Getting the Message About Noise Across, Loud and Clear

Controlling the position of an actuator to within a single atom distance can be a daunting challenge. However, such demands are becoming routine in applications such as scanning probe microscopy [ 1], data storage [ 2], [3], and semiconductor manufacturin g [4]. Good mechanical design and sensor choice are important; however, it is critical to have a comprehensive understanding of the noise processes involved with the system. The sensor noise, input noise, and disturbances will ultimately defi ne the resolution and performance. Being able to quantify noise sources and predict closed-loop resolution is a valuable skill that is sometimes overlooked in undergraduate control and mechatronics education. For a researcher involved with precision mechatronic systems, measuring and quantifying noise is often at the core of the research. For example, nanopositioning systems typically have a range in the tens of micrometers with a resolution down to fractions of an atomic diamet er [5]. To further complicate matters, modern applications are demanding ever-increasing speeds. The speed at which a positioner can operate is typically limited by the sensor noise. For example, a capacitive position sensor with a range of 10 μm typically has a constant noise density of about 10 pm Hz . The root-mean-square (rms) magnitude of the sensor noise can be found by multiplying by the effective bandwidth. For example, with a bandwidth of 100 Hz, the rms sensor noise is