Design and measurement tools shorten production lead times

This article originally appeared in Plastics Today magazine.

Three-dimensional scanning techniques combat shrinkage and warpage issues under the hood.

When Miniature Precision Components Inc. (MPC), a high-volume plastics injection molder in Walworth, WI, began making shield-like engine covers for Ford’s 2007 Lincoln MKZ, the inevitable shrinkage and warpage issues presented challenges prior to production. Using both laser scanning and point-cloud-based inspection, savvy managers at MPC quickly turned the solution into a competitive advantage.

In any molded part, as-molded final dimensions are affected by shrinkage and warpage, which is generally dealt with in one of four ways:

  • Changing the molding process.
  • Opening up the part’s tolerances.
  • Building a cooling fixture.
  • Modifying the tool that forms the part.

Uncontrolled shrinkage and warpage can leave as-molded surfaces several millimeters out of specification. Gregory A. Clark, manager of MPC’s Quality Assurance Measurement Div., points out that dimensions for the engine cover needed to be within a tolerance budget of ±5 mm (about .200 inch) to accommodate process variation; larger deviations are allowed for unsupported areas of the component.

Time is very important to both MPC and its OEM customers. Implementing any of the bullet-point remedies listed above consumes time. “Within minutes, scanning gives us all the surface data we need to choose among the production options,” says Clark. “This gives us new competitive advantages in uncompromising quality assurance, as well as speedy starts on production.
MPC uses scanning and point-cloud-based inspection systems from two Canadian companies ShapeGrabber Inc. of Ottawa, ON (laser scanners) and InnovMetric Software Inc. of Quebec City, QC (PolyWorks/Inspector 3D point-cloud inspection and reverse engineering software). ShapeGrabber and PolyWorks output in a verifiable report that goes back to the toolmaker and gaugemaker for any modifications, as well as to the makers of cooling fixtures.

Controlling shrink and warp

The difficulty of designing a part that will stay within dimensional tolerances is that shrinkage and warpage have separate causes. Shrinkage is mainly material related and ranges from 2% to as high as 14%. Warpage, on the other hand, is largely cooling related. However, shrinkage can be affected by cooling, and some warpage is due to the material itself.

The mold shops that make MPC’s tools for engine covers have worked closely with MPC for many years. When moldmakers design a tool, they start with the particular resin’s material safety data sheet (MSDS) that provides values for anticipated shrinkage. Toolmakers then add their own knowledge of warpage. Cooling is controlled by the sizes and placement of gates that release hot gases generated in the molding cycle. “Until a tool is built and run in a molding machine,” says Clark, “shrinkage and warpage are only educated guesses.”

Engine covers are often emblazoned with the automaker’s logo. These products are what automakers call appearance parts; when the vehicle hood is lifted the first thing seen is the engine cover.

The typical engine cover measures about 14 by 22 inches and its exterior or side is often domed, embossed, or fluted. Most have tolerances of ±5 mm. Initial quality checks, however, showed that shrinkage and warpage in the Lincoln cover were ±6 mm and even ±8 mm (approximately .240-.320 inch).

While available methods such as touch-probing with a coordinate measuring machine (CMM), moldfilling analysis, and photogrammetry are beneficial in dealing with shrinkage and warpage, they fell short in some areas.

Touch-probing is ideal for features like locator pins, holes and slots, and for customer buy-offs. But it is not as adept at measuring the smoothness of a curve or flatness and will miss any bump or sag that lies entirely between probed points. Touch-probing will also miss more subtle things,” Clark says, “like a shift in the pitch of a pair of X and Y points that twists another point out of its Z tolerance.

The other aspect of the problem is lack of time. “We have only a certain window of time to tweak and stabilize the molding process, and we have to cover a lot more than shrinkage and warpage,” he says. The CMM takes seven working days to set up a job, two more days to measure, and five days after that to crunch all the data,” Clark notes. And, because the CMM is always busy, we may need to wait another seven days to get a place in the job queue. With point-by-point methods, you have only two opportunities to get it right.

“With ShapeGrabber”, he says, “if we need new data, say, to determine exactly where an attachment point is in 3D space, we just click on the point in the data. Within seconds, it’s uploaded to the computer and we can query PolyWorks. We can go back and do this anytime,” he adds, “even three years later.”

Data in 15 minutes

MPC chooses to deal with shrinkage and warpage with laser scanning and point-cloud-based inspection, which also enables it to start full production in a reasonable amount of time. Because we can gather all the dimensional and tolerance data we need in 15 minutes, Clark says, €œwe have more time to engineer and analyze each job. That extra time is a huge help as we decide whether to seek looser tolerances, modify the tool, change the process, or have a cooling fixture made. With this type of inspection, we now have several additional days to make these decisions.

Most of MPC’s inspections are done with a full-color map of tolerances in the PolyWorks/Inspector suite that contains the IMInspect and IMAlign modules. The display uses the entire color spectrum from red (plus tolerances, too much material) to blue (minus tolerances, too little material). Green areas of the surface are ±0. Any surface region in gray is outside the preset tolerance band. The tolerance readout is a one-bar graph down the right side of the display, and its colors are keyed to the part display.

Because the color map’s tolerance band is adjustable in .5-mm increments, it enables what-if analyses of mold modifications. “The color maps let us see what is really going on in the mold, which surfaces are affected by changing a tolerance, and by how much,” Clark says.

“What the color map really shows us,” he adds, is how the plastic part floats inside the mold. Predicting the effects of shrinkage and warpage with these tools offers significant help in getting the process nailed down. We know the material will shrink into its proper dimensions and the warp during cooling will bring it into the shape the designer called for.

For MPC, ShapeGrabber and PolyWorks were the only systems that could:

Keep pace with production rates and inspection frequencies. €œNone of the other laser-based approaches could gather the necessary data in less than 2 or 3 hours,€ Clark says. That was too slow.

Pass standard repeatability and reproducibility (R&R) tests that certify inspection systems. In an R&R test, an operator measures a feature on 10 parts three times. A second operator repeats the inspection and the results are compared. “The first Shape Grabber and PolyWorks jobs were done in about 2 hours,” Clark notes, compared with three weeks using CMM and CAD methods in the past. By itself, that enormous gain in speed was enough to give MPC a two-week ROI in the ShapeGrabber systems and the PolyWorks software.

Time benefits

MPC says that laser scanning and point-cloud inspection outperform touch-probe inspection because the technology allows them to represent each surface with half a million points or more instead of just a few hundred. They also believe that scanning and point-cloud inspection outperform moldfilling analysis because they look at what actually happened in the mold, not what was supposed to happen.

The time gained by laser scanning and point-cloud inspection also means that Clark’s team and others at MPC can go back for a second look. This might include suspect areas in a second batch of sample parts. At this stage, touch-probing would probably generate sufficient data. “But that requires an entirely new CMM run,” Clark says.

“If we need to check something not originally specified in the inspection layout, we just add points by revisiting that part of the scan. The CMM job has to be programmed with the additional points. To add even a single point, you have to start all over again with the CMM, which never takes less than two days.”

Return on investment

The real test of the effectiveness of this technology is whether or not its benefits extend beyond the needs of users in their everyday tasks. In the larger enterprise, those benefits are threefold:

Speedier production starts stem from quick capability studies. These ensure parts will be molded as required before samples are submitted to customers. Users are able to eliminate numerous additional dimensional checks, which allows tools and processes to be locked down sooner.

More engineering and analysis time for production options such as loosening tolerances, modifying the tool, or changing the process and cooling fixtures. In turn, that allows more time for mold design and tryout, which translates into happier customers.

Increased flexibility of methods, which allows more production options to be evaluated. In the past, much of the production lead time was consumed in gathering data about the as-molded surfaces.

The ultimate gauge of a technology is, of course, whether or not it can boost competitiveness and transform the business. ShapeGrabber and PolyWorks score here, too. Instead of spending days to inspect surfaces with touch-probe methods, MPC gathers 500,000 to a million-plus points on a surface in just a few minutes. Confidence is high at MPC that the inspection improvement will lead to additional business opportunities.