Controlling the process

Careful consideration when implementing inspection processes yields efficiencies.

Manufacturers continually search for ways to increase profits and reduce costs. Trying to find cost savings where chips are made makes sense and is usually where the most-significant improvements are found. However, efficiency gains can also be made in nonvalue-adding areas, like inspection.

Introducing new parts to production often requires new inspection processes and measuring tools. An inspection process should be considered as carefully as a new manufacturing process. Inefficient, cumbersome and inaccurate inspection devices can hamstring a manufacturing process.

Joel Johnson, a quality engineer at Mitsubishi Hitachi Power Systems, checks calibration stickers on large micrometers used to measure the journals of gas turbine rotors. All images courtesy C. Tate.

For example, a former employer manufactured a family of parts in various lot sizes. Machinists were required to change from producing one part to another throughout the day. The company spent several weeks overcoming setup issues, tooling problems and other inefficiencies. It eventually reduced changeover time from a few hours to 5 minutes.

Unfortunately, the inspection process utilized a scanning technology that is better suited for R&D than production. Scanning each part took 45 minutes, and machinists could not begin the next run until they had received acceptable inspection data. Therefore, the benefit of the 5-minute changeover was not fully realized. The company eventually purchased a small coordinate measuring machine with a tactile-scanning capability to collect the necessary data. This reduced the inspection process from 45 to 5 minutes.

The Usual Suspect

Choosing the wrong measuring tool is a common problem and can cost companies significant amounts of money. For experienced metalworking professionals, selecting the best measuring instrument for a job often feels like a simple task and defaulting to past experience is often viewed as a sure way to achieve success. However, doing so without validation can prove detrimental.

Most of the time our instincts are correct—but not always. In one instance, I recall machine operators needed to measure IDs from 10″ to 14″ (254mm to 355.6mm). They only needed to measure three or four parts per day and tolerances were not very tight, so I selected what I thought were tried-and-true ID micrometers. There was a problem, though, because the parts were not round after machining, and ID mics don’t work well on out-of-round parts. I found the 10″ to 14″ ID mics were difficult to read because the machinists had to place them deep inside the part. Several erroneous measurements were made because the micrometer graduations were difficult to see. These errors, in turn, led to reworking parts that did not need it, along with production delays as we rechecked measurement data.