 |
New Frontiers in |
||||||
|
|
|
High-Conductivity Materials & Conformal Cooling Channels |
|
Why are almost all production plastic injection molding core and cavity inserts made from steel? Steel is certainly not the hardest material. Steel is also definitely not the most corrosion-resistant material. And steel most assuredly is not the best conductor of heat. So why are production tools made from steel? The real answer is probably because that's the way they have been made for many years. People are used to working with steel molds. The results are predictable. However, there are now commercially available inserts that achieve: (1) shorter lead times, (2) faster cycle times, (3) enhanced productivity, (4) lower unit part cost, and (5) reduced part distortion. A mold is effectively a thermodynamic engine. The analogy with an automobile engine is relevant. The well-known internal combustion engine sequence of "intake, compression, power, and exhaust" can be extended to plastic injection molding. In a motor vehicle, gasoline or diesel fuel is injected into the cylinder. In an injection mold, hot plastic is injected into the mold. In a gasoline engine, the fuel-air mixture is compressed to about 120 to 135 psi. In a diesel engine, the pressure reaches about 300 to 330 psi. In injection molding, the molten plastic typically is injected at pressures ranging from 5,000 psi to as high as 20,000 psi. The power in a car is derived from chemical energy stored in the fuel, which is transformed into thermal energy during combustion, and ultimately converted, through the drive train, into kinetic energy. In injection molding, the thermal energy of the heated, molten plastic must be removed in order for the plastic to cool and transform into a solid. Finally, in a car, the products of combustion must be exhausted from the engine. In a mold, the solidified plastic must be ejected from the press. In a passenger car, this cycle may occur from 800 to about 5,000 times per minute. In a Le Mans-winning race car, this cycle can occur as many as 17,000 times per minute. In an injection mold, the entire thermodynamic cycle may occur as infrequently as 0.5 to perhaps five times per minute. The name of the game in auto racing is high engine revolutions per minute (RPM) coupled with high component reliability and success negotiating the race course without problems. The name of the game in plastic injection molding is really the same: high PPM (i.e. parts per minute), long tool life, and few distorted or rejected parts. While considerable efforts have gone into assuring that injection mold tools are functional and will provide long tool life, far less attention has been devoted to the fundamental issue of their thermal management.
High Thermal-Conductivity Materials Figure 1 plots thermal conductivity for some relevant materials. Heat transferred from the plastic must be conducted through the material of the mold before it can be removed by coolant. Thus, the thermal conductivity of the mold material is critical to mold thermal management. Inspection of this figure immediately reveals one of the basic problems with steel. Here H-13 tool steel, having a thermal conductivity of 28 W/moK, was chosen to be representative of the broad class of "tool steels." As a point of interest, 316 stainless steel is even less thermally conductive, at about 20 W/moK. By comparison, copper at 390 W/moK is about 13 times as conductive as H-13 steel, and almost 20 times as conductive as 316 stainless steel. While pure copper is too soft to provide long tool life at the active mold surface, it is a terrific material for thermal management. Next, aluminum possesses roughly half the thermal conductivity of copper, but is also too soft for long tool life. Nonetheless, as it is easily machined, aluminum is often used for prototype or bridge tooling applications, requiring a few hundred to perhaps as many as 50,000 parts injection-molded in the desired engineering thermoplastic. If glass-filled plastics are required, aluminum tool life will be further reduced. On the other hand, nickel has a thermal conductivity of 88 W/moK-- more than triple that of H-13, and quadruple that of 316 stainless. Furthermore, nickel is very corrosion resistant, polishes well, is relatively hard, is abrasion resistant, can be textured, and provides excellent release characteristics. Combining a 2-mm-thick nickel shell at the active mold surface with a 4-mm-thick copper thermal management layer that encapsulates the conformal cooling channels can provide dramatic benefits. The resulting Ni-Cu composite has an effective thermal conductivity some seven to nine times that of conventional steel tools, and is capable of generating production part quantities. Conventional steel tools generally are CNC machined or electrical discharge machined (EDM) from a solid block of tool steel. Consequently, the cooling channels also must be drilled into solid steel. As a result, these channels essentially consist of a series of interconnecting straight segments, each having a circular cross-section. This operation inevitably results in two important limitations. First, because the cooling channels are "gun-barrel drilled," they cannot be made to conform to the curved shapes typical of injection molded plastic parts. The result is that some regions of the plastic are better cooled than other regions. The cooler plastic regions reach their solidification point earlier than the hotter regions. When the cooler regions solidify, they shrink. Somewhat later, when the hotter regions finally have cooled sufficiently to solidify, they also shrink. However, the material shrinking last is attached to the plastic that had previously undergone shrinkage. This delayed shrinkage, occurring after attachment, behaves like a bi-metallic strip. The result is substantial internal stress and part distortion. Thus, an important goal in plastic molding is to improve the uniformity of the mold temperature distribution over time. Finite Element Analysis (FEA) results show that conformal cooling channels (CCC), in conjunction with high conductivity mold materials, can provide major benefits. By positioning the CCC properly in x, y, and z space, it is possible to reduce mold temperature variance. A key measure of mold performance is oTmax, defined as the difference between the highest temperature of the plastic and the lowest temperature of the plastic at the instant the first portion of plastic begins to solidify and shrink. The lower the value of oTmax, the more uniform the shrinkage and the smaller the resulting part distortion .
Figure 2 shows a conformal cooling channel used in the injection molding of a Vaseline jar cap for Chesebrough-Ponds. Note that the CCC transitions from a straight vertical section into an oval shape in the horizontal plane, and back to vertical again. Machining a channel of this geometry in a solid block of steel would be impossible in a single piece, or prohibitively complex and expensive in multiple sections. However, when the active surface of the tool has been electroformed as a thin nickel shell, then positioning CCC behind the shell becomes straightforward. Second, conventional drilled cooling channels (DCC) naturally have circular cross-sections. From Euclidean geometry, it is well known that of all possible two- dimensional shapes, circles have the smallest perimeter for a given enclosed cross-sectional area. Coolant flow rate (e.g. gallons per minute) is proportional to the enclosed cross-sectional area. The heat transferred from the mold into the coolant is proportional to the perimeter of the channel. Thus, a cooling channel with a circular cross-section provides the minimum heat transfer for a given coolant flow rate. A range of CCC cross-sections should be explored to determine which shape provides the most effective cooling per unit coolant flow rate Figure 3 shows an automotive wire harness clip. This part was injection molded in nylon using two different molds. First, ExpressTool Inc., Warwick, RI, built electroformed Ni-Cu inserts with CCC for United Technologies Research Center, East Hartford, CT. Concurrently, a conventional H-13 steel production mold at United Technology Automotive (UTA), Dearborn, MI, also was used to produce the same wire harness clip. The clip is 60 mm long (2.38 inches) by 35 mm wide (1.38 inches) by 30 mm high (1.18 inches). The same mold base was used in both cases.
Two separate cooling channels were dictated by the wire harness geometry. One CCC was used primarily to cool the central regions of the part, while the second channel cooled the peripheral regions. Although the geometry of these channels may seem complicated, it is important to note that these CCCs are fabricated easily and placed behind the nickel shell prior to copper electroforming. In this way, the cooling channels are completely encapsulated in highly conductive electroformed copper. The heat from the hot plastic can flow through the nickel shell, through the copper thermal management layer, and directly into the conformal cooling channel, where it is transferred away by convection. After set-up, thermal stabilization of the tool, and optimization of mold parameters, the measured cycle time for the production H-13 mold at UTA was 21 seconds. This corresponds to 3600/21 = 171 parts per hour, assuming uninterrupted operation of the injection molding press. Again, after set-up, stabilization of the tool, and optimizing the mold parameters, the cycle time for the electroformed Ni-Cu CCC mold was 12 seconds, corresponding to 3600 /12 = 300 parts per hour, again assuming uninterrupted operation of the injection molding press. Note that 300 /171 = 1.75, or a 75% increase in mold productivity as a consequence of utilizing an electroformed nickel-copper tool with conformal cooling channels.
In the second case study, Figure 4 shows a CAD model of a standard Vaseline jar cap injection molded in high impact styrene for Chesebrough-Ponds. The performance of an existing H-13 steel mold built with conventional DCC was compared to the performance of an electroformed Ni-Cu tool with encapsulated CCC. After set-up, thermal stabilization of the tool, and optimization of the molding parameters, the measured cycle time for the production H-13/DCC mold was 15 seconds, corresponding to 3600/15 = 240 parts per hour, assuming uninterrupted operation of the molding press. Again, after set-up, thermal stabilization of the tool, and optimizing mold parameters, the cycle time for the electroformed Ni-Cu/CCC mold was 9 seconds, corresponding to 3600/9 = 400 parts per hour, assuming uninterrupted operation of the molding press. Note that 400/240 = 1.67, or a 67% increase in mold productivity when using an electroformed nickel-copper tool with conformal cooling channels. It is clear from these two case studies that the reduction in mold cycle time, and the consequent increase in productivity for Ni-Cu/CCC molds relative to conventional H-13/DCC steel molds is dramatic. FEA results provide an explanation for these substantial reductions in cycle time, as well as major improvements in mold temperature uniformity. To gain a better understanding of the fundamental phenomena occurring within an injection mold, ExpressTool began working with the FEA/Process Modelling and Optimization Group at the National Research Council (NRC), Boucherville, Quebec, Canada, under the direction of Georges Salloum. The temperature distributions presented here were developed through a collaboration with Michel Perrault of NRC. The calculations were based upon the latest version of the NRC FEA code. Starting from a CAD model of a specific part, Perrault modeled the geometry of the mold, as well as the geometry of both the DCC and CCC cases, respectively. Finally, he used representative thermal and mechanical properties for H-13 steel, as well as those for electroformed nickel and electroformed copper where relevant. If one cannot understand a simple problem, the chance of understanding a related but more complicated problem is greatly diminished. Thus, the part selected for the initial FEA thermal analysis is a simple circular disk, 3 inches in diameter, and 0.1 inch thick. The part geometry, while flat, has a round shape typical of many injection molded parts, and also has little intrinsic stiffness, as there are no supporting ribs or gussets.
Figure 5 shows a top view of the two cases evaluated by FEA. The first case corresponds to an H-13 steel tool with DCC, shown on the left. The second case corresponds to an electroformed Ni-Cu tool with CCC, shown on the right. For this case, the CCC geometry looks something like a keyhole when viewed from above. While in principle the CCC could have arbitrary cross-sectional shape, the channel cross-sections were assumed to be circular.
Figure 6 shows the model of the Ni-Cu tool developed at NRC that formed the basis of the ensuing FEA analysis. Specifically, the part was assumed to be gated in the center, the nickel shell was assumed to be 2 mm (0.080 inch) thick, the copper thermal management layer was assumed to be 4 mm (0.160 inch) thick, the copper was assumed to fully encapsulate the CCC, and the tool was assumed to be backed with aluminum filled epoxy having a thermal conductivity of 2 W/moK. Compared with a thermal conductivity of 88 W/moK for nickel and 390 W/moK for copper, a value of only 2 W/moK for the mold backing material effectively treats the latter as an insulator.
Figure 7 is a pseudo-color FEA image of the distribution of temperature throughout the core sides of the two tools for: (1) the conventional H-13 tool with DCC shown on the left; and (2) the Ni-Cu tool with CCC shown on the right. The difference in the two temperature distributions is dramatic. The H-13 tool with DCC shows a hot spot to the left of the cooling channel (near the sprue) and another to the right of the channel. Conversely, the Ni-Cu tool with CCC shows an almost isothermal temperature distribution. The value of oTmax for the H-13/DCC case is 12.5oC. In contrast, the value of oTmax for the nickel/copper tool with CCC is only 2oC. Obviously, the combination of high thermal conductivity materials and conformal cooling channels significantly reduces mold temperature variations.
Figure 8 is another pseudo-color image of the temperature distribution on the cavity side of the tools for: (1) the conventional H-13 tool with DCC on the left; and (2) the Ni-Cu tool wi th CCC on the right. Here the effect is even more dramatic than on the core side. The value of oTmax for the H-13 /DCC tool is 18.6oC. The value for the nickel/copper tool is only 1.9oC, or essentially an order of magnitude reduction in mold temperature variance.
Figure 9 shows the pseudo-color temperature distribution for the cavity surface of the H-13 /DCC tool at two-second intervals from 1 to 15 seconds after plastic injection. These images illustrate the cooling of the tool over time. Figure 10 shows the same information for the Ni-Cu/CCC tool.
It is evident from inspection of these two figures that the cooling rates for the Ni-Cu/CCC tool are much faster than for the H-13/DCC tool. In fact, the temperatures throughout the Ni-Cu/CCC tool only three seconds after injection already are lower than the corresponding temperatures for the H-13 /DCC tool after 15 seconds. This data begins to explain the reasons behind the extraordinary productivity improvements noted in the two case studies presented earlier. In fact, the only reason the productivity gains are not even bigger is that the cycle time includes not only the cooling time, but also the times to: (1) close the press; (2) inject the plastic; (3) pack the plastic; (4) open the mold; and (5) eject the part. These five time intervals are not affected by the thermal conductivity of the mold, or CCC vs. DCC. Thus, the reported improvements in productivity are only realized through the reduction of the cooling time. However, since the cooling time typically is the largest component of the overall cycle time, the benefits of reduced cooling time can, and are, quite significant. In conclusion, nickel/copper mold materials, using conformal cooling channels, produce more uniform mold temperature distributions. The value of oTmax was an order of magnitude lower for the Ni-Cu/CCC than the corresponding value for an H-13/DCC. Nickel/copper mold materials, coupled with the use of conformal cooling channels also result in much more rapid cooling times, which in turn lead to much shorter overall mold cycle times. The nickel-copper/CCC molds resulted in productivity improvements of 75% and 67%, respectively, relative to conventional H-13 tools with DCC . For more information, contact the author of this article, Dr. Paul F. Jacobs, Vice President of Technology, ExpressTool, 300 Metro Center Blvd., Warwick, RI 02886; Tel: 401-737-7900; Fax: 401-737-8223; pjacobs@expresstool.com.
|
