Injection in the 90's: how high tech?

Injection in the 90's

How High Tech?

There's plenty of debate on how far and how fast automation will penetrate--but none at all on the crucial role to be played by ever more sophisticated computer controls.

There are as many different opinions on how automated injection molding will become during the 1990s as there are experts willing to prognosticate. But numerous interviews with suppliers of injection machinery, controls, hydraulics, and screws have uncovered some trends that they're willing to agree on. For one thing, indications are that there won't be a lot of "lights-out" injection molding in the next decade, and that highly automated facilities will be found primarily among captive processors and large custom processors with dedicated production.

It also looks as if information flow--rather than parts flow--will be the target for more automation. Experts cite controls as the area where greatest changes will occur. Artificial intelligence (AI) is apparently getting ready to invade your control systems, and by the end of the decade, some experts predict that novel computer technologies such as neural networks will have an impact on the injection machine.

This report looks at what the experts see ahead in controls, and also in trends in automation, robots, quick mold change, coinjection, screws, drives, and large and small machines.

AUTOMATION NOT FOR EVERYBODY

Even if the lights are left on at most molding plants in this decade, sources at Husky Injection Molding Systems, Bolton, Ontario, say without hesitation that they expect an accelerated trend toward automation in the '90s, and most suppliers echo these sentiments. Says Michael Urquhart, Husky's v.p. of machine operations, "Inventory reduction will drive automation. People want to move parts directly from the machine to the truck. You're starting to see parts-removal robots integrated downstream with packaging equipment." he says.

Although Husky specializes in automated systems--70 to 75% of its machines are sold equipped with robots--the company believes it will take some time for automation to become widespread. Says Robert Schad, Husky's president, "I expect automation to take longer to reach custom shops, though we are seeing more and more dedicated custom shops, specializing in specific fields or products. In that type of plant, automation becomes much easier."

Sources at Ube Industries  America, Inc., Ann Arbor, Mich., who attribute the company's success in large machines for automotive to its automation capabilities, also expect to see lots of automation in the '90s. Ube sales manager Dan O'Keefe agrees with others interviewed that high levels of automation will be confined to captive and large custom molders, adding that "the price of automation will go down as software and hardware are proven and accepted. Then you'll see it trickle down to smaller companies. But that's a way off."

A similar view is held by Wolfgang Meyer, president of Battenfeld of America, W. Warwick, R.I.: "The degree of automation depends on many things. A problem is that each application dictates a different solution, and that gets expensive. A reduction in the cost of automation would result in better utilization."

Meyer is a proponent of automation for application-specific shops "where there can be a dedicated-cell approach," he says. Other proponents of this approach include Robert Hare, v.p. at Klockner Ferromatik Desma, Inc., Florence, Ky., who expects to see automation geared toward manufacture of particular products, as does Joseph Sweeney, v.p. at Boy Machines, Inc., Exton, Pa. "JIT and `lights out' will be prevalent in some companies," says Sweeney, "but most custom and even captive molders can't afford that level of automation. I think the highly automated companies will be large companies with one specific product, like CD's."

Dean Reber, director of product development for Cincinnati Milacron's U.S. Plastics Machinery Div., Batavia, Ohio, says "lights out" may never happen: "You can get 90% of the benefits of automation by spending 20% of the money on selective applications rather than total factory automation. I expect you'll see much more selective use of automation. Things like mold clamps, quick disconnects, and robots are all viable."

INFORMATION MAY FLOW MORE SMOOTHLY THAN PARTS

"Automation will be the biggest factor influencing injection molding during the next decade, but much of that activity will come in automating the information flow, as opposed to parts flow," says Bruce Bowman, v.p. of sales for Polymer Machinery Corp., Berlin, Conn. Others interviewed agree, including Jerry Boggs, v.p. of sales at the Reed Div. of Package Machinery Co., Stafford Springs, Conn. "During the Eighties we seemed to get tangled up with hardware development. In the Nineties we need to become more software-intensive," he says.

"By the end of the decade, most of the information flow will be integrated from the machine level right through to the office system level," according to Husky's Robert Schad. "Machine controls more and more will speak the same language as standard personal computers, so connectability to higher-level systems will get easier and easier."

Sources at Barber-Colman Co.'s Industrial Instruments Div., Loves Park, Ill., say that in the '90s controls suppliers will need to provide more than just process control. "We must be as cognizant of the need to provide overall manufacturing information as we are of controlling the process," says Dale Stefanac, engineering supervisor. Barber-Colman is one of a handful of controls suppliers that have expanded into process and production monitoring systems

MOLDING SMARTER

"Monitoring alone will not be enough. By 1995, eight out of 10 of today's molders will be doing more than just monitoring. The other two will be out of business," says Jerry Boggs of Reed Div. "Every machine has to be capable of becoming the brains and the heart of the process. During the Nineties, controls will do all that a computer can do. Rather than just monitoring, they'll have true diagnostic and corrective capabilities," he says.

Artificial intelligence, particularly in the form of expert systems, is expected to be the answer to the next decade's control requirements. "We'll take the top mold technologists and machine testers and embed their knowledge into expert systems," says Husky's Mike Urquhart. This may become especially valuable, as some sources see it, as the ranks of experienced molding technicians dwindle and replacements are harder to come by.

"By the end of the Nineties, microprocessor speed and memory capabilities will be non-issues. How to use the capabilities will be the issue. AI is going to control the machine," says Reber of Cincinnati Milacron. "For example, you'll be able to develop a predictive model that will tell you what's going to happen before you make a change in parameters. You'll find software like Moldflow integrated with machine controls, generating the initial process parameters. That sort of thing is about five years off," he predicts.

In truth, the capability to link flow and cooling analysis software directly to injection machine controls already exists. Plastics & Computer Inc., Montclair, N.J., has worked with five injection machine builders to permit automatic generation of initial setup recipes from completed mold analyses. This approach has been used to a limited extent in production, though the transfer of setup data to the machine control panel has been done only manually so far (see PT, Jan. '87; Jan. '89; Sept. '89).

AI will be used to solve a variety of molding problems, Reber adds: "It will be applied to anything that helps molders compete better. For example, controls will optimize a scheme to reduce energy consumption per part while keeping the parameters acceptable. You'll have to look beyond 1995 for this." Reber adds that the technology for such applications is already within reach. However, taking best advantage of it will be the challenge.

A prototype of what could be the first generation of AI molding aids was exhibited at the JP'88 show in Osaka, Japan. Fanuc, which builds the all-electric ACT machines that Cincinnati Milacron sells in the U.S., is developing a computer that can offer solutions to molding problems and will implement them if the operator approves its suggestions (PT, Jan. '89).

Other machine builders, such as Bernard O'Donnell, v.p. at Van Dorn Plastics Machinery, Strongsville, Ohio, and Robert Hare of Klockner, also look for AI to have a big impact on injection controls in the '90s. Self-adaptive controls with "AI will automatically adjust parameters to achieve a desired result." Hare expects AI will enable you to know more about what's going on in the cavity. "Temperatures and pressures will be adjusted based on what's occurring in the mold," he adds. To Dennis Richmond, general manager of Krauss-Maffei's Injection Molding Div., Florence, Ky., this means "controls will not only support production runs, but will also assist operators in machine setting."

AI is also expected to play a big role in machine troubleshooting and maintenance. "Future maintenance programs will not only pinpoint existing faults but also predict when problems are likely to occur," says Tim Triplett, president of American MSI, Inc., Newbury Park, Calif., a supplier of hot-runner controls. "This predictive ability is becoming available through AI, which can monitor key inputs such as current draw on motors, temperature sensors, and others, and then anticipate when these components will require maintenance."

BY THE YEAR 2000 ...

Neural networks are on the way, according to a few sources, such as Milacron's Reber. Neural networks are an emerging computer technology that should enable control systems to "learn" a process on its own. Says Triplett of American MSI, "a neural network functions like a biological nervous system. In injection molding, it will take real-time action as the process is running. It will be the biggest thing in the next 10 years," he predicts. "A neural network develops a database through experience. When a parameter varies from the database, it decides how to correct the problem itself." The big problem in developing such a system for injection molding is the large number of variables involved, notes Triplett.

HOW MUCH CONTROL?

Battenfeld's Wolfgang Meyer sees controls going in two directions: some molders want a simpler way to set up the machine; others want super-sophisticated controls to run the whole manufacturing process, including auxiliary equipment. "Many custom molders fall into the first category. Their needs for controls are somewhat less sophisticated. Usually they're using a machine that's more flexible, less dedicated." He says the other camp consists of mostly captive molders with dedicated equipment. "They require a controller with high-end capabilities. For us it's the Unilog 9000, which we'll develop to further ease setup," he says.

Sources at Barber-Colman Co. agree on this view of two very different groups to satisfy when it comes to controls. It's counting on the MACO 8000 VA (see PT, Mar. '90) to satisfy needs at the high-end, while it is developing the MACO 4000, a controller with fewer features, for molders with less-sophisticated needs. "We need a package for the little guys," says Ralph Box, product marketing engineer for Barber-Colman. "The 4000 will have straighforward statistical capability and less communication capability than the 8000 VA." The MACO 4000 is planned for introduction in time for the next NPE in June 1991.

CONTROLS STANDARDIZATION?

Hugh W. Silberman, CEO of Automatic Technologies, Inc., a well known custom molder headquartered in South Bend, Ind., addressed a complaint to a meeting of SPI's Machinery Div. last year: "The machinery suppliers have in-undated the industry with controls during the last 16 years, to the point where the only thing known for sure is that not one thing I learn today about how to run a machine will transfer into tomorrow, due to the lack of common terminology and common functioning between machine controls."

Cincinnati Milacron's Reber expects pressure for a standard controls interface to come from big processors, like Automatic Technologies. He also expects demographics to bring about some progress toward standard systems. "By the mid-Nineties, we're going to have a shortage of skilled people," he says. "Controls will have to address this issue. We're going to have to offer more user-friendly systems, with more uniformity in the interface."

Denes Hunkar, president of Hunkar Laboratories, Inc., Cincinnati, agrees that the biggest problem with control systems is that they all require different training--an even bigger problem considering the expected labor shortage in this decade. "The OEM's will be forced by the mid-Nineties to develop standard vocabulary, displays, and setup commands. There has to be more standardization of information," he says. "I think you'll see this start to take place through SPI within the next three to four years; then it will take another two to three years to implement."

Hunkar believes that an important first step is SPI's new communications protocol. "When that becomes a popular and daily thing, a processor will be able to look at his auxiliary equipment with a uniform display at all times."

Others are less optimistic about the possibility of standardization, although some trends toward that end may be beginning to develop. Some machinery suppliers have introduced standard personal-computer software and hardware for either direct machine control or as a front-end interface for their PLC-based controls, which could form the basis for a standard. For example, Hettinga Equipment Co., Des Moines, Iowa, recently introduced controls based on a factory-hardened IBM PS/2 with Presentation Manager software. And a new operator interface from Husky is based on a factory-hardened IBM PC-AT compatible computer.

Husky expects more suppliers to offer controls based on such popular hardware, but it expects controls software to remain proprietary, although all systems will probably take more advantage of color graphics and much simpler interfaces.

Talks with other suppliers indicate that more standard hardware, if not software, is indeed on the way. Battenfeld's Meyer predicts that most controls will probably become IBM PC-AT compatible within a couple of years.

Meanwhile, Husky is also heading in the opposite direction in one respect. The company sees the speed of hydraulics and programmable control loops as limiting factors in overall process-control technology. Consequently, Husky is developing its own application-specific electronic hardware and software as substitutes for off-the-shelf controls. It has already started designing and building intelligent coprocessor cards to run in conjunction with PLC's. These cards are designed to provide greater capability in temperature control, position control, closed-loop control for the clamp, and more.

SPC AT THE MACHINE?

Klockner's Robert Hare predicts that statistical process control (SPC) will become a standard feature, integrated with machine controls, in the next five years. It's already standard on one of Klockner's machines: a 100-tonner introduced at K'89 in Dusseldorf last November.

But while agreeing that SPC will enjoy widespread use, other suppliers interviewed have no plans to make it part of their machine controls, because they believe it overburdens them and slows their responsiveness.

Husky, like most, expects molders in the '90s to be archiving anywhere from a week's to three months' worth of processing information and providing SPC reports and other quality reports to their customers on a regular basis. "Mid-size molders are now starting to become customers for this type of technology. We see it in the packaging market," says Husky's Urquhart.

Milacron's Reber is looking beyond SPC. "We need to concentrate on providing tools to make on-line quality evaluations. Hopefully, we'll be evaluating every part as it's being molded, not just samples," he says.

QMC: R.I.P.?

Denes Hunkar predicts confidently that automated or semiautomated quick-mold-change systems are "going to drop dead." In his opinion, these sorts of QMC are designed for the mid-size molder, which is not a sufficiently large market segment to sustain the technology. "Only large companies invest in expensive automation like QMC," he explains. "In plastics, large companies are mostly dedicated to large volumes of high-quality parts. They don't change molds regularly. Custom molders, on the other hand, tend to be smaller companies with less financial resources, making a living on short runs that demand somewhat less quality. Where does that leave QMC?" he asks.

Others interviewed are more optimistic about QMC's chances. Says Milacron's Reber, "There will be some QMC because it allows molders to provide JIT while keeping their inventories at a reasonable level." Bernie O'Donnell of Van Dorn agrees that "quick mold and material change will be increasingly important to the JIT concerns of custom and captive molders, both in terms of fast deliveries and low inventories."

Battenfeld's Meyer concedes that while automated mold change isn't needed everywhere, "There's a point where QMC is justified for mold maintenance, like with very large machines."

Nevertheless, suppliers cited a few problems hindering the effectiveness of QMC, including the fact that these systems often lack the flexibility to be suitable for relatively small runs, and that earlier machine models were not designed to accommodate such equipment. Another problem is said to be the molds themselves, which are hardly standardized, when it comes to setup. Others say that the layout of many existing plants doesn't provide the correct logistics for QMC systems.

Of course, there are several levels of technology and cost today in QMC systems. The most elementary level, which is affordable by a wide range of processors, simply involves mounting hydraulic quick-change clamps on the machine platens. These clamps are available from a number of suppliers, and together with standard-size mold bases and quick-connect couplings for mold utilities, permit substantial time savings in mold changing, even with use of a hand-operated crane or mold cart.

FLEXIBLE SERVO ROBOTS

Husky, for one, expects to see lots more robots used during the '90s. "The key to robots is that they have tremendous flexibility," says Bruce Coxhead, supervisor of product handling for Husky. He sees more complicated applications in the future, beyond mere pick-and-place. "It's getting harder and harder to hire people to assemble products. We think this will lead to more sophisticated robotics," says Coxhead.

Suppliers' answers to the expected demand for greater versatility from robots in the future is exemplified by a new "jumbo" robot from Battenfeld. It's capable of lifting a payload of 220 lb and can be programmed for sprue picking, hot sealing or printing, and stacking. The robot has seven freely programmable axes that can operate simultaneously.

Battenfeld's Meyer says that speed is another key issue for robots, to the degree that the robot's cycle extends the open time of the machine. In his view, that means teaming up robots to perform multiple operations faster through a division of labor. "We're seeing increasing requirements for more complex post-molding operations together with higher speed of production. The robot that removes a part may be supplemented by a tandem robot mounted on the same beam, which will handle post operations without slowing down parts removal," he says (see PT, Nov. '89).

Meyer believes these additional demands on robots will lead to the demise of pneumatic robots by the end of the decade. Electric servo drives appear to be a strong trend among robot makers worldwide. "We're going to see an evolution toward servo drives in conjunction with rack-and-pinion or timing belts," says Meyer. "Timing belts have the strength and durability to support larger robots."

AUTOMATION & MACHINE DESIGN

How automated the injection molding plant of the '90s will become is a matter of debate, but machinery suppliers are already taking steps to make sure their newest designs fit into the "factory of the future." Some examples:

* The need to integrate downstream automation equipment led Boy Machines to design its latest machine, the Boy 80, so that its clamping unit is accessible from all sides, being supported only at the corners. This design is said to eliminate restrictions on the placement and attachment of downstream automation equipment, without adding to space requirements.

* Cincinnati Milacron reports that its redesigned large Vista series are now more suitable for automation. Modular construction enables the machines to be easily customized by placing components such as the injection unit, hydraulic tank, and controls so as not to interfere with robots or other automated equipment (PT, Aug. '89).

* Battenfeld offers its "Compact Design" (CD) series of machines with a high-precision scale in the discharge chute under the mold for in-line weighing of small parts to [plus or minus] 0.001 g. It's fully integrated with the machine controls, and automatically separates good parts from bad ones.

* The latest Arburg CMD-series machines from Polymer Machinery are said to have hydraulics capable of changing molds, injection units, and material hoppers as standard options. The injection cylinder and screw drive can be uncoupled quickly and automatically and rapidly. Automated cylinder change takes minutes rather than up to an hour (PT, Oct. '89).

* Klockner's W series of large machines is designed for easy application of all kinds of automation equipment.

* The new Series Eight line from Sandretto Plastics Machinery, Farmington Hills, Mich., is designed to accommodate five levels of automation, ranging from quick-mold locking to fully automatic mold change.

Prospects for increased automation have pushed injection machine suppliers to look beyond the machine and get involved with automating the whole factory. Sandretto management, for example, believes the next stage in the development of injection molding machines will require even greater commitment to turnkey projects, highly automated plants, and other high-tech areas--some far removed from the machine itself. Sandretto's recent purchase by Fornara S.p.A. of Italy gives it the resources to support massive projects like its outfitting last year of the plastics molding section of an automated appliance factory for Zanussi, part of the Electrolux Group in Porcia, Italy.

That project involves 15 islands of automation, with 10 880-ton injection presses, and five 1100-tonners. All have robots and automatic QMC systems with mold tables beside the presses and cranes to deliver molds to the roller tables. Parts will be loaded onto pallets by the take-out robots; the pallets are accumulated on roller conveyors at the machines, and then are removed by two shuttle carts on rails and an overhead conveyor to the assembly areas. An automated mold storage and retrieval system will be installed this year.

The machine controls have automatic good/bad parts discrimination; also, an occasional pallet of parts is automatically shuttled to a q-c testing area for human inspection. In case of a malfunction that persists for a preset time, the machines are programmed to automatically purge themselves and shut down. The plant also compounds its own materials, automatically feeding polypropylene resin, fillers, glass fiber, and colorants, from bulk storage.

The plant will operate on a just-in-time basis, with 18-20 hours' inventory. The entire molding operation will be supervised by four persons per shift. All setup recipes for the presses and auxiliary equipment will be transmitted from a central location. A DEC central plant computer will download production orders to the molding area and retrieve all production and quality information. Sandretto is the prime contractor for this entire plastics installation and is responsible for supplying the supervisory control software.

This is just one example. Cincinnati Milacron also is getting more and more involved in turnkey installations, which it plans to carry all the way back to bulk materials handling in the future. And part of the recent equity investment in Husky by Komatsu, Ltd. of Tokyo will be used for a lab devoted exclusively to factory automation. "We'll train our customers for up to six months and we'll have up to 10 pilot production facilities," says Bob Schad. Husky, he notes, is committed to growth in factory automation for a variety of industries, including packaging, automotive and data-storage devices.

MORE MULTI-MATERIAL MACHINES

Spokesmen for Battenfeld, Husky, Klockner, Krauss-Maffei, Mannesmann Demag, Ube and Van Dorn all predict an increase in the number of machines equipped with two or even three injection units in the '90s. One reason will be to make more sophisticated parts combining different materials, for which "inmold assembly" through multi-injection is a more efficient process than multistep assembly operations. "This means less labor through the elimination of assembly," says John Grigor, president of JMG Associates, Torrington, Conn., the U.S. representative for Mannesmann Demag. "I also think some of the new, more expensive engineering resins will be co-molded with cheaper resins for economics."

Another reason, all agree, will be the increasing popularity--or necessity--of utilizing recycled material in the core of a product, with virgin material on the exterior visible surfaces. "I expect to see molders specializing in this type of technology," says Robert Hare of Klockner. O'Donnell of Van Dorn sees especially good potential for co-molding of recycled materials in the waste-container market, as does Battenfeld's Meyer. Multi-injection unit machines are also finding their way into the large machine market because processors investing in such expensive equipment want maximum flexibility to use the machines for a range of applications, say suppliers. Ube recently sold a 5000-ton machine with two injection units to Bryan Custom Plastics in Bryan, Ohio. The machine is equipped with a 1000-oz injection unit for waste-container molding and a 400-oz unit for automotive.

Versatility is important in small machines, too. At K '89, Battenfeld introduced its first coinjection unit for machines under 110 tons, with a patented two-channel nozzle system. The two injection units are of equal size and can operate independently or together, by means of a special nozzle. The patented system allows flexibility in the positioning of the injection units, from parallel to perpendicular configurations.

MACHINES `BIGGER THAN BIG'

The '90s are already shaping up as the era of the really large machine, from several recent U.S. purchases of presses between 3000 and 9000 tons (see PT, July, '89, p. 132; Dec. '89, p. 122; Jan. '90, p. 130; March '90, p. 105; and this month's Industry Briefs). Says Battenfeld's Wolfgang Meyer, "We'll see a machine size range bigger than what's been considered big in the past. The waste-container market has only been tapped in limited areas." Waste containers have been responsible for several of the largest machine purchases--including the Battenfeld 9000-ton "twin" machine, which consists of two 4500-ton presses "stacked" together (PT, Dec. '89, p. 122).

Parts handling and mold mounting will be the challenges for such large machines in the next decade, says Meyer. "With stacked or twin machines, there are eight tiebars, and parts have to be removed from the side. The mold also has to be inserted and removed from the side. And the robot can't interfere with the mold change." Meyer says we may eventually get to the point where we see very large unstacked machines. But for right now, he says it's not practical because the size of machines is limited by foundry capacity. Not many foundries can handle a casting as large as the platen for a 9000-ton machine.

Krauss-Maffei's Richmond says large machines will need more rigid clamp systems in order to maintain the new higher standards in part repeatability. "Clamp systems need to eliminate all platen deflections and mold parting-line flash," he says. Krauss-Maffei assures clamp rigidity through multi-point mechanical ram support with thick platens.

Another issue to be addressed is safety, especially when it comes to mounting mammoth molds in these machines (see PT, April '88, p. 78).

At least one supplier of large machines expects the shortage of skilled labor to affect the large-machine market. "There's going to be a tooling shortage," says John Grigor of JMG Associates. "There are not enough new toolmakers coming along. This is especially a problem with large machines--there just aren't enough people making the molds." He suggests that as the large-machine business increases, machinery makers might have to provide incentives to moldmakers to buy new equipment and upgrade their facilities.

SMALL MACHINES, BIG FEATURES

Several suppliers, including Joe Sweeney of Boy Machines, expect small machines to develop with most of the same features as large machines. "At K'89, we made a dramatic change in the controls of our small machines," says Sweeney. "I think you'll see development along these lines in the Nineties. Many optional control features will become standard. Generally, small machines will incorporate the same technology as large machines."

Both Sweeney and Bruce Bowman of Polymer Machinery expect to see more use of small machines for small parts, as opposed to using larger machines with multi-cavity molds. "I think it has been shown that it's more productive to use small machines." says Bowman. And Sweeney says, "Companies are considering small machines with either a single cavity or one to four cavities rather than large multi-cavity machines, because they find small machines are more efficient and cost-effective. You can realize more good parts per hour."

ENERGY EFFICIENCY RESURFACES

Ube's Dan O'Keefe says that the '90s will see a lot of emphasis on energy efficiency. "Processors are looking for more efficiency and higher-cycle machines to reduce energy expenses," says O'Keefe "We expect to see more use of toggles. Our toggle machines use only 40% as much energy as hydraulics do. Domestic suppliers are investigating large toggles now. And we expect to see more high-performance motors to cut down on energy consumption. The machines we've placed at Ford are equipped with power meters," he adds.

Engel Canada, Inc., Guelph, Ontario, is also betting on large toggles for energy efficiency. Sales v.p. Kurt Fenske says that advantage shows up strongest in long-cycle automotive molding. That's one reason Engel has started up a plant in York, Pa., to make toggle machines of up to 2000 tons.

Klockner, too, has responded to demands for increased energy efficiency. At K'89 it announced a new drive concept for the W Series of large machines that's said to optimize energy efficiency by adapting to demand. It features two electric motors--a basic motor and an additional motor that can supplement the first, if necessary. Other changes include alarms on the control screen that alert the operator if screw or injection speed exceed the power level of the selected operating unit. It also displays the energy consumption in kw/kg (PT, Jan. '90).

Klockner's Robert Hare cautions that there's more to efficiency than energy consumption. "To be considered efficient in the Nineties, a machine will have to be competitive in terms of floor space, energy consumption, configuration, maintenance requirements, and a host of other factors," he says.

Officials at Cincinnati Milacron are also convinced that energy efficiency will get a lot of attention in this decade. Dean Reber says the firm is developing more energy-efficient drive systems using variable-speed electric motor (see PT, Jan. '90).

And for any molder that wants greater energy efficiency right now, Milacron provides the option of its all-electric ACT machines. According to Milacron v.p. David P. Hahn. some 1500 ACT machines have been sold around the world, over 400 of them in the U.S. He says these machines have gained such a following because of the energy efficiency, precision and cleanliness (no hydraulics). Energy savings are said to be at least 30-40% of what a standard machine consumes, and one customer claims he has saved "at least 50-60%." That customer, Gary Thompson, national sales manager for Kamco Plastics Inc., Schaumberg, Ill., goes so far as to speculate, "I can imagine that hydraulics could become a thing of the past."

While perhaps not going that far, Milacron's Hahn thinks it's not wrong to imagine much greater use of all-electric machines. One reason is that their cost has come down from a previous 35-40% premium over a comparable hydraulic press, to more like 15%, owing partly to changes in currency exchange rates. In addition to the energy savings, Hahn says the all-electrics offer substantial savings in maintenance. Milacron just introduced a new line of all-electric last month.

HERE COME THE `HYBRIDS'

There's agreement among some machinery suppliers that all-electric machines represent a degree of "servo overkill." They suggest that a "hybrid" machine--one that combines servo motors and hydraulics--may be the way to go. In fact, Cincinnati Milacron reportedly will introduce such a machine shortly, although it was unwilling to disclose details at press time.

Nissei Plastic Industrial Co. of Japan introduced an 80-ton hybrid at K'89 (PT, Jan.'90). Nissei president Tsukasa Yoda told PLASTICS TECHNOLOGY that using hydraulics on the clamp end, where ultra-precise control is less important than on the electric servo-driven injection end, helps keep the cost down while using advanced technology where it's needed.

Klockner's Hare also sees distinct advantages in the selective use of electric drives. "I think you may see that anything on the machine that rotates will be electric, while anything that's linear in motion will continue to be hydraulic. That's Klockner's way of thinking," he says. "We think hydraulic components have advantages over mechanical components. But we see some real advantages to having an electric drive for closures and some other applications." Still, he thinks there will always be a premium to be paid for servo motors.

Wolfgang Meyer says Batterfeld can be expected to make use of servo motors in the future, although he agrees that all-electric probably isn't the best approach. "We'll employ servo drives in certain places for certain applications, but we don't plan an all-electric machine," he says. Meyer says it's probably advantageous to have an electric screw drive for energy saving and to vary speeds. But he says certain configurations, such as combining toggles (which he consider as less accurate clamping system) with a servo, probably don't make much sense. "Our field sales show that interest in electrics has slowed somewhat. We don't expect the percentage to grow significantly. And we still expect there will be a premium paid for servo technology."

Husky reports it has no plans for an all-electric machine, but thinks it can make its hydromechanical machines out-perform an all-electric, even in a cleanroom environment. "Electrics have limited capabilities for injection molding," says Husky's Urquhart flatly.

SCREWS TO MATCH EXTRUSION

Jim Frankland, president of New Castle Industries, Inc., a screw builder in New Castle, Pa., says that injection molding screw technology will catch up with extrusion screw technology during the '90s, especially in areas like mixing and wear resistance. "There are lots of new materials, many of them incorporating reinforcements and many of them alloys requiring greater mixing. So we'll wear-resistant materials."

Frankland also expects to see increased use of barrier screws for injection, because they provide more controlled melting of crystalline polymers than conventional screws. Also look for improvements in vented screws, because they help eliminate some of the volatiles that are more common in engineering materials, Frankland says.

Other screw trends include better sizing of the injection unit to suit a particular application. "People are buying screws based on the shot and recovery required," says Frankland. "In response, many machinery suppliers are offering a building-block approach to injections units." The problems encountered in outfitting a machine with more injection capacity than required include degradation of resin from too much residence time and poor control of the shot, he says.

Most machine suppliers don't expect much activity in screws in the '90s. Husky looks for incremental improvements in screw performance during the decade. "There'll be more quick-barrel-change systems, and fewer general-purpose screws. The cost of extra screws is worth the improved processing" says Husky's Urquhart.

SERVICE-BY-MODEM IS COMING

"Maintaining a service staff is expensive. The OEM wants to cut his service costs while maintaining an acceptable level of service. Processors, on the other hand, would like free service. So we'll start to see a lot of service by phone through modem connection," according to Denes Hunkar.

Modem-based diagnostic allow machine problems to be solved over the telephone, saving the time and expense entailed in travel. Even when a visit is necessary, preliminary diagnostic work can be performed via modem. Sources at Control Technology Corp., Hopkinton, Mass., says that thanks to recent price reductions in high-speed modems, its customers can now use its Quickstep and other diagnostic software products over standard dial-up telephone lines to perform diagnostics on machines that may be thousands of miles away.

Boy Machine's latest models have a modem connection port built as an optional feature. And other suppliers including Husky, Cincinnati Milacron, Battenfeld, and Ube say they'll implement some form of modem service. "It means a big change in operations for the processor, but it can reduce the amount of downtime," says Hunkar, who predicts that most processors eventually will have modems, and that they'll be communicating not just with machinery OEM's, but with resin suppliers as well. "All these people will confer simultaneously," says Hunkar, who believes this will be happening less than five years down the road.

Modem-based diagnostics may also alleviate the training burden. With personnel turnover at the machine-operator level said to be high, modem-based diagnostics can solve problems more quickly than an inexperience operator.