"Air" on the side of smoothness

Semiconductor fab, high-resolution scanning, and high-speed machinery are just some of the applications pushing rolling-element bearings to technical limits. One reason is friction. Friction variations have always been the Achilles' heel of precision positioning systems, particularly when trying to initiate or stop motion precisely.

This is because plain bearings, and rolling bearings to a lesser degree, have a higher static than dynamic friction coefficient. In other words more force is needed to initiate motion than to maintain it. A motor-driven ball screw pushing a slide, for example, winds up before moving. Eventually the slide overcomes static friction then tends to overshoot the target. This stick-slip action is most pronounced in plain-way systems, though rolling-element bearings suffer from it as well.

Rolling-element bearings have a difference between static and dynamic coefficients of friction of about an order of magnitude less than that of plain bearings. For reference, a typical heavy machine-tool slide equipped with rolling-element bearings can be positioned to within 0.0001 in. But certain semiconductor fab equipment requires positioning accuracy better than 0.00001 in. In response, some makers of rolling-element bearings have reduced preload in what has become known as a "California fit" in an effort to meet these requirements, eroding stiffness in the process.

Air bearings are another option, though they have had a reputation for being expensive, delicate, and suitable only for laboratory test rigs. But air bearings made with porous technology are more durable, affordable, and easier to use than earlier designs. First, a quick review of air-bearing basics:

Gliding on air

Air bearings eliminate stick-slip because static friction is zero. Air-bearing friction is a function of air shear rate (from motion) so at zero velocity there is also zero friction, making infinite-motion resolution theoretically possible.

Friction also depends on viscosity of the fluid trapped between the moving surfaces. Viscosity of air is less than a hundredth that of oil which reduces power loss. Most large turbines today use oil-based hydrodynamic bearings, though many new microturbines employ aerodynamic bearings to improve efficiency.

Besides smoothing stop and starts, air bearings offer less resistance (friction) to steady-state motion. Friction generates heat which, in turn, causes spindles and other components to thermally grow, compromising precision. Air bearings generate much less heat in a given application than rolling-element or plain bearings in most cases. In fact, relative speeds must exceed about 100 ft/sec before air bearings generate any significant heat at normal air gaps. Air bearings excel in applications requiring tight velocity control such as scanning and wafer inspection, because they eliminate force ripples from recirculating ball bearings loading and unloading.

No contact, no wear

Advanced, high-speed, high-reliability machines may run a billion cycles annually. It is therefore impractical to do accelerated life testing on such equipment. Engineers typically estimate bearing life based on speed, acceleration, and loads. Air bearings don't require such life calculations because bearing components don't touch each other. Air bearings will work "as new" after billions of cycles. The only mode of wear in air bearings is erosion from the feed air itself so air cleanliness is important.

A lack of wear debris and no oil lubrication makes air bearings well suited for use in clean-room, medical, pharmaceutical, and food-processing applications. Air bearings also excel in dry, dusty, and corrosive environments such as salt or sugar factories where oil lubrication would quickly become lapping slurry. Air bearings instead self-purge and blow away light dry dust.

Stiffness, compensation, and load capacity

A common misconception about air bearings is that they lack the stiffness for precision applications. This is simply not true. For example, a 6-in.-diameter air bearing running at 60 psi supporting a 1,000-lb load has a stiffness exceeding 2,000,000 lb/in. Put another way, that's less than 0.0000005 in. deflection per pound of additional force. Stiffness rises nonlinearly with diminishing film thickness and is proportional to pressure and surface area. However, a factor called compensation also controls air-bearing stiffness.

Compensation is simply the restriction to airflow entering the air gap. The air gap itself must also restrict airflow; otherwise pressure would not build under the bearing. But how can a pressure reserve created by a restriction of airflow into the air gap provide stiffness?

Consider the case where orifice flow at 60 psi equals gap flow with a 150-lb load on a bearing flying above its mating surface at 300 µin. A 2.5-in.-diameter bearing has about 5 in.² of surface area so the average pressure under the bearing face is 30 psi. Increasing load to 200 lb and compresses the air gap to 200 µin., further restricting flow and boosting average pressure to 40 psi. The reserve pressure that had been held back by the orifice now increases gap pressure, creating a restoring force that gives the air bearing stiffness.

The two most common compensation methods are orifice and porous. Orifice compensation distributes pressurized air across a bearing face through strategically placed, precisely sized orifices and grooves. But scratches across a groove or near orifices may cause more air to escape than orifices can supply, causing a bearing to crash at normal air-supply pressures.

Bearings using porous surface compensation, in contrast, issue pressured air from an entire bearing face through millions of micron-sized pores so they are less susceptible to surface scratches. Moreover, where orifice-fed bearings have pressure gradients in the air gap, pressure in porous air bearings remains nearly uniform across an entire surface. Porous bearings also tend to be more stable than orifice-fed types because the porous media damps air flowing into the bearing. Orifice compensation is the most widely used, but porous surface compensation is rapidly emerging as the preferred method.

In short, compensation trades load capacity for stiffness and stability. But even without compensation load capacity is not simply surface area times input pressure. A load equal to this value would ground the bearing. This is because pressure is greater at orifice exits than at bearing edges where air leaks out to ambient pressure. Pressure profiles are influenced by the number and location of orifices and grooves, and especially by the type of compensation used, with porous compensation giving the most consistent pressure distributions.

With compensation factored in, average pressure under a bearing is typically about 50% of input supply pressure. However, both bearing shape and size modify this rule of thumb. Bearings with a high percentage of area near edges, such as small or narrow configurations, will have proportionately lower efficiencies, and vice versa. Further, air bearings don't have the load capacity of rolling-element types. However, air bearings handle similar loads per unit area as traditional plain bearings, making them suitable for high-speed, lightweight machines.