The Mold Filling Analyst, #14

Editor's note: This month's Mold Filling Analyst was written by Giorgio Bertacchi and Anne Bernhardt, of Plastics & Computer, Milan and Dallas. The company provides analysis service as well as selling faGaim software that performs mold analysis.

Optimizing a part molded using gas-assist is a challenge, but one that can be aided by the use of analysis software that models the behavior of the gas. Our example here is an ABS part that represents a plastics component of a type commonly found in the appliance and business machine industries. We're using it here as an example illustrating the procedure that may be used to achieve optimized part and tool design. The 600-mm-by-150-mm part is a flat panel that must be free of any surface blemishes. The project required a thick (14-mm) section to be hollowed out over its entire length, as shown in blue in Figure 1. A two-cavity mold was constructed, with melt and gas entry points in one side of the channel that was to be hollowed out.

Initial production was based on testing and intuition (no computer simulation was used) and yielded parts that were not fully acceptable. However, they confirmed that gas-assist molding would lead to the desired results.

The sequence of modifications illustrated here was based on results of computer simulations using the faGaim finite-element model. Whenever analyses are conducted on parts produced in existing molds, it makes sense to begin by modeling the problems and interactions already encountered during actual production before attempting to solve the problem at hand. When good agreement between the simulation and actual molding conditions has been demonstrated, as in this case, the modifications called for by the simulation can be carried out with confidence.

Figure 2 indicates the nature of the problem encountered. The amount of polymer in each section is highlighted. Red indicates the absence of any coring out (100 percent plastic), while the dark blue indicates the absence of resin. Light blue indicates the desired coring (residual plastic about 40 percent, or 2.8- mm skin thickness).

Zones of gas penetration, where cosmetic effects such as changes in gloss and color are generated by a particular type of gas penetration called "fingering," are shown in yellow. To overcome this flaw, it was decided to rerun the simulation using a modified mold, reducing the thickness of the area between the thick and the thin compact sections (shown in green in Figure 3). To our great surprise, the result was worse than the preceding evaluation, as illustrated by the gas penetration shown in Figure 4. The isochrones of the filling phase in Figure 5 show why. The restriction lateral to the thick section, created to avoid gas penetration, increased the flow of melt in the thick channel. It became completely filled with obvious consequences on the filling of the compact region. The gas cores out sections that have no resin left at their flow front, causing the undesired gas blowout.

The molding conditions for the simulation were then changed to overcome this problem: a delay between the end of filling of melt and the start of gas entry was programmed. The results of this delay are illustrated in Figure 6, which shows the evolution of pressure and flow rate. The green curve shows the increase in pressure during the filling of melt, and its reduction to zero for 3 seconds before the gas flow begins. Under these conditions the material cools and it prevents gas penetration, as highlighted in Figure 7. Unfortunately, the hollowing of the channel becomes insufficient. The green color indicates 60 percent residual plastic, or a 4.2-mm wall. This requires a longer cooling time, but more importantly, even with a very long cycle, the part warps due to the differences in wall thickness.

The solution to this problem was found by simulating an imaginative idea. The gate was moved to the proximity of one of the three little square openings situated almost in front of the original gate position. This balanced the melt flow as required for gas-assist, and is shown in the isochrone of Figure 8. The gas entry point was also moved and placed at the beginning of the thick section. An optimum core out was obtained, assuring parts of excellent flatness. Actual molding trials confirmed these results.