Robot specification sheets lead with repeatability, and for good reason. Modern industrial robots return to a taught point within a tenth of a millimeter or better. But a number on a data sheet describes the robot returning to a point, not to the right point.
The gap most spec sheets hide
Absolute accuracy describes how well a robot reaches a point defined in a global coordinate frame. For an off-the-shelf industrial robot, that is generally several millimeters or more. The causes are manufacturing tolerances in each unit, incorrect joint zero-mastering, flexibility, and gear backlash. When you add a tool that was not built exactly to design, a fixture that is slightly off, and possibly external axes, the combined error is larger still.
Why it matters now
More manufacturers want to program robots offline, clone programs between cells, mirror lines, and swap robots without re-teaching. All of these require the real cell to match the model. Without that match, plants fall back on manual touch-up.
What automated calibration delivers
Automated calibration closes the chain between the tool center point and the point on the part. In practice, the robot moves to a variety of positions, a measurement is taken at each, and the discrepancies are identified and compensated. Dynalog's DynaCal system does this for the robot, TCP, and fixture together, and can include external axes and positioners.
Staying accurate
Calibration is typically performed at installation. During production, dedicated in-line tools such as AutoCal and AccuBeam detect and correct accuracy issues so the cell keeps its installed accuracy.
Business benefits
Reduced manual programming time
Better process quality
Standard robots performing tasks usually given to more expensive equipment
Faster launches, cloning, and robot swaps
Bottom line
Repeatability tells you the robot is consistent. Calibration tells you it is also correct.