Close-up on 'lost-core': a puzzle with many pieces.

Close-Up on |Lost-Core'

The first two large-part commercial applications for so-called "lost-core" injection molding in the U.S. start up this month: automotive engine manifolds at Dunlop Automotive Composites Inc., a Dunlop/Ford venture in Benton Harbor, Mich.; and integrated fuel-rail/manifolds at Handy & Harman Automotive in Dover, Ohio. Meanwhile, Tomco Plastic Inc. in Bryan, Ohio, which became the first commercial U.S. lost-core molder of smaller parts a year ago, is looking at several new automotive applications, including brake housings for a German auto maker (see PT, Sept. '91, p. 49 for a review of automotive "lost-core" developments). Dozens more molders are developing parts for cars, airplanes, plumbing, bicycle wheels, and even footwear using lost-core technology. These processors expect to enjoy handsome advantages molding complex hollow parts that would not be moldable in one piece any other way.

But the stakes are high for newcomers to lost-core molding, a technology that involves inserting a core (generally of cast metal) into an injection press, overmolding it with a thermoset or engineering thermoplastic material, and then melting or washing the core out of the part.

Major problems with lost-core parts are the long development time and high cost of coordinating the many facets of production. An automotive lost-core part can take four years - two years to reach the prototype stage, two more to production. Nonautomotive parts can be quicker: only 18 months were needed to develop a new two-way body valve at Johnson Controls Inc. in Erie, Pa.

Those that have tried it warn that costs for entering this technology can be huge. It can take an investment of up to $5-8 million to get involved in lost-metal-core (also called fusible-core) molding technology to produce a four-cylinder engine manifold, says consultant Siegfried Niedermair of NPT Inc. in Newmarket, Ontario.

Not everyone agrees that the costs need be so daunting. A molder with existing injection equipment can make 250,000 manifolds/yr with core-casting and melt-out equipment costing roughly $500,000 (not including robotics, tooling, and about $50,000 for core metals and wash-out solution), says Mark Batista, v.p. and general manager of Electrovert MDD, a unit of U.K.-based Cookson PLC, which supplies core metals and core-casting and melt-out equipment.

It appears that differences in cost estimation depend partly on overall approach, to a great extent on complexity of tooling, and partly on intangible factors. "The cost situation is uncertain because not every design is feasible," observes Dr. Willy Hoven-Nievelstein, director of applications development and technical service at BASF Corp.'s recently opened plastics applications center in Wyandotte, Mich. He is referring to the fact that some lost-core projects proceed relatively easily through the prototype phase, but prove unsuccessful at molding in production - sometimes after considerable expense in time and money.

That question of predictability may be yielding to technology. Use of computerized flow-analysis software (in this case, from Advanced CAE Technology Inc.) reportedly has helped one Taiwanese molder develop a successful lost-core application.

A number of important technology issues remain unresolved. Only a handful of companies claim success in lost-core molding: Montaplast GmbH, Mann & Hummel, Freudenberg, and Klifa in Germany; Dunlop in the U.K.; Meccaplast in Monaco; Solvay & Cie's STMP unit and MGI-Coutier in France; and Tomco in the U.S. Freudenberg, which purchased Woodland Molded Plastics Corp. in Broadview, Ill., and Dunlop both entered the U.S. market this year; and Montaplast and Solvay are considering such a move. In the Far East, Apollo Industrial Co. Ltd. in Korea and Yue Loon Co., a Taiwanese auto maker, are also near production. These early leaders are both secretive about their success and concerned that new processors rushing into the technology will stumble at any of several points and make bad parts. This, they say, could jeopardize the auto industry's initial acceptance of plastics in critical underhood applications.

"Right now everyone is hyper to get into this technology, but there will be some surprises," predicts BASF's Hoven-Nievelstein. "Some companies will find it's a very difficult technology." Only one company, Klockner Ferromatik Desma GmbH of Germany (represented by KFD Sales & Service, Inc.), builds complete turnkey lost-core manufacturing systems. And it trains customers for six months in Germany, then sends an engineer to work with the customer for a year to familiarize operators with running a system that has many more variables than injection molding alone.

On the following pages, PLASTICS TECHNOLOGY presents the first in-depth look at the lost-core process and outlines some of the pitfalls of this demanding technology and the options available to get around them. This includes discussion of the capital costs for different techniques and other investment risks; how costs and cycle times differ with varying press configurations; the tradeoffs of using different resins and core materials; and some of the unsolved problems of melting out cores. Robots and process controls present still more questions that space does not permit full discussion of here.

ALTERNATIVE PRESS DESIGNS

Five years ago, the first lost-core parts were made on standard horizontal clamping presses, but loading and unloading 100- to 200-lb cores with these machines was cumbersome and slow. One solution has been to switch to vertical clamping with either rotary or shuttle action to facilitate core loading and unloading. Though a vertical-clamp rotary press costs 35% more than a stationary one, takes up more space, and requires two bottom molds (adding 40% to tool costs), it needs only one $200,000-240,000 robot instead of two to load and unload and has over 30% shorter loading cycle, according to Rudi Loehl, marketing manager for Battenfeld of America. However, small lost-core parts in the 1-5 lb range are mostly still molded horizontally.

Vertical 600- to 800-ton presses are needed to mold a six- to eight-cylinder manifold; 500 tons for a four-cylinder, says Loehl. A Battenfeld four-tiebar rotary press cycles in 7 sec, making a six-cylinder manifold with 120-lb core, says Loehl. Rotation time for a Bucher rotary thermoset injection press reportedly is about 7 sec. Klockner, meanwhile, offers a vertical press with rotary table and three tiebars that cycles in 4 sec when making a four- or six-cylinder manifold, says NPT's Niedermair. (The latter press was originally custombuilt to specifications from Montaplast, a German automotive molder known for technical innovations. German lost-core molder Mann & Hummel also uses Klockner equipment.)

A vertical press with shuttling bottom mold costs 15% more than a fixed vertical press, takes two robots to load and unload, and has a 21-sec mold-open cycle, Battenfeld's Loehl says. But it saves the cost of a second bottom mold. Thermoset molders Dunlop and Freudenberg in Europe both use vertical shuttle presses from Bucher and Battenfeld for lost-core molding.

Fixed-clamp presses, whether vertical or horizontal, don't need to be as large, heavy or expensive as shuttle or rotary models, Battenfeld's Loehl says. But fixed-clamp presses may take two robots and a 28-sec loading cycle (depending on core geometry) he notes. Horizontal presses will also need more mold sliding action to lock cores in molds.

Handy & Harman chose a fixed-clamp horizontal press - a Nissei FH 860 (960-ton) - for molding its manifolds because of a new clamp design that takes half the room of comparable 900-ton presses. Handy & Harman's first lost-core molding cell will use three robots from ABB Robotics Inc. to assemble, load and unload cores, plus two Electrovert LMD 2000 core-casting machines and two induction melt-out systems purchased through Electrovert and based on induction coils from Inductoheat.

COMPLEXITIES OF MOLD & CORE DESIGN

For starters, lost-core molding is so complex that standard mold filling and cooling analysis software is often of little help, processors say. Montaplast managing director Albert Stulz says, "We make all our tooling in-house, using flow and cooling analysis software, but sometimes solutions, like the way the sprue is gated, are actually contrary to what the software tells you."

Lost-core deisgn is inherently difficult, with two mold sets involved (for core and part) in making very tight-tolerance parts. "In effect, you're using a new mold at every shot," says Phil Leopold, president of Bucher Inc., since the core or male part of the mold must be replaced on each cycle. "The core can move and any variation from one shot to the next has a serious effect on the part." This is not made simpler by the fact that you're molding with resin nearly twice as hot as the melting temperature of the core metal. Nylon 66 melts at about 500 F, typical tin-bismuth alloy at 281 F - though this is variable by tailoring the alloy ratio. An alternative to metal is a soluble acrylate polymer core material from Belland Inc. (U.S. arm of Belland AG in Switzerland), which typically melts at around 275 F.

Gating is problematic because filling with too much pressure collapses hollow polymer cores or distorts metal ones. Metal cores are only 10% as rigid as mold steel, according to Solvay, which has had lost-core production at STMP since 1989. Even if the core doesn't distort, just creating surface waves on it will spoil the airflow within a finished manifold. To solve this deformation problem, Electrovert's Batista says resin should enter "parallel to the plenum and flow around the core."

With soluble polymer cores, gating should be spread out and flanges used to break resin flow to the mold, says Belland president Lloyd Willey. "The outermost skin of a soluble core softens for maybe 60 sec during molding, but it doesn't hurt the core's dimensional stability," he adds. Soluble cores are envisioned for some aerospace parts in development, which will require overmolding with liquid-crystal polymers at 600 F melt temperatures.

Tool design also must solve handling issues like how to section a core. Handy & Harman's new manifold uses a 200-lb metal core set in five pieces, all assembled and loaded robotically. "One nice thing about tin-bismuth is that you can force-fit pieces together. You mold a plug and hole and assemble them. It works very well," notes Tomco v.p. of sales and marketing Peter Bowers.

But what really makes success or failure in lost-core tool design is whether the tool makes a part with absolute repeatability in an acceptable cycle time. To ensure repeatability may require conventional slide-action cores to lock cores in place, mold complex shapes on the outside of a part, and reduce fusible core mass (the more you use conventional sliding cores, the less fusible core is needed). You may even need "cores within cores" - core pulls inside the fusible core - in order to keep long cores from bending. As an example, Handy & Harman's integrated fuel-rail/manifold tooling has seven sliding actions, including some double movements, says technical center director Robert Sinacola. Such tooling can cost over a million dollars, though not all lost-core tooling is so elaborate. Tooling for a two-way valve molded with fusible cores by Johnson Controls reportedly costs less than $20,000.

Core design must allow for precise robotic assembly and loading. The first choice for locking cores into the mold is to utilize all available manifold holes, shutting off around 1.5 in. at the ends of runners. When this isn't possible, designers create holes in the part that have to be covered later by spin welding a plug, and use conical, self-centering pneumatic slides from the mold to support the core.

Added to all the above complexities, lost-core-molded part design goals constantly get tougher. Newer lost-core parts include long, thin, hollow shapes and integrate more functions. In Germany, an air-channel system behind a developmental instrument panel integrates pumps and thermostat housings, says Klockner export sales manager Gerhard Neidinger. New resins are being tried, like PPS and rigid thermoplastic urethane (Dow's Isoplast) for manifolds and fuel rails, or soft elastomeric TPUs for athletic-shoe "pump" bladders.

Help in overcoming these difficulties can be obtained from materials suppliers. BASF, as a maker of nylon 66, offers assistance in lost-core part and tool design, as well as prototyping and trials at its Ludwigshaven, Germany, lab. BASF currently is involved in developing "maybe 10 different manifolds" for BMW, GM/Opel, Fiat, Volvo, Porsche, and Detroit's Big Three, says Hoven-Nievelstein. Other makers of nylon 66, like Du Pont Co., Bayer AG in Germany (parent of Mobay Corp.), and ICI in the U.K. (parent of ICI Americas), also offer lost-core design and development support.

MOLDING THE CORES

Tomco looked at core-material alternatives over a nine-year R&D period, including a developmental sand/salt program of Du Pont's and developmental starch-based polymers from a University of Toledo, Ohio, program, and concluded that metal cores maintain part tolerances better. A disadvantage is that a metal core inside a plastic part acts like a heat sink. "You have a highly thermally conductive material inside a part which isn't at all conductive, making it very hard to control the dissipation of heat," says Tomco's Bowers. "So allowing the injected polymer to cool without inducing stress is difficult."

Metal core making is also unfamiliar to plastics processors. Bismuth-tin mixture (typically 58% bismuth, 42% tin) is the most commonly used eutectic alloy because it expands slightly (typically 0.003 in. per dimension) as it cools, packing the core mold without external pressure. Electrovert offers bismuth-tin, tin-lead-silver and tin-lead-antimony alloys with 98-800 F melt temperatures. MCP Systems, a unit of Mining & Chemical Products SA of Switzerland, also offers a complete range of alloys. Bismuth-tin alloy melting at 281 F is commonly used for molding nylon 66. Tin-lead-antimony melting at 396 F is used for thermosets. Resins with other melt temperatures may require custom alloys. The core alloy developed for the athletic-shoe prototype, for example, melts out in boiling water at 220 F.

The two main sources of core-casting equipment are Electrovert MDD and Klockner. Electrovert offers computerized core-making and melt-out equipment, as well as robotics and its metal-alloy core materials. Klockner offers core-casting systems (using metal injection nozzles and metal pumps from Bachmann GmbH in Germany), melt-out systems, and robotics, as well as its own molding machines; and it has an association with MCP for core alloys.

A key difficulty to avoid in casting metal cores is trapping air bubbles that can cause part flaws later. To avoid bubbles, Solvay's STMP division built its own "two-pressure" casting system. It fills core molds from the bottom at a slant, applying pressure at the same time to the surface of the molten metal. Klockner and Electrovert say the bubble problem can be corrected by varying temperature, speed, and the part of the core mold that is filled first. Electrovert's Battista also says cores from its system are weighed to "within 2-3 grams on a 100-lb core," in order to check for voids.

Klockner's core-casting machines rock slowly on an x-y axis during casting at 281 F. Cores come out of its system still fragile, with an approx. 1/8-in. skin solidified at about 220 F, and are put in special steel fixtures that are custombuilt to provide support and conduct heat out. Cores then travel into a cold-air cooling tunnel and come out at 70-170 F.

Klockner's system is designed to cast cores "just in time" for part molding. Company sources note that there is a tradeoff of part quality for cycle time: When making glass-reinforced nylon parts, hot cores make a smoother inside surface, and gradual cooling lessens part stress. But a cold core saves 8-10 sec of cycle time, says NPT's Niedermair. Electronic controls keep metal cores moving continuously between core casting and molding, so cores enter the press at a constant temperature.

Tomco built its own core-making equipment because "what was commercially available at the time didn't meet our needs. It wasn't fast enough or precise enough," says Tomco's Bowers. One press making a large part with a multi-section core may need two core-making and melt-out systems to provide a continuous cycle, while a small part like a BMW impeller water pump uses one core-making system and a 15-ft glycol melt-out tank for three presses.

The main alternative to metal cores is Belland's soluble acrylic-polymer core material filled with 25% talc. It has been used to prototype parts with relatively simple internal surface detail. Soluble cores are generally molded in two 1/8-to 1/4-in.-thick, hollow halves and welded ultrasonically, rather than molded in a solid piece. Hollow cores sometimes deflect, or can even cave in, when over-molded with high-temperature (500-600 F melt) engineering plastics. However, a new foaming technology is in development that makes solid foamed cores in one piece with greater surface definition and three times faster melt-out, says Belland's Willey.

Cascade Engineering Inc., Grand Rapids, Mich., has prototyped office furniture and plumbing parts with soluble-polymer cores, but has found no commercial applications yet. Mitsubishi Plastic Products Corp. in Japan has also prototyped a four-cylinder manifold with soluble-polymer cores.

Because applications are strictly developmental, processors haven't yet had to address the question of recycling soluble core polymer. Belland says this can be done by precipitating the polymer out at an alkaline pH, followed by spray drying and repelletizing with a devolatilizing extruder. Belland director of applications development Lee Wielgolinski says recycled material degrades slightly during reprocessing, so 10% new material must be added. (Disposing of nonrecycled polymer into sewers isn't permitted in some areas for environmental reasons.) Some technicians who have molded the Belland material also say it gives off a sweet smell that gives them headaches.

However, soluble polymer cores offer one clear advantage: a molder can prototype a soluble-core part without buying new equipment, since core halves and overmolded parts can be molded on a standard injection press, and cores can be washed out in a sink, says Belland's Willey. Dow United Technologies in Wallingford, Conn., and McCourtney Plastics in Minneapolis are doing such development work with soluble cores.

Capital cost of a polymer-core production molding cell with two injection presses, one for cores, one for parts, is about $1.5 million, Willey says. This has clear advantages for small production runs where volumes could never justify the cost of a metal-core set up.

CHOICES IN CORE

MELT-OUT

"Melt-out is probably the least well developed and least understood part of the process on a commercial scale," says Steven Haycock, marketing manager of Freudenberg-NOK in Plymouth, Mich. Techniques for voiding metal cores are hot-bath (glycol solution at 300 F) or induction melting or a combination of both, depending on core size and whether a part is thermoset or thermoplastic.

A hot bath typically consists of glycol-water solution at 300 F. BASF also sells a phenol-based heat-transfer liquid for this use, called Lutron HF3, for about $2/qt. Hot-bath melting can help complete the cure of a thermoset part and avoids problems of induction heating metal inserts, but molders face a potential environmental disposal problem when parts are rinsed in water or the glycol bath needs changing. And melting a large core with a hot bath alone is uneconomically slow - 60-90 min. "This cannot be considered a viable industrial solution," claims Solvay's STMP. The longer time required means that a longer bath containing a greater volume of solution, as well as a greater volume of cores, is required in a closed-loop system.

In hot-bath melting, parts with cores may hang from an overhead conveyor or ride submerged on a stainless-steel conveyor through a melting tank. A tank for 100 manifolds might hold 25 cu yd of solution (four manifolds/cu yd is a rule of thumb). The solution bath must be cleaned or disposed of once a year - every six months if induction heating is also used. Klockner's line includes a centrifuge to clean the solution.

Getting all of the core to drain out of a part with a complex shape can be a problem, and some parts may have to rotate in the bath. A V-8 manifold, for example, may need to rotate for a whole hour to drain completely. A wedge-shaped tank bottom funnels melted alloy into heated pipes that carry it back to the core-casting machine.

Induction heating is needed to melt metal cores sufficiently quickly out of thermoplastic parts, because prolonged heat could warp the parts. Induction reportedly can cut melt-out time to 1-3 min. Induction coils wrap around each part about 1-4 in. from its surface, allowing room for grippers to tip the parts so they'll drain. But induction alone often won't remove all the metal particles, so parts usually must be submerged in glycol or brushed afterward.

The big problems with induction melt-out are the cost and design uncertainty of custom-built coils. "For each new part you have to start all over again, and some parts are much, much more complicated to ensure that all the alloy is removed than others," warns Freudenberg's Haycock.

Solvay says it has developed software to calculate the size of the induction coil and how to adapt it to the generator for maximum efficiency, and has applied for patents on proprietary melting machinery developed by its lab in Brussels. Solvay says its new system melts a 20-lb turbocharger core in 40 sec.

Handy & Harman is setting up what is believed to be the first induction melt-out system in the U.S. Its 200-lb manifold core will require 5000 lb of alloy (at $5/lb), assuming 10 cores between melt-out and casting and 10 waiting to be injection molded, plus 20% extra for safety. Cores are cast in just over 1 min, and parts molded in 1.5 min, but induction melt-out takes 3 min, necessitating two induction tanks to service a single injection press.

NPT's Niedermair warns that metal inserts in parts can present problems during melt-out. Brass or steel inserts can react with the alloy, destroying the inserts and oxidizing the alloy. To solve this problem, one nylon 66 BMW manifold is manufactured with inserts that are pressed into the thermoplastic after core melt-out.

NOVEL TECHNOLOGIES

EMERGING

New core technologies are beginning to emerge that could mitigate some of the hurdles to lost-core commercialization.

* "Stage-four" ice: In Taiwan and Japan, supercold "stage-four" ice is being molded for prototype cores using distilled water and liquid nitrogen. Nissei in Japan developed the equipment, said to be used by two custom injection molders of small parts for Nissan cars. Ice at such low temperatures attains such a tight crystalline structure that it freezes the molten resin without creating steam in the injection mold. This could obviate the need for metal handling, and all the expensive environmental issues associated with the glycol melt-out system.

* Encased ice: Cores made of ice-filled, blow molded PPS are being developed by Patent Products Corp., a technology development company associated with an injection molding firm. With this method, the ice reinforces the hollow blow molded part during injection. Then the ice is melted out and the PPS core remains inside the part. Thus the inside of the blow molded shape becomes the inside surface of the finished part, says inventor and owner Royce Husted.

The beauty of it is simplicity and low cost. Husted says his first sample cost $700 to produce, compared with over $1 million that it would have taken for either a metal or soluble-polymer lost-core prototype. And injection tool design may be lower, too: molding over ice-filled PPS also requires fewer gates than other lost-core materials because ice is so rigid - more so than metal, says Husted.

Commercial refrigeration equipment is used to freeze industrial distilled water inside the cores to below zero F. Processing requires air-conditioned, dehumidified conditions to prevent condensation on cores before molding. Ice-filled cores have a "core print," or lock-and-key pattern, molded in to hold them in position in the mold cavity.

Part design is restricted to the physical limitations of a blow molded part, but with multiple-sectioned cores, complex shapes are possible - e.g., a two-cylinder, lawn-mower manifold with several "legs," to be molded of 55% glass-filled TP polyester, is in development.

* Starch-based polymers: Vegetable-starch-based polymer from Novon Products is reportedly being tried for cores in several small developmental parts for computers and sewing machines. The cores could be either melted out at low temperature (275-300 F) or dissolved out in water. Advantages would be an entirely polymeric system, eliminating metal casting, and a core material that could be disposed of without environmental problems.

PHOTO : Loranger Manufacturing's prototype intake manifold with integrated fuel rail uses Electrovert bismuth-tin cores. Lost-core manifolds are cheaper, lighter and more fuel-effecient than previous aluminum ones.

PHOTO : A vertical rotary press like Klockner's can load and unload cores three times faster than a fixed press and saves one robot. Klockner's core-casting machine is another integral part of its turnkey system. The core-casting machine tilts to fill molds without voids or bubbles which can cause part flaws.

PHOTO : Getting precise, repeatable cores is critical for lost-core molding. Pictured is the sharply detailed two-piece alloy core set for the first production U.S. lost-core part, a thermostat housing by Tomco Plastics. (Core and cavity tooling by C&A Tool Engineering Inc., Churubusco, Ind.)

PHOTO : Lost-core molding is like installing a new mold with every shot. As robots lock 100-200 lb cores into place in the mold cavity, any slight change of position will spoil the part.

PHOTO : A problem with large manifold cores is melting them away at commercially viable rates. One answer is induction heating, which can cut melt-out time to 1-3 min from 60-90 min. Below, a balsa model of Handy & Harman's first lost-core cell shows the first induction melt-out system in the U.S. It includes two Electrovert core-casting units molding five core parts, and two induction melt-out units for one press.

PHOTO : Voiding metal cores out of thermoset parts with hot glycol alone saves expensive, customized induction coils, and won't overheat metal inserts. But a glycol bath is slow and may pose environmental disposal problems.