured points.
For illustrative purposes, suppose a command signal versus time, such as a step function position command that lasts for some arbitrary time, is injected into a galvanometer scanner system. The signal has infinite slope, which only occurs in ideal and not practical agreements. Because of inertia, the system can not respond instantly—it accelerates as the command is applied. There is some latency because it takes time for the system to detect the command and the magnitude level. At some future time the position of the load sensor reaches the desired position.
In general, because of inertia—similar to a mass spring system—the system acts as a tortional spring on each end of the shaft with respect to the motor. The ideal situation seeks to minimize latency and settling, and the stiffer the system the more ideal the system performance.
The calibration or recalibration process is typically done by calibrating the galvanometer before or during a pause in the manufacturing process. It is necessary to build the head and set up the system to perform a point by point array in the field of view. An average sequence may start with 64 points consisting of corners and middle points, measuring these points by the various methods known in the art. The field of view may consist of 10
6
points, so when a particular point is commanded it is necessary to interpolate from the look-up table to obtain the corrected position. The number of resolvable points are much greater than the calibrated number of points and it is therefore necessary to interpolate from the look-up table to obtain the best fit gain and offset.
Gain and offset are the two components or factors that control where in the box or field of view the command is pointing. From initial calibration measurements, initial calibration data is converted into gain and offset components. A look-up table is generated in ord
GSI Lumonics Corporation
Luu Thanh X.
Maine & Asmus
LandOfFree
Continuous position calibration for servo controlled rotary... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Continuous position calibration for servo controlled rotary..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Continuous position calibration for servo controlled rotary... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3237111