Plastic injection molding is possibly the most widely used, yet least thoroughly understood manufacturing technologies in use today. Injection molding of thermoplastics began in the 1930s, long before the scientific study or understanding of the behavior of polymer melting under pressure. It was an empirical, trial by error industry, and the critical element of low-cost, high-value part making—the mold—was designed by engineers who learned more by experience than from textbooks. It was something of a black art.
Today, everything is different. The combination of good mathematical models of the rheology of resin melts, a better understanding of metallurgy and heat transfer, and the formalization of years of “rules of thumb” has allowed specialist engineers to train specifically in plastic mold engineering.
Iterative mold development, however, is still very much a tooling strategy. The difference today is that this is done virtually with simulation software. But why? Modelling of fluid flow in closed channels has been done for decades. The reason is in the resins, says Caitlin Tschappat, Moldflow technical specialist with Autodesk and a specialist polymer engineer.
“People don’t realize how complex plastics are and how they don’t behave like metals. They flow by non-Newtonian principles, with very different flow properties. For example, they’re highly shear sensitive. It’s complex. Many people look at a simple part and think ‘oh, that’ll be easy to fill.’ That simple shape could be one of the hardest parts to manufacture because of many considerations, such as warpage,” she explains.
Even Simple Parts Can Require Complex Mold Design
At its simplest, injection molding is about orienting cavities in three-dimensional space with a parting line that allows free ejection of the cooled, solid resin parts. Simply determining the location of the parting line can be nontrivial. Parts with zero or negative draft angles may be impossible to eject with conventional ejector pins, forcing a designer to use fewer cavities, or nonoptimal cavity orientation in a mold to get clean ejection. In some severe cases, there is no way to orient the part to facilitate ejection, and core pulls must be used, adding complexity and cost.
The complexity doesn’t end there. Coinjection, over molding—especially of TPU’s over commodity thermoplastics—and in-mold decorating all complicate the mold design process.
If very high-volume production is necessary, such as in the packaging industry, stack molds may require complex designs with multiple parting lines and a need for fast, clean ejection. Symmetry helps, and a 64 or 128-cavity small part mold may be a simple matter of design recursion. A family mold, however, or parts that require special features such as the popular “living hinge,” can be very difficult to design. There are also multiple other issues involving gates, runners and other essential mold features.
The advantages of simulation are obvious. Tschappat is an industry veteran with experience in the packaging and automotive industries and has seen this complexity up close.
“Think about a larger part,” she says. “The automotive industry, for example, has many long, thin parts. For these, you must consider the aspect ratio with respect to part length to wall thickness, as it would be difficult to fill uniformly without the inclusion of complex runner systems, hot drops and valve gates. On the other hand, small parts, such as those for medical or electronic applications, may also require the same complexity with respect to the runner systems, as they too can be difficult to fill due to small features and limited filling pressures.”
Modern mold design partially addresses the complexity issue by the use of off-the-shelf mold bases, inserts, gates and other standardized components wherever possible. Advanced simulation software such as Autodesk Moldflow works with these components to allow accurate approximations of mold performance in gating and cavity filling.
Cavity balance is always a high priority when filling multi-cavity molds, which can sometimes be remedied through the use of mold cavity symmetry. These instances are also able to simplify mold simulations, such as how other Finite Element Analyses use symmetry for model simplification.
For family molds, or large single cavity molds with complex shapes, simulation makes the difference between a productive and cost-effective tool, and a design that is revised so much it “goes through the alphabet.”
Before simulation became more widespread, it was not uncommon to alter important mold components on the fly, such as quick-fixes to gates for improving mold balance. Unfortunately, these are just as it says—quick-fixes—leading to effects on other aspects such as part quality like jetting, knit lines and even dimensional instability. That kind of experimentation may solve a problem, but it frequently requires a total rethink of the overall molding strategy, with new machine parameters that may require hundreds of shots to perfect.
The ability of simulation to minimize rework not only reduces time spent on the mold, but also the learning curve on machine set up with a new job. For shops with a captive press operation, the savings for reduced downtime and improved machine scheduling ability are obvious, but for mold shops there are additional benefits. Rework costs and delivery delays are reduced, customer satisfaction is improved, and the too frequent finger-pointing (who pays for that modification?) can be greatly minimized.
The Solution to Complexity Is Simulation
Complexity is a given with modern injection molding and many jobs simply can’t be attempted without advanced simulation. Tschappat describes how she uses Moldflow to address higher level problems, saying, “We have many modules in the software depending on what questions you need answered, whether it's co-injection, two-shot, over-molding or insert molding. For example, with gas assist, we can help identify where the gas void will settle within the cavity, or with coinjection how two plastics are going to bond together in the analysis.”
Simulation helps address the previously mentioned topic of cavity balance for multi-cavity tools.
"We see so many tool cavity layouts when talking with our simulation customers. But the one thing that always seems to throw people off is if they can get away with molding quality parts from family tools, where each cavity may be a different part, and technically would need different processing conditions,” says Tschappat.
“Using Moldflow simulation, we model up the mold layout and predict the part quality of the different cavities. Then we experiment with artificially balancing the filling through the use of changing runner diameters between cavities or changing the runner design, to attempt a more uniform filling from cavity to cavity. Even if the filling is balanced, other factors such as shear-induced imbalances can occur as a result of the different geometry features. This is where simulation is really cool—seeing something that we can't even see with our eyes when actually molding the parts at the press," she adds.
The types of runners are another source of complexity. Hot runners are standard for volume production due to zero or minimal wasted material, but cold runners allow for re-purpose of their runner material toward regrind to add to material savings through reusing small percentages. But as Tschappat observes, “based on your part, what is the best gating scheme? Depending on how big your part is, maybe you're using a fan gate, or if it's a smaller part it might be a small pin gate. Then the pressures and the pressure drops are a factor throughout the runner system. These are all things that make it complex, yet you have to weigh them out at the end of the day to determine what is best for your business.”
Even a seemingly simple change in a commodity resin can introduce issues. “They teach you in school that if you are going to be using a different resin, particularly very dissimilar materials to those that you’re used to, you should build your mold to that material,” Tschappat says. “We all know that that's not necessarily the practice. You may start out with a non-filled material and then decide to switch to a glass-filled material after manufacturing a few shots, for example. Now you have to worry about more abrasive wear on the tool over time. What do you do in those situations? More frequent tool inspections and reworks, maybe even welding up the gate and re-cutting it so it allows for less shear of the fibrous material. These sorts of things just take extra time out of the process and slow you down, but it's common practice.”
The ability to rerun a simulation with a new material virtually can flag a molder about potential problems before they translate to expensive tool rework and downtime. Warpage is another common reason for modifications to the tool design. A good designer who has familiarity with simulation packages such as Moldflow can gain insight into root causes of these problems.
“Once you really get comfortable, you can actually isolate causes of the warpage to better understand why something is warping,” states Tschappat. “For example, if you're getting a lot of warpage due to a thicker cross-section because your material's not freezing off and you have a lot of shrinkage in that area, Moldflow allows you to visualize that and then make a design change and rerun an analysis to see how that design change reduces overall warpage. This is why we are working more and more with part designers. They can use simulation as they design the part to flag these problem areas before escalating to the tool designer, leading to fewer iterating between the two, making them look like they are superstars!”
The cooling issue is critical and is often more difficult to perfect than the cavity and runner design. Although 3D printing promises truly conformal cooling, most production molds are cooled by drilled and manifolded channels carrying coolant, usually water or oil, which carries the heat away by the thermalator.
There’s an old rule of thumb that for efficient cooling, set the machine to eject the part at 80 percent of the part’s heat distortion temperature. However, for complex parts, multiple cavities or complexity added by factors such as thermoplastic elastomers, coinjection or gas assist, rules of thumb are quickly replaced with empirically derived settings during the mold runoff. Cooling is usually the determining factor in overall cycle time, so here time very much is money.
Simulation of cooling channel layout and flow capability can be equally or more important than efficient cavity filling for a high capacity mold, and in the world of injection molding, very few molds are not thought of as high capacity, meaning cooling is almost always a critical factor.
Simulation Helps Even Before the Design Process
Mold designers are frequently faced with customer requirements that may be difficult or even impossible to achieve. Most production shops understand their press plate size and tonnage, chiller capacity and target cycle time, but know little about the mold.
Customer expectations can sometimes be unrealistic. “Part designers and mold engineers have to work with one another,” says Tschappat. “Time is money, and everything needs to be finished yesterday. It goes through several phases. How are you going to lay your model out? You have to consider where you're going to inject resin and locate your gates. And then what type of gate you're going to use. What type of runner systems? Is it going to be a hot manifold or are cold runners going to be sufficient? Or is it a hot to cold runner? What's the ejection unit look like? Where can you squeeze cooling in?”
“Often in my job, you get parts from customers and you ask them, ‘what's your cooling layout look like?’ You may get an answer to that question, or they may not know. Often a designer will just squeeze cooling in there wherever they have extra room left over,” Tschappat adds.
Even the capability of the machines that run the mold can be assessed critically with mold simulation. Simulation allows a mold designer to run “what if” scenarios that may show a customer where expectations are unrealistic and nudge them to a better mold design without an argument.
Math is definitive. “If you're limited in the types or how large your injection molding machines are, then you're limited to the size of the mold that you can put in that unit,” Tschappat says. “And then of course, how much pressure it is going to take to fill these parts? Is the machine large enough to not just fill out the cavities, but hold the pressures needed to pack those parts? If it's not, you're going to have some problems and you're either going to have to redesign the tool or reconsider it. Or buy a new machine.”
This isn’t a theoretical consideration. Tschappat has seen customers who were spared the cost of a new machine when shown the benefits of a better optimized mold design.
Simulation and the Tooling Designer Work Hand-in-Hand
Does simulation replace the tooling designer?
“No,” declares Tschappat. “I would say there's still an art to it. When I'm in conversation, I regularly tell people they should be using simulation as an extra data point or another data point, and use their experience. This actually ties into that aging workforce, too, as we're starting to lose that skillset. People are retiring, so how do we build that experience base up? Simulation is a really good way to help. Young people need to be sponges; if you're a toolmaker or a tool designer, work closely with the older generation to pick up some of their skills.”
Simulation software for injection mold design has progressed from “nice to have” to a “must have” for cost effective tooling. Autodesk’s Caitlin Tschappat is an enthusiastic Moldflow advocate.
“I think it did wonders for the industry because you're able to build an ROI case and see where you can improve cost savings at the end of the day. Everybody wants a more complex part cheaper, faster, quicker, and by incorporating this sort of technology into your workflows and work processes, we can help achieve that goal for our customers.”
For more information about Autodesk Moldflow products, click here.