One molder's private-eye work yields fine results

Black specks were bumping off the good parts yield of molded PC cuvettes. What was causing the contamination? Could the yield be improved in time to satisfy the growing global demand for the parts? A savvy troubleshooter performed the customary third degree on all the likely suspects, including the mold, the machine, and the auxiliaries. Marginal improvements were made. But the breakthrough came when he saw an ad in the trades for a micro-sized fines separator (see figure below). The separator led him to the real culprit—the material. 

"I did not have a fines separator in mind when I was presented with the black speck problem. That happened on my second day on the job," says Cornel Onu, plastics process engineer at Ciba Vision Corp., a Novartis Co. (Des Plaines, IL). He began his career as a plastics engineer in 1968 in his native Romania. Now a Canadian citizen, he works here as a free-trade professional under NAFTA. Ciba Vision used to be known as Wesley-Jessen Corp., the first mass marketer of contact lenses. 

The black specks in the molded cuvettes were throwing off an artificial vision system that controls a fully automated contact lens packaging line in the Des Plaines plant. The cuvettes hold the lenses in solution as they are conveyed through various conditioning and washing stages prior to packaging. 

Before Onu's arrival, cuvette scrap was a sufferable problem. But when the company opened a lens plant in the U.K. to serve the European market, volume demands dramatically increased. Good part yield had to be improved. Sufferable had become intolerable. 

Good Is Not Good Enough 
"The machine in-house was down when I started, and the mold had been sent to a regional custom molder whose plant was only partially a cleanroom, which is not a good molding environment for such parts," says Onu. "I told my boss to bring the mold back so I could analyze it and plan." Onu visited the shop, pulled the mold, worked all night, and got 2000 good cuvettes out of it. "People were excited, but I had just started my investigation." 

Onu's plan was to investigate all the other probable causes for the black speck problem. He faced several challenges. "There was no PM program in place for the machinery and the average age of the molding machines was nine years. There were platen parallelism problems, and the cuvette press had a GP screw—a GP screw for molding an FDA-approved, 15 MFR polycarbonate resin. Imagine that." 

During the first months, he was working with a Milacron Act 75-D all-electric. It was a prototype of a since-discontinued line and there were no specks. After checking over the press himself he sent the injection barrel, end cap, and check valve out for work. He also contracted for a new PC screw. 

"You have to learn how to handle polycarbonate," Onu explains. "It can damage screws. You can't squeeze it. You have to find exactly where the material wants to melt, so you have to have adequate space in the transition zone. Good-parts yield went up to 35 to 40 percent on the best days just with the screw change. They said that was good enough, but I knew these were still the first steps." 

Pursuing Perfection 
Next came the dryer. "We were using a 15-year-old dryer. Dewpoint readings were off by 10 to 15 deg F. I put in a request for a new one." Onu then changed the machine nozzle, putting a longer one on to reduce shearing. There were higher stress levels in the shorter nozzle because the exit melt temperature of the material was lower. "With the new dryer and nozzle, and a little more tweaking, we got up to 50 to 55 percent good-parts yield, but we still needed high packing pressures and speed even with these modifications. We still had black specks and parts were sticking in the mold. There was no mold venting." 

He suspected a runner balance problem and performed a moldfilling analysis. After analysis the main runner, subrunner, and sprue diameter were increased overnight at a local moldmaker, as was the nozzle tip diameter. Mold vents were added, which also helped to open the process window. Injection pressures were reduced by 30 percent. Parts stuck only during startup. Good-part yield was now up to 55 to 65 percent—sometimes 80 percent—but Onu was not satisfied. There were still black specks. 

Around that time Onu ran across an ad in IMM for fine separation systems from a company called Technology & Trade Inc. (T&T, Chicago, IL). Discussions began. F.E. Gerlitz, director of engineering for T&T, reconfigured a standard system, equipping it with a stainless steel dosing unit and a static eliminator to meet the rigorous demands of the project. Consistent good parts yield shot up to 85 percent. Case closed? Not for a seasoned plastics process engineer like Onu. 

"You cannot achieve perfection, but you can pursue it," he says. "I had looked everywhere, but I overlooked one thing—the material." After spotting black specks in more than one lot of his FDA-approved PC pellets he started looking around for another FDA-approved material. "But then I thought, this is an optical part first, then medical. We needed an optical-quality, fast-cycle, FDA-approved grade." He found one in Bayer's Makrolon 2458-1112. Cuvette yield is now an Onu-approved 90 percent. 

He estimates that the capital investment in the entire project was around $35,000 to $40,000. T&T's separator paid for itself in about three months.