| Abstract
ID |
|
2935 |
| Title |
|
Real-time Periodic Error Compensation with Low/Zero Velocity Parameter Updates |
| Category |
|
Optics and Interferometry |
| Primary
author |
|
|
| Organization |
|
|
|
| Content |
|
Periodic errors caused by polarizer leakage in displacement measuring heterodyne interferometers can be compensated in real-time [this technique has been presented at previous ASPE conferences]. The approach is to acquire error parameters during target motion in 1 millisecond intervals and apply the current error parameters through digital compensation in the next millisecond, while again measuring the error parameters. In this way, the compensation is applied in a leap-frog manner with 1 millisecond latency. While compensation can be applied at any travel velocity, updating the error parameters currently requires that a minimum travel speed be met or exceeded. There are applications, however, where the minimum speed is seldom or never reached in operation. Therefore, the parameters cannot be updated even though they are continuously changing. An example of this situation is the x-axis interferometer on a stage moving mainly in the y-direction (such as the stage position feedback for a lithographic stepper machine used in semiconductor fabrication). Mirror imperfections cause beam shear in the x-axis interferometer, which alters its periodic error magnitude even though there is effectively no x-travel.
In this work, it is shown that by monitoring the signal strength, it is possible to update the periodic error magnitude even at low/zero velocities because there is an inverse-proportional relationship between the two. Periodic error, as represented on a phase vector diagram, is the angle caused by a perturbing vector appended to the intended vector. The magnitude of the perturbing vector depends on the polarizing leakage within the beam and is constant; it is unaffected by beam shear. The intended signal, on the other hand, is diminished by beam shear. Therefore, by monitoring variation of the intended signal magnitude, it is possible to calculate the variation of the periodic error. This relationship is experimentally verified using a setup which includes the ability to apply known misalignments, measure position, and monitor the frequencies and magnitudes of the interference signal components. |
|
|
|
