CMMs have traditionally performed the quality assurance function in labs and gage rooms off of, or away from, the production floor. Today, however, we are hearing much more discussion about putting CMMs into the factory environment. Indeed, CMMs are finding their way into production with varying degrees of success, but often the implementations are the result of wishful thinking rather than clear-headed rationality. For anyone who wants the benefits of using standard products and software for the flexibility, reusability, and potential costs savings derived from placing a CMM into production, the factors for success require fresh thinking.

Unskilled users, temperature variations, dirt, vibration, speed, and to a lesser extent humidity, are enemies of the CMM. These elements are plentiful on a machine shop floor, whereas they are highly controlled in a gage room or lab.

The "customer" for a production shop-floor CMM is typically a manufacturing manager or engineer, not a quality manager or technician. He/she will be more process-oriented, and less focused on accuracy specifications or all the software can do, thus the CMM may be viewed as an operation sequence in the line, i.e., "Op 40," for example. While the function of the CMM is to ensure that parts remain within production tolerances, in these applications it is as much about monitoring key operations and understanding the real-time effect of process variables on the part dimensions. The customer wants to keep the process in balance, and the CMM is a sensor that monitors that balance. The users are not CMM technicians, but production line operators who may have no more than a few minutes of training to load and run the part. Simple user interfaces are a necessity up front, while the metrology/programming interface and password security protection must reside underneath. Initial training will be forgotten after the few needed programs are in place, so a turnkey approach including programs and fixtures, with ongoing support, is best. The shift in mentality from lab coats to aprons has blindsided more than a few CMM providers.

Machining produces significant heat in a part, moderated slightly by coolants, and thorough soaking time is usually not practical. Most CMMs offer some form of temperature compensation, but these schemes can be highly suspect. The CMM can be modeled in varying temperatures, but a sensor for a workpiece will only report the temperature at the point where it is attached. Most parts are not thermally symmetrical in 3-D, so heat deformations are not evaluated, and may have to be factored in initially. Sometimes a pre-CMM cleaning station will be needed, and this can help to stabilize part temperature. Inspection cycles are typically short, from a few seconds to a couple minutes, so time-relative temperature changes are not an issue. If a plant is not air-conditioned, A/C for the computer/controller may become a necessity, and incoming airflow must be filtered.

Machining fixtures can produce thousands of pounds of clamping force, yet CMM fixtures produce very little holding force. Materials may be deformed when machined "correctly," leaving features distorted when released.

Dirt in the environment will certainly result in erroneous measurements if not removed from the part, but it is a bigger challenge for the CMM. Most suppliers have by now rediscovered mechanical bearings for their plant-floor CMMs because they are more durable, and are better able to handle the inertia of faster speeds and cycles, since air bearings will bottom out. Covers, bellows, and other protections generally do a good job most of the time, but cast-iron machining can be particularly troublesome. Additional protection in the form of housings or enclosures may become inevitable for the sake of the CMM.

Machining, stamping, and other operations produce plenty of vibration, and those processes should be floor-isolated. Still, energy can find its way into the CMM as a source of errors, so some type of vibration isolation is needed. Passive rubber pads or pneumatic doughnut types are good, but active, self-balancing, or floating isolation may not be suitable where external automation like load/unload robots is used.

The CMM rarely stands alone, and must interact with the outside world at some level, if only to release clamps or open a door. CMM hardware/software must have an I/O side that facilitates communication and data export.

The CMM production arena is evolving, and seeking out a supplier with experience and the flexibility to respond to your unique needs will save much grief in the end.