provided by: 
Statistical process control determines whether a process is stable or changing by taking samples of historical data and predicting performance based on pre-set control limits. SPC software can automate that data collection and interpretation to help predict problems before they occur.
Fair enough.
Yet a common challenge with SPC has remained for years. Different machine controllers and software feed data in different file formats, with various headers, trailers, right or left margins, invisible characters and so on. The challenge has been getting all these formats to agree, to make comparing apples to apples possible.
DYNAMIC MASKING
"In general, we still have to deal with a large variety of equipment and communicate in a variety of ways," says Frank Tappen, vice president of solution delivery of SPC-software-vendor DataNet Quality Systems, Southfield, Mich.
Even within the same basic file formats issues still arise, says DataNet President Ned Greenberg. "The layout of that file could be tab or comma delimited, or other organizational methods."
Consider the coordinate measuring machine. To be used with SPC software, different CMM files must be "masked" to ensure SPC reads the correct data-points to, again, compare apples to apples. One common method involves "positional masking," which instructs the software reading the file format to only look at data entered at specific positions and character lengths, such as "4th row, 16th column, 10 characters long." Another method, dubbed "key identifier" masking, instructs software to read only certain words, such as, "Look for the word 'Datum 1X,' and whatever follows is the measurement."
This by itself can be very time-consuming. Add to that the common tweaks to part programs on CMMs, and the problem becomes worse. CMM files can involve literally hundreds of dimensional data points for a single part, so it's little wonder end-users demanded an automated solution.
"Users complained that they could import spec limits into SPC software, but they couldn't always count on the locating points being in the same spot every time," says Tappen. "Typically, CMM operators will insert or remove a datum if they receive a new feature to test. That change moves everything around" within the file.
One solution involves a technology called dynamic masking in which software algorithms detect patterns among different file types. It parses out and formats certain information to be used within a plant- or department-wide SPC program.
Parsing utilities can detect part names, specifications and other raw data in CMM-brand-specific output files required for SPC data sampling; operators no longer must manually re-mask CMM output files for different setups and configurations. It takes brand- and technology-specific files from CMMs and draws like information from all of them—ensuring the SPC receives consistent data formats for statistical sampling.
BEYOND CMMs
A need to collect data from different technologies doesn't lie only within CMMs. Quality data from one machine can directly affect the operation of another. For instance, one machine may drill holes onto plate but with a slightly off initial hole location. Now consider a rod assembly, machined separately, designed to fit in those holes. If the holes are slightly off, SPC software can see it and send instructions to make adjustments to the rod assembly, so the final product will exceed customer quality specifications—despite the initial process error—and bad parts are minimized.
"By itself, the drilling machine can't do that," explains Tappen. "It is only focused on the immediate part it's cutting. This is where software comes into play. It can see the historical data from any piece of equipment on the floor."
Yet such an ideal still requires the masking of myriad file formats from various systems. In the long-term view, automated masking could be called a first-step solution to a larger problem: the lack of interoperability between technologies on the shop floor.
THE QUEST FOR STANDARDIZATION
Tappen stays involved with organizations like the Automotive Industry Action Group that trumpet the need for standardization and interoperability, especially among inspection equipment. As part of the effort, industry has developed a dimensional markup language, or DML, that would serve as a standard format for all inspection files.
For the end-user, such standardization would bring big savings, with no translations and no masking of files required. Says Greenberg, some studies estimate the automotive tooling industry alone spends $450 million a year linking disparate shop-floor systems together.
Yet as with any technological standardization, competing factions exist. Standardization can lead to commodization, which, in turn, hinders innovation. The free market spurs innovation that differentiates proprietary inspection platforms. Those give unique selling points that lead to sales and profit growth, which in turn leads to more innovation. Without that profit, innovation diminishes.
Nevertheless, Greenberg explains, users will be the ones who ultimately move the industry toward file-format standardization, but he expects it may take years to happen. In the meantime, current technology can provide an automated solution so that various file formats are masked properly—and the final data elements used compare apples to apples.
Editor's Note: Artwork courtesy of DataNet Quality Systems, www.winspc.com