Monday, May 14, 2012

Stepper Scanner Control

If you know me at MIT, your probably know that I really enjoy controls and mechatronics. Even though my degree says "Mechanical and Ocean Engineering" I have done most of not all my requirements studying control theory and practical applications of control.

This blog post is about the controller design for a lithographic stepper. A "stepper" is a essentially a read head that moves in steps on a silicon wafer while the wafer is exposed to UV radiation. The catch is that the error in position during the half second of UV curing is on the order of tens of nanometers. For my digital controls class, we had to come up with a design for the controller structure.

Although we did not have the opportunity to implement the controller in real hardware, our professor included a lot of non-linearities and real world problems and parameters (he even brought in an old stepper to class) which made the problem incredibly interesting.

Here is a picture of the model for the stepper.

yep... its pretty convoluted! but luckily the professor helped us derive the state space model in class and thus we could focus on the control design of the stage.

We used completely classical control techniques (screw LQR and Kalman filters!) including decoupling and trajectory generations to get optimal performance. Most of the analysis was done is simulink, here is a screenshot of the final simulink model!

The final controller took over 36hrs of work to figure out but it worked great (in theory!) and I learned a great ton about design in controllers and how to be able to compensate for disturbances, noise, saturations and more of those great things that pop up in the real world. 

This was an awesome project to do and hopefully a good intro to my maybe job this summer at ASML!

Thanks to Prof. Trumper and all my controlfreak buddies for the collaboration during 2.171!

No comments:

Post a Comment