Nanotechnology and its corresponding unit—the nanometer—are causing some major technical challenges for the controls community as they relate to ultra-precision systems. When I discussed this in my August 2004 column, I addressed accuracy, repeatability, and resolution and gave the "target shooting" example. Two things from that column are integral to understanding what I address here:

1) If you are target shooting, accuracy would be your ability to center on the bullseye regardless of the spread on the target; repeatability would be your ability to cluster shots tightly regardless of their location on the target; and precision would be the size of the holes that you put in the target;

2) There are a billion nanometers (nm) in one meter.

Now, let's look at implementation of an ultra-precision machine that has 1 nm resolution over the range of a meter, or one part per billion. From an implementation perspective, there are a few issues we need to consider. Let's think about the relationship of range to resolution for a minute. If we are controlling location to one part per 1 billion (1 nm) over a single meter, that means we need quite a counter, 30 bit minimum, for our system. If we were using an interferometer or a glass scale linear encoder traveling at a rate of just 1 mm/s, our controllers would see an encoder pulse every 1 µs. If we were traveling at 1 m/s, we would see a pulse every nanosecond. The speed at which the encoders must be read in conjunction with the range of the encoder counts is incredible, and most controllers in use today cannot handle these requirements. Thus, only very high-end controllers are used in such applications.

Let's look at one other control problem that faces ultra-precision manufacturing systems. The thermal coefficient of expansion of steel is about 12 parts per million per degree C. That means a bar of steel 1 m in length will grow 12 µm for every degree Celsius that it is heated. This does not sound like much, but it is huge from a precision engineering perspective. If a machine axis is 2 m (6 ft) long, a single degree increase in temperature will make it grow about 24 µm, or 24,000 nm. Such growth, known as thermal growth, can significantly reduce the capabilities of ultra-precision systems. Thus, motion control is not the only critical control issue in ultra-precision systems, but environmental control is critical as well. In the most precise machines, ambient temperature is controlled to less than 0.1 °C, and machining coolant temperatures may be controlled to an even tighter level.

In the near future, the ability to control systems to ultra-precision levels over long ranges at relatively high speeds is a significant challenge and should be left only to the most powerful controllers. Only the highest-end sensors and actuators will be able to provide feedback and actuation to the levels demanded by ultra-precision systems.

Furthermore, controlling the environment around the system can be as important, if not more important, than the motion controller. As a result, you will not see cheap ultra-precision machines available. However, as technology improves and more capable controllers become available, you can expect the capabilities of new systems to greatly improve without significant cost increases.