Here's the first in an occasional Better Buying series from PURCHASING magazine. For expert advice on buying plastics parts, Purchasing went to 40-year industry veteran and plastics engineer Robert A. Beard. Beard's top advice—

"Let's discuss several scenarios," Beard says. "If an OEM has ten molders using the same resin poundage, then each molder will probably be purchasing resin from a distributor at a higher cost than if the OEM negotiated a contract with either the distributor or resin manufacturer for the yearly quantity, which then would allow each molder to purchase the resin under that contract.

"Note that the molder, not the OEM, is purchasing the resin, which is an important distinction," Beard emphasizes. "We live in uncertain times. If there was an interruption in our oil supplies similar to the oil embargo of 1974-1975, there would be an interruption to the resin supply." During the oil embargo of 1974-1975, resin allocations were given to the companies that actually bought and paid for the resin. An OEM that wanted to move a job from one molder to another was not able to do so, because the resin allocation stayed with the molder [who paid for the resin] and not with the mold.

"It would be extremely prudent for an OEM negotiating a resin contract to have both the resin company and the distributor present at the same table at the same time," continues Beard. The reason for having both present is that the OEM needs to know how the resin manufacturer would allocate resin to the distributor and how the distributor would allocate to its customers. Clear rules for resin allocation should be negotiated into the resin contract.

If an OEM is single-sourced with one molder, there may not be a financial advantage to negotiating a resin contract or buying the resin. There still would be a business consideration to negotiate a resin contract for spelling out the rules for resin allocation.

"However, if an OEM is concerned about being off the market for lack of resin, then it might want to purchase resin and store it as a hedge inventory," suggests Beard. He once had a situation where a key plastic component was used in a diagnostic test that represented $50 million per year in sales. The resin supplier wanted to discontinue the particular resin. "We discouraged the resin supplier from dropping the resin, then we purchased a year's supply and put it in a warehouse. We determined that a year would be sufficient time to develop, test and approve a substitute resin."

Most specifications center around part dimensions. However, the part is only going to be as accurate as the mold that produced it. Beard recommends having a mold measured by an outside company using an XYZ coordinate measuring machine before the mold is sampled. This does several things. It puts whoever is going to build the mold on notice that their work is going to be measured before sampling the mold. "This stops the argument between the mold builder and the molder as to who is responsible when the part does not meet specification," he explains. It costs more money to do this and adds several days to the delivery time. However, it avoids needless weeks of trouble-shooting and delays to the market.

All molders will have the equipment to manually measure molded parts. Others will have made the investment in a shadowgraph or XYZ coordinate measuring machine to more easily measure parts and molds.

If the application is one that demands a high degree of repeatability, such that the reject rate is measured in parts per million, then the molder should be using statistical process control (SPC) to monitor the process, according to Beard.

If the application requires that the part perform under a required load, then the OEM should develop a functional test that the molder can perform on the part in his plant immediately after the part is molded and also 24 hours later after the part has equilibrated and has stabilized. "The goal here is to give the molder the same yardstick that will be applied at incoming inspection," explains Beard.

"My personal benchmark is that a molder is state-of-the-art if his molding machines are five years old or younger," Beard reports. Typically this type of molder will trade in 20% of its machines every year to accomplish this. He considers a commercial molder to be one if his molding machines are ten years old or younger. "If the molding machines are older than ten years, it is indicative that the company is not staying up with technology, may have higher rejects and costs, and may not be able to mold tight tolerance jobs," he continues.

Most molding machines built in the past ten years have closed feedback control, meaning they sense during an actual shot that something is going wrong and correct it in the same shot, avoiding a reject. "The machines are improved every year, which allows the holding of ever tightening tolerances," adds Beard.

If the part is going to take an impact load, then impact strength is going to be important. If the application is an optical one, then light transmittance at various frequencies will be important. If certain resin properties are important, then the molder needs to have a resin certification from the resin supplier on every lot, certifying the min/max values of the important property. "There is no sense in specifying anything unless you are going to randomly test to see that you are getting what you paid for," Beard points out.

In the healthcare industry, for example, it is common to have a statistical sample of every lot of plastic parts coming from a supplier tested with an infrared spectrometer to make sure the parts are made from the correct resin. "Recalls are very expensive in both the healthcare and automotive industries," adds Beard.

He has run into a number of instances over his career where molders shipped perfectly shaped parts molded out of the wrong material. "They took it upon themselves to change the material without telling the customer," he explains.

Almost all new parts require a new mold, according to Beard. Some OEMs believe that they should buy the mold and provide it to the molder, thus avoiding the molder's markup on the mold. This can be done, but only if you have experienced mold makers on staff. "The major disadvantage to this approach is that it sets up the proverbial 'finger pointing' between mold maker and molder when something goes wrong," he cautions.

The better approach is to contract with the molder to build the mold, even if he has to subcontract the actual building to a mold-making company. In this approach, there is single source responsibility. The OEM can still exert considerable control in the specification of the mold without relieving the molder of his responsibility. Beard has seen very few OEMs that have technical staffs on board to specify their own tools.

"In my opinion, most OEMs do not give molders the tools they need to prevent downtime," replies Beard. He recommends that the OEM purchase 25% spare cores and cavities at the time of building the mold, if it is a high volume job. By having spare components, the molder can quickly change out a damaged core and cavity and still maintain shipping schedules. In the meantime, the damaged component can be either repaired or replaced without affecting the production schedule.

A typical mold change is two to four hours, according to Beard. It depends on the size of the press. Some take a shorter time, some longer. Depending on the mold size, the press size, and the type of quick-mold change system, a mold can sometimes be changed in about five minutes.

If an application is going to be molded in small lot sizes (less than 1,000 parts), this will increase the number of mold setups. When this is done, it increases the reject rate and the potential for damage to the mold. "The rejects are usually produced in the first four hours of starting up a mold," explains Beard. "Most of the damage done to a mold is done during mold setups." As such, if you order in small lot sizes and only run the mold for two to four hours, the process will never get to a steady-state situation.

"The more savvy molders have inventory systems in place that their customers can access 24/7, either over the Internet or on a direct computer call-in," says Beard. This makes the supplier's inventory very transparent to the customer. However, there is a potential danger to this feature. "If the customer micromanages the inventory, he may be increasing the number of mold setups," he cautions. "Most mold damage is done during setups." And again, most rejects are made in the first two to four hours of the mold startup. "The OEM may be improving inventory turns, but at an increased manufacturing and maintenance cost," says Beard.

"When I was the general manager of a custom molding division, I always found it was more cost-effective to perform secondary operations in-line with the molding process rather than decoupling the secondary operations and performing them off-line," says Beard. The advantage is that if a reject occurs that will affect secondary operations, it is discovered immediately, rather than making hours, or days, worth of rejected parts and having them end up in in-process inventory.

This is where plant layout is important. "If the molder has his presses located so close together that he cannot accommodate an in-line production line, he may be forced to decouple the production line," he explains.

This requires different approaches depending on if companies are making 5,000 parts per year or five million per year. It requires coming up with a plan of options, having the options quoted in parallel so as not to lose time, and then directing the supplier as to which option is most beneficial. "What I strive for is financial justification for the course of action that is chosen," Beard emphasizes.

The advantage of not having your own captive molding is that you can shop for molders or suppliers that are experts in the technology that your particular part requires. "If you do have your own captive operations, you are limited by the knowledge that your internal people have, which can be a serious limitation," he cautions.

As such, when it comes to multi-process capability, you should be going to the molder that has the best experience and best technology in molding that particular part. "Your engineers may be experienced parts designers," he points out. "However, they are not mold designers or plastic process engineers. They also usually have very limited knowledge of resins and their properties."

Unfortunately, according to Beard, few molders will want to offer design services. The reason: If the molder is responsible for some of the product design and warranty service, it will expose itself to a lot of product liability. "I don't think most molders want to take on this risk, especially since they are not being compensated for it," he explains.

Robert A. Beard, P.E., is president of Robert A. Beard & Associates Inc. (Kenosha, Wis.) (www.plastic-solvers.com), one of the nation's leading experts in relationships between plastics suppliers and customers. Beard has been involved in all aspects of the plastics industry since 1964. In his role as a consultant, Beard claims he can typically identify savings of 20% on total annual plastics parts purchases by extensively reviewing each part, mold and process. He teaches a seminar called Purchasing & Quoting of Plastic Parts.