is part of the Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.


Articles from 1997 In January

Selective foaming for strength and speed

Coralfoam, a patented technology recently developed in England for selective foaming of molded plastic parts, greatly increased its visibility and accelerated its licensing activities at the recent Interplas exhibition held in Birmingham. Combining an endothermic blowing agent with specific mold design and process control technology, Coralfoam allows the foaming of selected areas of a part, thus changing the shape of the molded part during the mold cycle. Key benefits of the process include reduced part weight coupled with increased rigidity, faster cycle times, and thermal insulation, plus the ability to design parts not previously possible.

Polypropylene cups are only one of many applications that might benefit from the increased rigidity, reduced part weight, and faster cycle times allowed by the Coralfoam process.

The beverage cups shown in the accompanying photos were produced during the Interplas show on a Mannesmann Demag Ergotech ET 35-120 Compact machine using Elf Atochem's Appryl high-clarity polypropylene. The tool was an existing one modified for the Coralfoam process. Samples of the nonfoamed cup were available for comparison. The foamed version is 6 to 7 percent lighter and three times stiffer than the previous version, and it is molded in a shorter cycle time. Polished windows have been added to the tool's original matte surface to show the level of clarity possible.

Peter Clarke, the developer of the process and technical director of Pentex Sales Ltd., which is handling the marketing, noted that while modifying a current mold will work, designing a mold from scratch will yield even more dramatic results. For example, a 50 percent reduction in weight is possible while still increasing rigidity. Cycle times are fast, since they are dependent on the thin-wall sections rather than the thick, foamed areas. Foaming actually occurs in fractions of a second after the mold is opened, and temperature differential between mold halves is an important factor.

Normal Molding Machinery
At another Interplas stand, a 4-inch flower pot was being molded using Coralfoam technology on an 80-ton Netstal machine. Cycle time was 3.7 seconds for a part with a wall section of .5 mm and a 6-mm foamed lip. According to Clarke, a quality standard injection machine with a precision NC controller is all that's needed, plus a good dosing unit for the foaming agent. The injection machine needs no modification. There is also the added cost for the endothermic blowing agent, which is Safoam from U.S.-based Reedy International, but that is more than offset by the reduction in overall material costs.

Pentex/Coralfoam has been actively working with container manufacturers in the U.K., but sees lots of other markets and products that could benefit from the process. Pallets could be a major market. Then there are crates and boxes, coat hangers, razor handles, safety helmets, closures, cutlery handles, toilet seats, and more. Car bumpers made of one material instead of the usual skin plus core could greatly simplifying recycling. Interior auto posts could have a thickened sealing bead. Pentex has already molded against another material to create a thin skin with selectively placed strengthening ribs.

From the design point of view, Coralfoam offers the ability to place the foam where you want it, and to have foamed sections with a high-gloss surface. You can place foam selectively in transparent parts for functional or esthetic reasons, or have 10-mm undercuts with no need for split cavities, or 10-mm-thick ribs along with a 6-second cycle time and no worry about sink marks. The temperature control can be extended to multi-cavity molds.

Alternative to Gas Assist
Since Coralfoam can place either foam bubbles or voids in the selected areas, it can be seen as a competitor to gas-assisted molding for creating certain types of lighter, stiffer parts. Pentex notes the lack of the nozzles, gas pressure generators, and controls required for gas-assisted molding. Being a relatively low-pressure process, Coralfoam can also save costs on mold construction and permit use of a lesser tonnage clamp. For example, a production mold for a 2-liter ice cream container was modified for Coralfoam. The shot weight went from 56 to 46g and cycle time from 6.3 to 4.9 seconds. Before modification, the mold needed more than 250 tons of clamp; with Coralfoam it moved to a 175-ton press.

Selective foaming creates both insulation value and increased rigidity, without extending cycle time.

Production costs, says Clarke, for a mold developed from scratch will be relatively the same as if it were for normal compact material molding. The tool modifications mentioned above cost about $8400 per tool. In terms of materials applicability, Clarke says that virtually all polyolefins have been tried with good results and that the process is equally applicable to engineering materials. The company is eagerly looking at applications where PP can replace styrene. That gives the obvious advantage of lower raw material cost per pound, plus a lower specific gravity for more parts per ton, all while increasing rigidity. A styrene coat hanger converted to PP Coralfoam processing weighs 30 percent less and has double the stiffness.

Marketing and Licensing
Coralfoam and Pentex have been busily marketing the new technology, thus far primarily within the U.K. Their partnering relationships, which are aimed at providing broad support to licensees, include Netstal and Mannesmann Demag for machinery; Reedy, Montell, and ATO in materials; plus toolmakers and product designers. They have conducted numerous demonstrations and seminars, generated a lot of interest, initiated a number of developments, and will continue giving seminars, both group and private, to potential Coralfoam molders and end-users. Details of their entry into the U.S. market will be decided in early 1997. In general, they are expanding in line with their ability to give full support to licensees.

Originally, they were considering licensing production sites for the process but have changed the strategy. They are now seeking to license processors for the production of specific product groups, for example coat hangers, in a specific geographic area. They are also offering negotiable options that could extend up to a year to allow development. Cost of the option would vary from roughly $50,000 to $336,000, depending on the size of the market being considered. The cost of the license itself will also vary according to the market, but will consist of an initial purchase price plus royalties on production.

Shot inventory, or, if this is a 24-oz machine, why can't I run a 24-oz shot?

Old, tired pelletsWhy is shot inventory/residence time so important? The rule of thumb in shot capacity is to use no less than 20 percent and no more than 80 percent of the rated capacity of your barrel. Another way to say this: you should have no more than five shots and no less than one and a quarter shots in inventory in the barrel. Take the total part weight from all the cavities, add the sprue and runner weight, and divide this into the machine's rated shot capacity.

A shot inventory of greater than 5 means you are at the low end of the barrel's capacity. The inability of the machine to accurately control moving a few hundred pounds of steel (the screw) with any degree of precision over a short distance will explain why your parts are remarkably inconsistent.

Molders generally compensate for this by turning down the heats. This puts unmelted material in the front of the barrel, which then is melted through shear heating. This process degrades the physical strength of the part in two ways. One is the mechanics of ripping the material apart; the other is the injection of cold material into the mold, where it cannot properly weld. If you're in this fix, one solution (although I'm generally opposed to it) is to make the runner system larger to take up the volume. To offset the cycle time increase, be sure you have a cooling line under the runner or directly in the runner plate.

A shot inventory of less than one and a quarter means you are over the top of the machine's ability to adequately melt the plastic and hold a cushion. Here is where a hot sprue or a hot runner can really help. Because the sprue/runner is eliminated, there will be less volume in the usable shot, and you might be able to reduce it to result in a higher inventory.

Many people are in the microprecision part molding business—their parts weigh in fractions of a gram. Routinely this means 80 to 90 percent of the shot is the runner system. The shot still uses only 1/10 of the machine's shot capacity. Ideally, this seems like an application for a hot runner. However, there are other considerations.

Many companies sell hot runner systems and only a few molders really know when and how to use them. While virtually any mold can be converted or originally built as a hot runner, when does it make sense and when is it more trouble than it is worth?

The first consideration is figuring out the volume lifetime of the part. Products with a low volume generally don't justify hot runners unless the design demands it because they're expensive to install. The auxiliary equipment can also cost a lot.

The next consideration is the press. I don't know why, but this is usually the biggest failure of communication between the molder, the mold builder, and the salesman for the hot runner system. When we install a hot runner system we are, by necessity, eliminating the sprue and runner. This means no regrind, which is usually considered good. It also means a much longer mold is needed to accommodate the runner system, and the press must be able to open and close.

At the same time, you have a much smaller shot size. In many cases these concepts are mutually contradictory. Since we have to get the mold open and closed, most molders will live with an extensive shot inventory in the system. What if the shot inventory is above the shot inventory level of 5? You need to consider some options:

  • Since the thickest portion of the molded shot is often the intersection of the sprue and the runner, eliminating it will usually improve the cycle. You have two choices: make a smaller sprue so the intersection is smaller, or use a heated sprue that will eliminate the volume of the sprue, leaving you with only a runner. However, eliminating the sprue has now increased your shot inventory. You'll need to offset this by making "poker chips" (round flat dummy parts that take volume but no cycle time).
  • If you're molding in small machines, you might consider a still smaller machine or a smaller screw/barrel combination that will put you back into a good shot inventory.
  • Some parts lend themselves to a semihot runner system. This is a full hot runner system that fills into a foreshortened conventional runner system—ideal for parts that need multiple gates. It saves dramatically on the material usage of a full runner and sprue conventional system.
  • Sometimes fate smiles on us. Look at your cavity layout. It might be possible to put in more cavities with either a full hot runner system or a semihot runner system. You are no longer bound by the "power of two rule" in conventional sprue-and-runner mold layouts that dictates two, four, eight, or sixteen cavities. With a properly balanced runner system, you can make a five, seven, or 13-cavity hot runner tool!
  • For those on an extremely limited budget, there is still some compromise available: What if you used an extended machine nozzle and a short sprue? What if you used a really extended machine nozzle and didn't use a sprue at all? Either of these solutions costs only a few hundred dollars and has a remarkable payoff.

Now, what if the shot inventory is less than one and a quarter? Using almost the entire shot capacity of a machine to fill a mold results in a high amount of sinks, short shots, and warp. This is because you usually cannot melt enough material to pack the part in a single shot. Also, it's hard to hold the cushion to pack the shot. A hot sprue can help to eliminate the amount of plastic required.

The use of a hot runner, hot sprue, or extended machine nozzle is an obvious solution for two reasons. First, by saving material consumed, the shot inventory will go up, hopefully higher than one and a quarter. Second, having the inventory up will save on scrap and cycle time.

When do you not want to use a hot runner? Remember that hot runners and heated sprues are merely an extension of the machine. Every overheating problem you get with a machine you'll get with a hot runner system. Thus, parts with extreme cosmetic requirements are not good candidates for heated runner systems. This is equally true for materials that are not particularly heat stable (medical grade vinyl) or those with thermally sensitive colorants, such as pastels or fluorescents. This is not to say you cannot run any of these materials with a hot runner system. However, your expertise and understanding of the system must be good and your people highly professional.

With multicavity tooling, it is a common practice to block off a cavity. This is bad for several reasons. The major damage is to productivity and profit. But blocking off a hot runner tool also leaves in the mold idle material that is constantly cooking. This burned material will never be truly cleaned out with purging compound until the mold is sent to the manufacturer, disassembled, and properly cleaned.

Many people still think that with a hot runner/hot sprue there is no such thing as a cold slug or the need for a slug well. Do this mind experiment. If you actually had the machine nozzle filling the part directly, each time the mold opened the part would exhibit a string. If it didn't, this would mean the plastic has slightly frozen and the part broke away. This frozen material is your cold slug. If your mold is designed properly, you have a cooling circuit next to each tip. This stops the heated tip from trying to heat the mold and forms the cold slug in the heated nozzle. You therefore need to design a slight depression in the part to accept this slug as part of the molded part. If not, you'll get it anyway as a cosmetic defect commonly known as blush or smearing.

Running quality parts is like carpentry. While you can put up a house with only a tack hammer, there are better tools. Think about your application, then make the proper choice. Everything has its place, which, when done properly, maximizes profit.

Modulus has designs on customers' success

Aeron Chair
Aeron Chair design
The Aeron chair by Herman Miller won IDSA's design of the year award. Computer-aided engineering helped Modulus designers optimize the chair frame design, select a paint-free PET material and thus remove a step in manufacturing, and still exceed the customer's strength requirements.
Business as usual at AlliedSignal's design center, where the Modulus Design Group makes camp, is advanced enough to satisfy the most demanding standards. A recent visit confirmed that Modulus, believed to be the only design engineering group in the industry with a trademarked name, is a powerhouse of computer-aided activity. At one workstation, CAE engineer Dawna Schultz conducts an FEA on a power tool housing with a nonlinear package from Ansys. Elsewhere, Craig Scott, a CAD engineer, finishes up a 3-D model in I-deas and prepares it for mold filling analysis.

Most resin suppliers offer design support services; however, in recent years, those services are free only to large customers and, in some cases, must be purchased. Modulus appears to be one of the few that remains available to all customers. So just what does Modulus offer?

What it doesn't offer, you probably don't need. Modulus' services include 3-D solid modeling; analysis of flow, gas, cooling, and warpage; tooling design; FEA for structural, impact, acoustic, fluid, or heat conditions; design optimization; rapid prototyping (both SLS and SLA); and development of composite tooling.

"Our business philosophy is to provide total customer service from concept to part delivery," says Linda Chamberlain, director of research and technology. Her department consists of 80 people who support the three technology functions that make up AlliedSignal's Research & Technology organization—Technical Service, Modulus Design, and New Product Development. "That means supplementing our customers' capabilities from injection molding to a full range of analyses, including structural, mold filling, and cost. Sometimes, it means starting from scratch and completely redesigning the part."

Modulus team members work together to find the optimum design, or redesign, and material for a product. A majority of the time, team members can be found onsite at molders and OEMs helping to train operators and simplify production startups. The group also holds seminars to help customers optimize in-house testing, analysis, molding simulation, and design.

AlliedSignal recently upgraded hardware in the design center with a $600,000 investment in new Hewlett-Packard Unix workstations. As far as software goes, Modulus is equally well stocked with the most recent versions of Cadkey Professional, Catia, I-deas Master Series, Pro/E, Abaqus, Ansys, Hypermesh, Optistruct, Moldflow, MF Warp, C-Flow, C-Warp, and TK Solver. And all systems are linked via local area network for quick communications.

X-core 5
The X-Core 5 is a one-piece hollow wheel molded by the lost core process and marketed as a direct replacement to conventional wire spoke wheels. Modulus designers selected an impact-modified nylon 6 with 25 percent glass fiber for strength, stiffness, and long-term durability in this wheelchair application.
Power tool fan
Makita worked with Modulus to redesign this fan for a power tool in Capron HPN, a high-flow nylon 6. Cycle times were reduced by 30 percent without sacrificing performance properties.

Putting all of this design muscle behind some recent products underscores why Chamberlain says, "We've made these investments in Modulus to provide greater customer support early in the design/development cycle." For example, Ford Motor asked its supplier Comcorp to design a plastic accelerator pedal, replacing a steel one, for the 1999 model of a popular vehicle. Both OEM and molder turned to Allied for help, and Modulus came up with a material and a design methodology to do the job.

The material was a 33 percent glass-reinforced nylon 6 (Capron 8233). The methodology is called "proprietary structural design optimization process analysis," and allows Modulus designers to do computer searches for designs with the least volume and highest performance. This led to a C-channel pedal with a reinforcing rib pattern that consists of angled parallel ribs crossed intermittently with diagonal ribs. Ford saved 10 percent on cost and 50 percent on weight.

Makita had been using the same material in a fan for one of its power tools. During a redesign, however, Modulus designers substituted Capron HPN (a high-flow version). The switch cut cycle time by 30 percent and improved aesthetic appearance, all without sacrificing any performance properties. Molder IKKA reported that the material processed well at 15 deg F lower than did the original nylon 6, saving energy as well.

Another automotive customer worked with Modulus to develop an air intake manifold, also for a 1999 vehicle. The part will be vibration welded, so the group performed weld-strength testing to customer specifications.

Herman Miller's Aeron chair, also a Modulus project, represents one of the most stylish applications for Allied's paint-free Petra, a PET grade designed to give great aesthetics in an unpainted condition.

Modulus designers optimized the chair frame structure to the point where its strength exceeded the customer's expectations, according to Chamberlain. She also mentions that Allied is now establishing groups closer to customers. "We now have a Modulus group in Southfield, MI in order to better serve our automotive customers. Recent areas of support include air intake manifolds and accelerator pedals," she notes. It contains four design engineers and two CAE engineers. Another, located in Heverlee, Belgium, serves European customers.

In a related development at Allied, Chamberlain told IMM that the online ordering system announced this year is being prototyped with select customers. Once proven effective, the system will be expanded and will guarantee a 48-hour delivery for 60 percent of Allied's product line; the rest will remain made-to-order. "We're taking the time now to ensure there won't be any security issues at play, and to make sure that orders can be verified and cross checked for accuracy," she notes.

Mold design wizardry goes full throttle

Figure 1
Tomco Plastics uses a warped tool to produce this straight part: a glass-filled PPS intake manifold throttle control housing for the '97 Lincoln Mark VIII.
When awards were handed out in the most recent SPE Automotive Innovation design competition (December 1996 IMM, p. 48), it was hard to decipher why certain finalists went away empty handed. For example, one such process nomination involved "tuning" a mold to compensate for the way plastic cools in order to maintain tight tolerances and flatness. Unusual? We think so. Effective? You decide.

Custom molder Tomco Plastics (Bryan, OH), Ford Electrical and Fuel Handling Div. (EFHD), and resin supplier Phillips worked together to design and produce a glass-filled PPS intake manifold throttle control (IMRC) housing for the '97 Lincoln Mark VIII. Briefly, the IMRC restricts air flow in an engine's intake system so that air velocity increases at low engine speed for more efficient combustion; at high speeds, the throttle valves are opened for maximum air flow and power.

IMM spoke with Bob Wisler, market development manager for Tomco, who confirmed that the IMRC housing was a team effort among all parties involved: "We had dedicated engineers to work with at Ford, starting with Walt Fedison (lead engineer, EFHD), and the team collaborated together on every aspect of this project. This shows what can happen when your agenda is to manufacture a quality part that saves money." Design Analysis

To match the flatness of machined aluminum, Tomco has developed a creative way to achieve the tolerances. Rather than making the plastic conform to dimensions, explains Wisler, Tomco applies a proprietary process to contour the mold so that the plastic part will "cool" itself to the required flatness. Starting with a steel-safe mold that could be contoured later, the team produced hundreds of parts. Mold cavities contained perfectly flat surfaces at this point, and owing to the material's nature, the parts warped slightly upon cooling. Tomco engineers then took the detailed layout data from these parts and ran it through a computer program developed in-house that converts the data to mold dimension. Toolmaker Kellems & Coe received the required dimensions and cut them into the mold steel, resulting in molds "tuned" to compensate for warpage.

Tomco then began production on a 300-ton Van Dorn HT. Ford thought that the 1-inch-thick parts needed a higher tonnage machine, Wisler recalls, but Tomco's injection molding expertise resulted in a flat, 3-mm-thick PPS part weighing about 1 lb. Parts hold tolerances to .20 mm for sealing surface heights and .25 mm for the injector pod true positions, matching the accuracy of former machined aluminum versions. Bottom line: the unconventional approach saved Ford 30 percent in both weight and cost.

Says Wisler, "It wasn't just the computerized mold contouring process, it was the can-do attitude that said we know we can hit tolerances this tight. We feel it's our job to understand the polymer and its properties, plus how it reacts with the proper injection molding process. When you start molding something for dimensions' sake only, you sacrifice both product and process integrity. We mold the product for the product's sake. As much as possible, we needed to be able to mold a stress-free part, letting the material shrink and warp the way it wants. Our history with these types of polymers proves that the parts will shrink the same way every time."

On the drawing board at Tomco are several related projects: a 2-inch-thick version of the IMRC for another Ford engine; another manifold component; and fuel rails. "We apply this technology to everything we do, but we've never had a part this complex with drop-in machined metal tolerances," Wisler notes.

Sliding dies for molding hollow parts

Die Sliding during Injection (DSI) was developed by The Japan Steel Works Ltd., Tokyo (JSW Plastics Machinery Inc., Santa Fe Springs, CA) to overcome problems JSW technicians believe to be inherent in other processes that are used in the production of hollow parts, such as blowmolding, lost-core molding, and welding. DSI is designed to produce high-performance hollow products with inner ribs and bosses, like high-quality/low-cost automotive parts and E/E assembly encapsulations, some of which might be impossible to make using other processes. Wall thicknesses are adjustable using DSI. Work continues to improve DSI's applicability to more general-purpose hollow parts production and multimedia molding.
Figure 1
Figure 1. How the DSI molding system works.

JSW technicians say fluctuations in wall thickness, particularly at corner areas, is a weakness in blowmolding: heterogeneous parison expansion and the inability to accurately control uniformities in specific local dimensions cause such fluctuations. As a result, blowmolded hollow products can lack strength. Regarding lost-core molding, JSW technicians say the process is unsuitable to cost-effective mass production, since preprocessing is required to mold the cores, and since postprocessing is required to remove them. What's more, additional processing is required to make lost-core molded hollow products airtight. And welding? They say welding is unsuitable for producing products with complex shapes in high volumes. Defect rates may climb.

As you can see in the accompanying drawings (Figure 1), a hydraulic unit installed on the fixed platen operates the sliding die in the DSI molding system (A). First, both halves of a product are simultaneously molded in two cavities (B), and the mold opens (C). A sliding die moves one half of the molded product to a position where the parting-line surfaces of both halves face each other (D), and the mold closes again (E). Next, at the second injection stage, what JSW technicians call a "circumference part" is overmolded to combine the two parts produced in the first injection stage (F). A single hollow product is produced. DSI cycle times can be anywhere from 1.3 to 1.5 times slower than straight injection, but costly secondary processes are eliminated and complex hollow parts with surface appearance equivalent to straight injection molded parts can be produced. Defect rates plummet. And, hollow parts can be stronger with molded-in ribs and bosses. Three part-design possibilities are illustrated in Figure 2.

DSI processing requires mold and molding machine modifications. Molds must have two cavities for molding product halves. Particular attention has to be paid to mold design, since the two molded product halves are not removed from their cavities after the initial molding sequence and are held in place. JSW has patents pending on its DSI mold design technology. Also, JSW injection molding machines equipped for DSI processing come with controllers smart enough to handle the different sets and additional sequence steps of molding conditions for molding product halves and the circumference part.

Figure 2
Figure 2. Sample product designs molded using the DSI system.

JSW offers 11 models of its E series machines, ranging from 30 to 850 metric tons, with DSI systems as standard equipment. Yet company sources say DSI can be adapted for use with other machines bearing the company marque. A wide range of commodity and engineering materials can be used, including FR and foamed grades, as can materials combinations. At IPF '96 in Japan, JSW demonstrated its latest two-shot/two-material M-DSI process on a twin-barreled 150-tonner, model J150EII-P-2M. Products with ABS as the core and a circumference seal of TPE were molded at the show. JSW also demonstrated how DSI technology can be used to produce two different types of parts designed for subsequent assembly in the same material. A 40-metric-ton vertical clamp/ vertical injection model JT40REII-20V molded two different parts in a 30 percent glass-filled LCP using two different mold faces carried on a two-station rotary table.

Analyzing plastics with FEA: Part 4

Sample Geometry
Figure 1. Sample Geometry.
When designers consider using finite-element analysis, they often want a tool to simulate the effects of loading or heat. There are other, fairly well-known benefits of the finite-element method for plastics, including mold filling and cooling simulations. But there are not a lot of people modeling the gas-assist process this way. If the results of a recent study are any indication, more designers may use FEA to optimize gas-assist designs.
Figure 2. Effect of volume fraction on gas penetration.
Volume Fraction

Tom Kowalske of DuPont Automotive (Troy, MI) realized it would make sense to simulate the effect of process variables on gas penetration so that customers could cut down on blow-through and related defects normally requiring a good deal of trial and error to correct. To this end, Kowalske contacted Shrikant Oak and Ross Nordin at another Troy-based company, Feamold Inc., that specializes in injection molding process simulation and design optimization for the automotive industry.

Oak used a finite-element/finite-difference, or hybrid solver, to analyze how the following variables would affect gas penetration: volume fraction, melt temperature, gas pressure, packing time, delay time, and channel dimensions. In performing the simulations, Feamold measured gas penetration as the length of primary and secondary penetrations, change in volumetric fill time, and uniformity of gas/skin distribution.

Gas Penetration
Figure 3. Effect of melt temperature on gas penetration.

Both Kowalske and Oak shared the results of their experiments with IMM, adding the caveat that similar studies performed by other researchers have yielded contradictory results. "The results of this study are specific to the combination of geometry, material, and processing window we used," notes Oak. Geometries consisted of various lengths of a long, flat plate similar to an "A" pillar with a gas channel running along one long edge of the part, Figure 1. Material properties used for all the analyses were those for PET (DuPont Rynite RE5309 BK515). Processing conditions changed for each variable analyzed. And now for the findings.

Volume Fraction
Primary gas penetration length decreased with increasing volume fraction, or prefill percentage, in all cases. At 70 percent prefill, the gas raced beyond the melt front and blew through, as illustrated in Figure 2. Check the straight lines in that graph, and you'll see that gas penetration decreased in a linear fashion as the percentage of prefill increased.

For part one of this experiment, fixed inputs included a melt temperature of 585F, mold temperature of 200F, gas pressure of 1000 psi, gas injection of 10 seconds, and delay time of 0 seconds. The part being modeled was 33 inches long, with a 1/2-by-1/2-inch gas channel. For part two, the only changes involved were a 2000-psi gas pressure and a 66-inch-long part.

Skin Thickness
Figure 4. Effect of melt temperature on skin thickness distribution.

Results of both experiments showed that there are two factors that affect primary gas penetration more than any other parameter. They are volume fraction and melt-front advancement pattern. Volume fraction can be used to control the amount of gas penetration, while melt-front advancement tailors gas direction.

Melt Temperature
Results indicated that secondary gas penetration and total gas penetration increase linearly with melt temperature, Figure 3. This makes sense intuitively, because a higher melt temperature will induce greater part shrinkage; thus, more secondary gas penetration occurs to compensate for the increased shrinkage.

While this analysis showed melt temperature having little effect on primary penetration, Oak cautions that highly temperature sensitive materials may behave differently: "They can experience a greater change in viscosity as temperature rises, and this in turn can reduce primary penetration and affect volumetric fill times."

One area that showed distinct changes as melt temperatures went up was skin thickness distribution, which becomes more uneven with increasing melt temperature, Figure 4.

Figure 5. Gradients across polymer-gas interface (sectional view).

Gas Pressure
For a 100-inch-long part with a 1-by-3/4-inch gas channel, gas pressure was set at 1000, 2000, 4000, and 6000 psi with all other inputs remaining the same as in the volume fraction experiment. Experiments were run at 80 percent prefill with a resin injection time of 3.2 seconds, after which gas was injected.

As gas pressure rose, primary and total gas penetration decreased along with volumetric fill times, while secondary gas penetration went up. To explain this phenomenon, Oak says that the polymer-gas interface can be illustrated as made up of multiple layers of plastic, as in Figure 5. The layer next to the mold wall is coldest, while that in the center is hottest. Colder layers have higher viscosity, needing greater pressure before they will flow. So when gas pressure goes up, more layers begin to flow. As they become more mobile, says Oak, the cross-sectional area in front of the gas bubble grows, decreasing the length of penetration.

Gas pressure
Figure 6. Effect of gas pressure on gas penetration.

Gas pressures must be above a critical lowest value before any plastic can be pushed. Notice how there is no gas penetration below a certain gas pressure in Figure 6. This value, says Oak, depends on geometry, flow length, material, and process conditions.

As gas pressure rises above the minimum value, there is an increase in gas penetration similar to conventional molding in which flow lengths increase. But too often, if molders still see insufficient gas penetration at this point, their reaction is to increase gas pressure even further. Oak cautions that beyond a critical maximum pressure, boosting the gas only decreases the length of penetration.

If, at reasonable pressures, parts aren't getting enough gas penetration, the solution probably lies elsewhere. "Take a look at part design," says Oak, "and check the melt-front advancement pattern and end of fill."

Packing Time
As packing times go up, secondary penetration increases in linear proportion, corresponding to the rate of cooling related shrinkage. Both primary penetration and volumetric fill times are unaffected. Skin thickness, however, became more uniformly distributed as packing times were boosted. PET gas process parameters

Delay Time
Increasing delay time boosted primary gas penetration up to a critical value, in this case .5 second. Further raising delay time past this point caused a decrease. Oak explains: "Remember the layers (Figure 5) at the gas-polymer interface? As delay time increases, temperature across those layers goes down. This reduces the number of layers in front of the gas bubble that can be moved at that level of gas pressure, so the primary penetration length increases. Once you reach the critical value, however, the plastic freezes excessively, and there are fewer fluid areas remaining for the gas to push. So penetration length decreases."

Summing It Up
Feamold's Oak stresses the importance of analysis for gas-assist part and process designs. "An acceptable or target level of gas penetration must be defined. This process is extremely sensitive to process variables," he notes. For example, if gas is injected too early, it reduces the melt cushion available and increases the potential for gas blowthrough. If gas is injected too late, the plastic cushion reaches the end of the flow path before the gas has advanced far enough, leading to highly cored out areas close to the gas pin and solid sections away from it. This uneven thickness distribution means different cooling and shrinkage rates, thus highly stressed parts that can warp unacceptably. Table I summarizes the Feamold findings for PET.

Autofact '96: CAE now sees through plastics lenses

Wireframe-to-solids converter
CADKey '97 will feature a wireframe-to-solids converter so that moldmakers can bury part designs in a block, then subtract them to create mold surfaces.
As little as five years ago, the world of computer-aided design, manufacture, and engineering that convened for this expo presented a real paradox: here were the most advanced design and manufacturing tools in the world; yet, in general, their focus was still firmly on traditional materials—namely metals. Perhaps the dollar value of products still made from metals was a factor in this myopic view, but it made little sense to nearly overlook an emerging market such as plastics.

This year at Cobo Hall in Detroit, Autofact exhibitors restored our faith in the intelligence and keen eyesight of the CAE community. As a group, it appears to have overwhelmingly grasped the fact that more engineers are designing with plastics, and that computer tools are nearly essential for the job to be done effectively. Of course, there have always been vendors with this mindset, especially those marketing products such as Moldflow and C-Mold. What has changed are general-purpose CAD/CAM packages as well as FEA offerings, which now reflect the significant position occupied by plastics in automotive, medical, IT, and a litany of additional markets.

MF/Flowcheck analysis package
MF/Flowcheck, a preliminary filling analysis package, is now integrated within Autodesk's Mechanical Desktop.

Other changes are afoot in this industry as well. Unlike its splashier, consumer cousin Comdex (for coverage, see pp. 30-31), the mix at the Autofact show is changing. Formerly large exhibitors such as Parametric Technology Corp. and Autodesk were either completely absent or represented only as a partner in a "group booth." On the other hand, we saw several product design and development firms here for the first time. Those that specialize in plastics included Lexmark's Plastics Technology Center (PTC), Compression Inc., and Plynetics/ Prototype Express. PTC's Steve Spanoudis noted that his company finds that design engineers in general increasingly seek to outsource this type of plastics design expertise.

Mergers and acquisitions among many of the industry's key players have contributed to a scaled down expo floor. But although show size was smaller, sales of software appear to be healthier than ever. Market research firm Daratech forecasts those revenues will top 18 percent growth for 1996, breaking previous records.

Formula 1 Minardi
The engineering team for the Formula 1 Minardi uses Cimatron's design, analysis, and drafting modules.

IMM made several interesting discoveries during this year's exhibition, hosted as always by the Society of Manufacturing Engineers. First, hot topics on the Autofact conference agenda have also been covered recently in this publication (see sidebar, "For your information"). Secondly, we tracked down a clear, concise explanation of recent advances in solid modeling (see sidebar, "What's in a solid?"). And finally, we found computer tools aimed directly at plastics designers, molders, and moldmakers, all of which are detailed below and on the following pages.

Solid Updates
A third version of Autodesk's design-through-manufacturing solution, the Mechanical Desktop 1.2, will be available this month, according to Guri Stark, Autodesk director of industrial marketing. Customers such as Trane, Siemens, and Philips are using the hybrid modeler to unite 2-D wireframe and 3-D solids, reports Stark.Autodesk has also assembled fifteen software vendors that it calls MAI solution partners for Desktop, including Moldflow and Ansys, to offer users a complete suite of art-to-part capabilities. Similar to other modelers, this is a feature-based, parametric solids package with bidirectional associativity and Nurbs-based surface modeling. Unlike competitors that run only on Unix, however, Desktop is PC-based.

Euclid Quantum modules
All of the modules within Euclid Quantum, including Designer, Analyst, and Machinist, are linked to Euclid Desktop for access to the Web, intranets, and easy data exchange.

The new release will feature a simplified user interface along with edit-in-place capabilities—that means users can work in assembly mode, adding parts as necessary. An interesting aside: Stark noted that "Gunslingers," an expanded quality program in which customers spend a week with developers to find and correct bugs, will be performed twice for each release of Desktop.

After recently unveiling the Euclid Quantum series (November 1996 IMM, p. 66), Matra Datavision further underscored the value of its integrated design, analysis, and machining package for concurrent engineering environments. The same technology underlying Quantum, which Matra calls Cas.cade, will be applied to other products, the company confirmed. This includes Strim for plastic part design, analysis, and prototyping as well as Moldmaker with its built-in library of 3-D parts.

CADKey, a Windows-based solid modeling package, is now owned by Baystate Technologies. President Robert Bean told IMM that 30 percent of CADKey users are moldmakers. By the end of January, the company will release CADKey '97, with integrated ACIS modeler and wireframe-to-solids converter. "This will allow moldmakers to bury a solid model in a block and subtract it to produce core and cavity sets," says Bean. Surfaces can then be output via IGES. Baystate is committed to upgrading CADKey with a new release every six months, he notes.

Intergraph's Solid Edge, a Windows-native solid CAD package with parametric part modeling, will feature open profiles in its newest version 2. This means part features will be connected to, and can exchange information with, adjacent features. In addition, the program will offer more intelligence in geometry creation. It will know, for example, that a rib follows across multiple faces and features. An assembly mode lets users create new parts and reference the geometry of existing ones.

Ansys/AutoFEA 3D embedded within Autodesk Mechanical Desktop offers tighter integration between design and analysis.

Analyzing Designs
Moldflow and Autodesk have integrated MF/Flowcheck within Mechanical Desktop for seamless flow-analysis capabilities, according to Moldflow's Paul Bordonaro. "This gives product designers a 'will-it-fill' guideline," he notes. Analysis results will show fill, flow pattern, and location of possible weld lines and air traps. Each node will be assigned a confidence factor based on color—blue means the part will fill under normal injection pressure, green means it may fill, and red means it will not fill. When two or more colors are present, the software determines an answer based on the dominant color and importance of its location.

Another Desktop solution partner, Ansys, has integrated AutoFEA 3D for early, upfront design validation. "Users can answer the questions 'Will this design work?' and 'Will it perform when forces are applied?'" notes Dick Miller, vice president/general manager of Ansys Design-Space. By embedding the system, he adds, Ansys ensures that validations are performed directly on the CAD model for data integrity. AutoFEA 3d runs in AutoCAD r13 or Mechanical Desktop 1.1 under Windows95 and NT platforms.

Engineering Animation Inc. introduced VisMockUp for digital prototyping of large assemblies. The product is due for release in the first quarter of this year.

Moldmaking News
Cimatron Technologies Moldesign module, integrated within the Cimatron CAD/CAM product, lets moldmakers read in a design file, perform detailed mold development, and output a precise NC milling file, all within the same environment. Peter Bolger, vice president, tells IMM that 80 percent of Cimatron's customer base is involved in moldmaking. "These are people who need true CAM integration with elite solids and surfaces," Bolger says, "and these are the areas where Cimatron excels." The package is open to all platforms, PC and workstation alike.

In version 8.0, developers created functions just for moldmakers, such as knowledge-based NC. This means the software automatically knows where the unmachined areas are. Also, users can build libraries of machining strategies for similar parts so that they don't have to continually reinvent the wheel, Bolger notes. Cimatron also offers hybrid modeling—rather than converting the entire IGES wireframe file, users can just pick a reference point or one surface for conversion to parametric solids. "This is especially time saving for fillets and corner blends, where you can literally save hours," he says.

Step-to-IGES converter
International TechneGroup Inc. displayed a Step-to-IGES converter for importing Step-based CAD files into IGES-based CAM programs.

Cimlinc is focusing its efforts on helping automotive moldmakers re-engineer their process for growth. Jeanné Naysmith, vice president, believes the company's Prospector package is the first process optimization software to address all mold machining issues. "It not only verifies paths created by the CAM system, it also allows operators to quickly optimize those paths on the shop floor or create new toolpaths to take advantage of objective conditions," says Naysmith. Prospector runs under Windows NT and uses a knowledge base to let operators determine the best tool, tool holder, and adapter for each mold. The system also maintains associativity between toolpaths and the surface models contained in the part's CAD file.

Reverse Engineering
Mold repair via reverse engineering just got simpler. Imageware's Surfacer links digitized model data from a variety of sources (touch probes to CMM equipment) to existing CAD/ CAM tools. Surfacer will get a facelift of sorts as the company adds on a new module to speed the time it takes to create surfaces. The new approach simplifies the process, allowing someone who isn't a Nurbs "guru" to quickly create complex, multiple-surface Nurbs models, says founder Kurt Skifstad.

Tooling for encapsulated molding

Figure 1
Figure 1. Proposed mold modifications to uniformly encapsulate a metal ring in one shot. Pneumatics or hydraulics can be used for pin retraction.
Earlier this fall Paul White, product development engineer for research and development at Finish Thompson Inc. (Erie, PA), posted this challenge on the Internet at the Plastics Network and in this magazine (October 1996 IMM, p. 117): "I need to totally encapsulate a steel ring with plastic in one shot so that there will be no seams where acid can get to the metal ring. The ring must also be centered in the molded part no matter which way you look at the part."

Finish Thompson manufactures, among other things, industrial pumps, the kind that pumps highly corrosive chemicals or acids used often in cleaning, plating, semiconductor, and photo processing applications. The company has two of its own presses, one 225 tons, the other 650 tons. The 4-inch metal ring mentioned by White is a pole piece for a set of rare-earth magnets. The magnets are made of Neodymium, a relatively rare and rarely used 34-karat inert metal you can find at the bottom of the periodic table. This magnet is molded into an impeller hub, an important part in the company's line of seal-less pumps.

The impeller hub comes in direct contact with the chemicals that pass through the pump, and protection from corrosion is paramount to the pump's long life. To protect the ring Finish Thompson currently encapsulates it in 40 percent glass-filled polypropylene, or 15 percent fiber-filled polyvinylidene fluoride (PVDF) in a two-shot process. The first shot encapsulates about three-quarters of the ring; the second shot finishes the last quarter.

The problem with the two-shot process, White says, is that the ring is not always completely sealed. The effect is domino-like: Chemicals violate the seal, the chemicals corrode the ring, the plastic encapsulating the ring swells, the plastic rubs against the barrier surrounding the impeller hub, the barrier is destroyed, the pump leaks, the pump fails. The failure rate is less than 2 percent, but any failure is unacceptable, White says.

One-shot Molding
White wants to encapsulate the ring in one shot to ensure that the ring remains sealed. Responses to White's query all followed the same vein, with some slight variations in technique. The mold design White wants to use is pictured (Figure 1) and illustrates the concept.

Four spring-loaded pins, located every 90°, hold the magnet ring in the mold. The spring pressure is countered by a pneumatic cylinder to maintain pin position. During injection, as the mold fills and molten plastic starts to enclose the ring, cavity pressure will trigger a relief valve in the pneumatic cylinders, retracting the pins. White expects that the melt already in the mold will be stiff enough to keep the ring secure, but still fluid enough to fill in the holes left behind by the retracted pins.

Figure 2
Figure 2. Encapsulation of a cellular antennae coil involves a two-shot process.

There are several ways to trigger pin retraction: via a timing mechanism, cavity pressure, machine pressure, or screw position. "I'd rather not have it be a function of time," White says, indicating his preference for cavity pressure retraction. "I'd rather have the plastic itself do the job."

There are also several ways to actuate the pins: hydraulically, pneumatically, with spring action, or a combination, as White chose. White says he thinks the tooling modifications will cost his company as much as $5000. "But," he says "we're saving money by not losing pumps. So it should more than pay for itself."

Two-shot Molding
Although the two-shot process does not work all the time for White, that method still carries merit for many applications. Alex Kondor is senior project engineer at Advance Dial, based in Elmhurst, IL. He was one of the respondents to White's challenge and says the two-shot process works quite well for one of his products, a cellular telephone antenna.

Advance Dial is a custom molder operating four different facilities. The Elmhurst plant in which Kondor works has 33 machines ranging from 20 to 170 tons molding parts for consumer electronics and auto underhood and interior applications. Kondor says about a third of his facility's presses are rotary insert machines.

The heart of the cellular antenna is a gold coil that is fully encapsulated in a thermoplastic or a thermoplastic elastomer. A coil such as this one is not easily positioned in a mold with pins. Kondor says the best option is a two-shot process. In the first shot the "core" is filled and the coil itself is partially encapsulated (Figure 2). The idea is to attach to the coil a series of wart-like positioning posts. These are used in the second shot to position the coil evenly in the mold where it is completely and uniformly encapsulated.

Kondor says both shots are done on one two-cavity mold. In the first run the second cavity is blocked and a series of gold coils is given the first shot. Then the first cavity is blocked and the same coils are given the second shot. Kondor has this advice for molders who are contemplating molding in two shots: "I think the mistake most people make with the two-shot approach is they try to encapsulate as much as possible with the first shot, then finish it off with the second. Instead, use the first shot to mold position posts for the second shot, which does most of the encapsulating."

One-shot encapsulation, however, is also performed at Advance Dial, Kondor says. One of the parts is a printed circuit board for automotive use. The edges of the board must be completely encapsulated to protect it in the relatively harsh underhood environment. Unlike White's plan at Finish Thompson, for this application Advance Dial performs pin retraction based on timing, and actuates the retraction hydraulically. Kondor says the rule of thumb at Advance Dial is to start pin retraction when the cavity is about half full. He says the tricky part is to not let melt seep into gaps around the pins; this can be alleviated by careful pin placement.

Advice on Encapsulation
IMM ran all of this by a man whose career is encapsulation. He is Tom Boyer, a representative for development technical service at DuPont Engineering Polymers in Wilmington, DE. Boyer's job is encapsulation R&D. He's cut his teeth encapsulating everything from golf balls to antilock brake system wheel speed sensors. His thoughts and advice on the subject follow.

One-shot encapsulation. Almost every encapsulation project can be best molded in one shot. Boyer strongly recommends pins to position and hold the part in the mold. Even coiled parts like the cellular antenna at Advance Dial, he says, may be able to be held with pins by treating the coil with a bonding agent that hardens the material, making it stiff enough to withstand the force exerted by the pins.

Trigger pin retraction. Boyer says he mostly uses cavity pressure or screw position to trigger pin retraction. On a well-tuned machine, cavity pressure is the most reliable. Other machine components can vary in performance, but cavity pressure is the truest indicator of the status in the mold, where the part sits. Boyer also uses screw position. When the screw reaches the cushion, he assumes the part is almost fully packed and pulls the pins; typically the melt is fluid enough to fill the pin holes, but stiff enough to maintain part position and uniformity. In any case, Boyer likes to see the pins pulled as late as possible in the process. Of Advance Dial's strategy of pulling halfway through injection, he says he would be concerned that the part could shift in the mold, causing nonuniform melt distribution. But he says each application should be treated separately. Later pin retraction may minimize nonuniform melt distribution.

Retract with hydraulics. Boyer says pneumatics work but do not provide the precision and predictability inherent in hydraulic systems. He estimates that on a typical mold, adding the pins should cost about $5000 to $10,000, as White indicated. Boyer says the primary parts required are four hydraulic cylinders (pancake cylinders take up the least space), two on top of the mold, the other two on the bottom; these are attached to two plates with the pins attached. Finally, if you are using cavity pressure to trigger pin retraction, install a cavity pressure transducer.

Pin size. If the pins are large and the part is relatively small, the holes left by the pins after retraction may alter flow fronts, creating voids, leak paths, or structural weaknesses. "The smaller the pin, the more likely you are to pack that part out," he says.

Material. Nylon 6/12 is a favorite of Boyer's because its flow properties make it ideal for encapsulation; it's also strong and relatively resistant to corrosive chemicals. No matter what you choose, if you are encapsulating a metal part, Boyer says to be aware that the melt can affect the inductance of the part. High melt temperature and the injection pressure are the two biggest culprits of inductance change.

Get a vertical press. Boyer says anyone who is serious about encapsulation should invest in a vertical injection molding machine. He points out that when encapsulating on a horizontal machine, gravity is your worst enemy. It can lead to nonuniform distribution of plastic around the part. Also, a horizontal machine generally restricts you to a one-cavity tool as the operator cannot position more than one part in the mold at a time.

Boyer recommends a shuttle or rotary vertical press. This provides four main advantages, each of which saves time and money:

  • You are restricted to one top half of the mold, but the shuttle or rotary system allows multiple bottom halves of the mold. While parts are being injected and encapsulated on the first mold the operator can remove parts from the second and prep it for injection. Says Boyer, "This is the optimal use of time and material."
  • Tooling costs are reduced. Again, you have one top half of the mold, but multiple bottom halves. Every extra bottom half is money saved by not having to machine a top half for it.
  • Because the molds lie flat, multiple cavities are easier to handle. The operator lays parts for encapsulation in the mold, instead of trying to position the part vertically in a horizontal press.
  • Because the vertical press injects down into the mold, gravity is an advantage. Parts are more likely to be uniformly encapsulated.

For molders who balk at the cost of a vertical press, Boyer says, "The cost savings seen by investing in a vertical machine would more than justify the cost of the machine."

Don't be afraid.
"Encapsulating is not black magic," Boyer says. The most common myth he hears comes from molders who traditionally use epoxy and thermosets to encapsulate parts. He says they're under the false impression that pressures introduced by thermoplastic injection molding are too high. He points out that an injection molding machine can easily be adjusted to mold a good part at reduced pressures with a slower injection velocity. There are few reasons why a part cannot be encapsulated in thermoplastic materials.

Sound Off: Let's design hot runners into the molding machine

Matt Lofgren
Matt Lofgren, Kendall Healthcare
Editor's Note: Matt Lofgren is a molding supervisor at Kendall Healthcare Products Co. in Ocala, FL. This article is his description of a design for a new type of injection molding machine. Read it over, give us your feedback. Is it old hat? Does it make sense? What are the strengths of his design? What are the weaknesses? Does it have practical applications today? Could it work?

We may be at a crossroads in injection molding machine technology. Since the advent of the reciprocating screw to replace the original ram or plunger design, there have been few technological leaps in the machines with which we process today vs. the machinery molders ran 20 or 30 years ago. We have seen incremental advances in equipment such as microprocessor control, all- electric designs, and tiebarless machines, but the basic design has remained the same.

Hot runner systems compensate for some molding machine deficiencies, and we rely on them to the point where they are nearly an extension of our machines. Unfortunately, this expensive, complex system is useful with only one mold. Want to install another mold that benefits from the features a hot runner system offers? You must design, purchase, and fabricate that next mold around another hot runner unit.

It seems like we are reinventing the molding machine every time we see a new iteration of the latest hot runner system. Granted, hot runner technology has come a long way in the last 10 or 15 years, to the point that many molders swear by it. Hot runners serve many molders quite well; our molding department in particular runs several hot runner molds, for the most part successfully. But here are some of my observations of the drawbacks of hot runners:

  • Hot runner systems are expensive. A typical eight-cavity hot runner manifold system can add $20,000 to $50,000, or more, to the cost of a mold. Even the most rudimentary systems start at more than $10,000.
  • Hot runner tools are quirky. A molder can run into trouble whether starting up, running, shutting down, or cleaning hot runner systems. They are complex systems, particularly valve-gated hot runner molds.
  • Hot runner molds tend to take longer to install and remove. Many shops around the country have gone to or are considering going to a JIT molding environment. While this approach certainly looks good on paper, it puts extra burdens on the molder as far as quick and efficient changeovers go.
  • Hot runners are not suitable for certain materials. I don't know many molders who run 45 percent glass-filled PET through hot runners. I'm sure it's been done, but such systems would have to be fabricated from exotic steels to withstand abrasives in filled materials.

Hot runner systems do a great job if they are properly designed and used for what they were intended. Their inherent advantages make it easy to justify the expense and hassle of these molds. What I'm advocating is to take a new approach: Let's put the money we're spending on hot runner systems into the injection molding machine itself.

Figure 1
Figure 1. The big difference in this machine is six small, but conventional injection units, three high and two wide in a tight pattern in front of the fixed platen.

The Idea: "MultiJect"
I propose a new design for the injection molding machine. The machine is a multi-injection unit with an integrated, standardized mold base that I have dubbed the "MultiJect" injection molding machine.

The "MultiJect" looks like this: Start with a standard 250-ton base injection molding machine, toggle or hydraulic. The machine remains as is from the movable platen on back. The big difference in this machine is six small, but conventional injection units, three high and two wide in a tight pattern in front of the fixed platen (Figure 1). This configuration is variable; it could be a four- or eight-injection unit system or other clamp tonnage of the base machine. The point is, a separate unit injects each cavity of a mold.

The mold itself is also unique. The customary center sprue bushing and runner system are obsolete, replaced by either direct injection of material into each individual part, or a small runner system gating to two (or more) cavities. The result is a molding machine that has all or most of the advantages of a hot runner system built into the machine, with none of the inherent disadvantages.

The individual injection units are independently controlled by their own hydraulic (or electric) circuits to allow for separate control of all normal injection unit functions. Each injection unit is also designed so it swings out independently of the other injection units to allow for maintenance, even while the other units are producing.

The Mold
The mold is unlike anything to which we are accustomed. Instead of a locator ring surrounding the sprue bushing used today, I invision several locator or guide pins to properly align the mold into the fixed platen, automatically squaring up the mold. For the purposes of this description, assume a six-injection unit design. Six small, individual sprue bushings are located to coincide with the tips of the injection units. With extended nozzles we inject directly into the part, or into a simple runner where no sprue would be attached after ejection of the part/runner.

All mold features are still used, including tunnel gating, three-plate mold systems, unscrewing molds, water lines, and others. Just think of it as six individual mold details built into a larger mold base. The mold is clamped to the platens conventionally, or by using Arburg-style bolts through the platens.

This design mimics six small-tonnage injection molding machines running one- or two-cavity tools. It eliminates the expense of six separate presses and operates in the same footprint as a standard 250-ton machine. The biggest benefit is full control of each individual cavity (or two cavities if running two per injection unit). A molder could run a six- or 12-cavity tool with the efficiencies of a hot runner system without the per-tool expense.

The Benefits

  • Single-cavity control of injection pressures, velocities, and times. This design injects fresh melt into the cavity each time, not melt that's been degrading in hot runners. It also eliminates problems of flash or shorting.
  • Standardized tooling. There is no investment in individual hot runner systems for each mold; we invest in the machine up front and use standard tools to fit the machine. Smaller tools, e.g., less than six cavities in this case, could also be used as long as the tooling is designed for this type of press. Though a single-cavity tool may not be cost-effective using this design, bolsters such as those used to support the molding plates during high clamping pressures could be installed across the face of the platen to prevent tiebar strain should this configuration be required.
  • The ability to purge or perform screw and barrel maintenance on individual barrels while still running other cavities. How much more efficient would your operation be if you only had to lose one or two cavities while changing out a defective heater band while your press kept running?
  • Faster cycle times. As with a hot runner system, we have little if any runner to cool.
  • The capability to shut off single cavities, but with no ensuing balancing problems; nor would a molder have to adjust the shot size after dropping a cavity.
  • The ability to run different color materials in the same mold, on the same shot cycle. This design can mold six different colors simultaneously, eliminating the need to run separate colors on separate machines and mix the product before packaging.
  • Family molds would be much more practical. It's difficult to get good fill on two different sized cavities on the same tool, as most of the family molds are today. With the "MultiJect" system, simply program in separate shot sizes and process parameters for each individual injection unit or cavity combination.
  • No material limitations. Unlike most hot runner systems, any material compatible with today's machines can be molded. Each injection unit could even run a different material if needed.
  • Component parts for the injection units themselves are less expensive than the cost of a comparable, conventional full-sized injection unit. Components—screw, tip, barrel—are standard off-the-shelf parts for smaller conventional presses.
  • Better control of the heat profile. Material runs in a small barrel to feed a single cavity, with a small runner.
  • Balanced pressure distribution. The mold itself is more stable as pressures are evenly divided throughout the tool. The "MultiJect" system practically eliminates unequal flexing of the tool, easing the strain on tiebars at the same time.

The purchase price of the machine would be higher; I estimate about twice what we pay for a machine today. This would be a one-time purchase though. Our controller technology is up to the task of individually controlling each injection unit and displaying those results on a CRT. The hydraulic circuits would be complex, but most of the circuits would be a duplication of what exists in most machines today.

The tooling would be unique to the machine. The mold base industry and mold designers would have to be sold on the concept. Unfortunately, conventional tooling cannot run

on this type of machine, unless adapters are designed to accommodate them. The tool in my design is offset from the centerline of the platen, and there are limits imposed by the small volume of the injection unit itself. Given the large base of existing tooling, this is the biggest stumbling block (aside from designing and building the machine itself).

Where to Now?
One can see that my mythical molding machine has numerous advantages over the traditional molding machine and its hot runner counterpart. There are disadvantages, but to me, they do not seem insurmountable.

Is this a concept that could catch on in the molding community? Would there be enough support for this concept to have a company take the risks involved to see the "MultiJect" system to its conclusion? I don't have the answers to these questions, but I would like to have some feedback (pro and con) from anyone who has kept an open mind to this point.

Has it been done already?
Matt Lofgren proposes a design for a new type of injection molding machine, but it's not really new. Battenfeld produced a four-barrel injection molding machine in the early '70s. Battenfeld, according to one of the machine's owners, apparently only ever made four such presses before it stopped producing them. According to Battenfeld records, the machine was given the model number PFG 400/4X 3000.

At least one of those presses is still in operation today. The machine started life with Playskool in the early '70s before it was purchased in the mid '70s by now-defunct Keolyn Plastics in Chicago. Jack Glatt, formerly with Keolyn, says the 440-ton press molded structural foam in four-cavity and single-cavity molds with shots up to 60 lb.

In the early '80s Keolyn sold the press to FM Plastics in Rogers, AR. Dave Shallenberg, manager of manufacturing services at FM Plastics, says the four-barrel machine is one of his most reliable presses. "It's become a real workhorse," he says. FM Plastics molds products for computer, business, and medical industry applications using polystyrene, Noryl, and PP. Shallenberg says he uses some single-cavity molds on the machine, but most of the work is done using multiple cavities with shots in the 4- to 7-lb range.

Shallenberg says he's rebuilt much of the machine, adding a controller for each barrel, replacing the original pair of controllers that monitored two barrels at a time.

One of the advantages Lofgren describes in his design idea is the ability to perform maintenance on one barrel or screw while the others continue to run. Shallenberg at FM Plastics says the compact design of the four barrels on the Battenfeld, combined with mold use, prevents him from performing such maintenance.

Snapfit software closes the loop

Of the many reasons designers choose molded plastic, part consolidation is perhaps at the top of the list. And falling under this category are integrally molded fasteners, also known as snapfits.

Designing the various types of snapfits to adequately withstand stress, however, can be tricky due to the nonlinear nature of plastics in which stiffness changes with applied loads. Does this mean you need to perform nonlinear finite-element analysis every time you design a snapfit, or is there another solution?

Cantilever snap tab designAnnular snapfits
Figure 1. Developers predict that this will be the most popular use for Snap Design—namely, to design cantilever snap tabs. The left window lets users input parameters and shows calculation results, while the right window shows the 3-D view at any angle specified.Figure 2. Annular snapfits, commonly seen in pen caps and bottle closures, are analyzed with red used to represent the elastic element and blue to model the rigid one.

Dennis Que of Closed Loop Solutions (Troy, MI) tells IMM that the answer for snapfits is not so black-and-white. "We recommend using closed form equations widely, backed up by nonlinear FEA when required," says Que. "In fact, designers often learn what's needed during FEA by applying closed loop solutions first." Furthermore, he adds, to avoid overstressed designs, each snapfit feature should be analyzed in some way.

Torsional snapfits
Figure 3. Torsional snapfit features are not common, but can be used for releasable drawer stops molded in the line of draw.
Snap Design software
Figure 4. In addition to its generic materials database of common thermoplastics, Snap Design lets users add their own material data. A nonlinear material model is used to accurately predict the stiffness of the snapfit.

Que's newly formed company has developed a software module called Snap Design, which is based on closed form equations and parametric geometries. With it, users can bypass some of the iterations that FEA requires. To use FEA, you need to build a model; apply boundary conditions, loads, or displacements; specify material properties; and then wait for the computer to solve for deflections and stresses.

These values must be compared to the desired values, and inputs must be adjusted iteratively until all of them are acceptable. With Snap Design, if you know any two of the three inputs—loads, material, or dimensions—you can solve for the third unknown without building a model or adjusting inputs.

This isn't the first time plastics designers have seen software and guidelines for easing the task of snapfit design, Que admits. AlliedSignal, Bayer, BASF, GE Plastics, and Dow have all released their own versions in the past. But the new package builds on the existing knowledge base, and relies on some heavy-duty contributing technical advisors, including Paul Tres of Montell, a noted plastics design authority.

Accurate results in a short span of time appear to be this software's strong point. For example, using the software to solve for a material, users select the snap type—cantilever, annular, or torsional as in Figures 1 to 3—then enter dimensions, allowable deflection, and deflection force. After crunching numbers through the closed form equations, materials that satisfy the inputs are displayed in a summary window. To narrow the materials choices to a manageable size, users can fill in known criteria such as the material's generic name or operating temperature range. The package also includes an integral materials database with tensile stress-strain curves; users can expand the database easily with their own material properties, as seen in Figure 4.

Snap Design sells for $349. The package runs on Windows95, NT, 3.0 or 3.1.