Case Study: Vision for the Future
Examining first pass yields.
Using real-time statistical process control (SPC).
Employing Six Sigma Black Belt engineers who use lean manufacturing protocols to eliminate waste and remove variation from manufacturing processes.
Validating special processes to ensure repeatability, minimize inspection time and reduce cost.
Training employees to continuously improve critical skill sets and integrate quality disciplines into every aspect of the company’s business.
ISO 13485: 2003 certification and FDA QSR compliance.
With quality being so central to everything the company does, one might think that selecting vision-based inspection equipment would be a routine and straightforward exercise. According to Maurice Leger, senior process quality engineer at Tegra’s Dartmouth, MA, facility, this was not the case.
Dartmouth has just upgraded its vision inspection capabilities by adding a new The L.S. Starrett Co.’s (Athol, MA) Galileo EZ200 manual vision system and an automated Starrett Galileo AV 300 multisensor system. This acquisition allowed Tegra to augment or replace reliable older microscopes, comparators and vision systems that have been used at the plant for many years, with systems that are far more accurate, repeatable, and easier to program and use.
Leger, who led the search during which his team evaluated several brands of multisensor systems, explained that new vision-based systems would be involved in diverse applications. These include first piece inspections, traditional in-process part validation, random roving inspections, mold validation studies (for the plant’s newly launched clean room) and Six Sigma manufacturing process qualifications, in addition to generating geometric data and analyses that could be shared across multiple divisions and with the company’s Genesis Tech Center in support of rapid response manufacturing projects. So finding vision-based measurement systems that provide the best fit for all of these present and anticipated uses was a very exacting process.
The primary driver in acquiring the multisensor measurement capabilities was the expansion of the division’s focus from the precision grinding of needle products into Class 8 clean room over-molding of completed products incorporating the needles and subsequent medical device assembly and packaging, including Class III medical devices.
Leger says, “The Galileo AV 300 was bought in parallel with the decision to steer the business toward clean room injection molding. Mold validation studies performed at Tegra Medical involve making tens of thousands of parts very quickly and making hundreds of measurements at well-defined intervals. The object of the game is to determine which settings in the machine I will use to get the best results out of every cavity. For this purpose we needed a system that is fast and easy to program, extremely accurate and capable of automating the measurement of large numbers of parts in a very small window of time.”
Today, with the AV 300 system it takes a few hours to write the program to set up a large mold validation during which the same two or three CTQ (critical to quality) features are measured on a large number of parts. Programming is simple because it keeps all the needed tools, directly on the desktop just as they are with modern computer-aided design (CAD) systems. Because the system feeds all the data into an Excel spreadsheet, transcription errors are eliminated. It takes about a day to measure the parts, collect the data and complete the study. This reduces the mold validation process by days while improving accuracy and eliminating many sources of potential error.
The AV 300 system at Tegra Medical is equipped with a tactile probe that can reach in and collect data from areas of the part where the geometry obstructs the video sensor. In the past, inspectors had to cross section parts to measure these areas.
Cross sectioning is a very labor intensive and delicate procedure that adds hours to the measurement process. Tegra hopes that the AV 300 will dramatically reduce need to cross-section to mold validation and process studies even faster.
While the company was evaluating advanced video measurement systems for its laboratory, it was looking to upgrade the manual measurement capabilities that relied on optical equipment of a previous generation. Operators employ these systems on the shop floor, so they have to be rugged, fast and easy to use. Leger discovered that it was possible to emulate the cross-hair alignment mode of operation used by the older style equipment with Starrett’s Galileo EZ200 manual video measurement system. Therefore, operators got the benefit of higher accuracy, repeatability and more magnification with a system they could be comfortable with after just a few minutes of training.
Because quality operations management spans all of the Tegra Medical’s divisions Leger wanted to confirm that this style of measurement would suit the needs of everyone working on projects that span multiple divisions. So shortly after installing his systems in Dartmouth, he had the quality team from the Franklin, MA, plant come in to look at what he was doing. They liked what they saw. Soon they were asking Leger to do some profile measuring for them. It was not long before the Franklin plant installed a Galileo EZ200 system of its own.
Now Tegra’s Dartmouth facility has very accurate video measurement tools that it can use not only to meet today’s needs, but also adapt to future requirements as the company customizes and fine-tunes manufacturing processes to meet the needs of its exacting customers. Leger says, “When the Galileo AV 300 first came in I told just about everyone we were about to land a spaceship in the middle of the plant. It is much more accurate and capable than systems of the past. At present, we are only using a fraction of those capabilities, but we can easily turn them on or off as needed going forward to accomplish our goals.”
The L.S. Starrett Co.
BenefitsWith the AV 300 system it takes a few hours to write the program to set up a large mold validation during which the same two or three CTQ features are measured on a large number of parts.
Programming is simple because it keeps all the needed tools, directly on the desktop just as they are with modern computer-aided design (CAD) systems.
Because the system feeds all the data into an Excel spreadsheet, transcription errors are eliminated.