Filling the void in lost foam patterns: eliminating the steam flow entirely in thin sections (less than 7 mm) of polystyrene patterns stops void formation and reduces the scrap produced by lost foam metalcasting.
Casting dimensions depend on pattern dimensions," and some lost foam casting defects have been associated with variation in bead and pattern properties. But serious pattern defects, like voids (Fig. 1), cannot be eliminated by adjusting process settings alone.
[FIGURE 1 OMITTED]
A pattern void is a hole or space in the foam pattern that would normally be filled with beads. In general, it has been accepted that pattern voids are caused by poor bead fill. As a result, fill gun pressures often are adjusted to optimize the fill process for a given set of conditions, stopping the defects from forming. Unfortunately, the improvement in pattern quality is often short lived because the root cause of the defect has not been eliminated. In many cases, additional fill guns are added to a molding tool to improve its resistance to pattern defects, but this too has proven to be a short-term solution. A more permanent fix now has been found.
How the Void Forms
To gain a better understanding of how pattern voids form, a test pattern, known as a grate pattern (Fig. 2), was studied. A high-speed camera, working at rates of up to 2,500 frames per second, captured how the beads, followed by steam, entered and filled the molding cavity.
[FIGURE 2 OMITTED]
Patterns were made using the tooling in a vertically-parted, horizontally-acting molding machine. The grate pattern was made up of five bars connected at the top and the bottom of the pattern. The center bar was set up with inserts on the stationary and moving tools to make a foam section 0.197 in. (5 into) thick. The bars to either side of the center bar were set up to be 0.394 and 0.787 in. (10 and 20 mm) thick. The outermost bars were 1.57 in. (40 ram) thick, and the pattern was filled with three fill guns (Fig. 2).
Pattern scrap data were collected with three input variables--thickness of the center bar, fill gun air supply pressure and venting type--in an effort to discover the cause of defects. A replica of the moving side mold plate was machined out of a clear acrylic sheet to duplicate the shape and venting of the moving cavity (Fig. 3) while allowing the camera to capture the white side processes.
[FIGURE 3 OMITTED]
Locating the Voids
When the tooling was operated in its normal configuration, it experienced scrap levels varying from 20-80% due to pattern voids. The voids appeared almost exclusively in the center of the 0.197-in. bar.
Typically, the thinner sections of foam patterns are more susceptible to void formation. As a result, the threshold thickness at which the voids appeared was evaluated. The thickness of the center bar was adjusted from 0.197 to 0.276 in. (5 to 7 mm) in 0.039-in. (1-mm) increments. It was apparent that the scrap rate due to voids decreased with pattern wall thickness. Unfortunately, wall thickness is not something that can be changed frequently in a pattern that is already in production. It was suspected that either the beads were not filling the cavity adequately in the 0.197-in. bar, or they were moving around.
Voids that formed in the foam patterns generally were accepted to be a result of poor packing density during the fill step due to insufficient air or bead flow, water on the tooling, poor venting, etc. The EPS beads were colored to identify, which fill guns the beads came from, and it was observed that the voids occurred frequently where the bead streams would meet. The voids also tended to form at the same locations as the vents in the molding cavity.
The high speed videos of the fill step showed that voids did not form during the filling of the cavity. In addition, no voids remained at the end of the fill step. From this, it was apparent that the voids did not form during the bead filling process, but rather in the steaming process.
What was obvious was that the packing density of the beads was lower in the center bar than in the other bars. In addition, if the fill air pressure was reduced, the number of beads in the same area of the center bar also was reduced. In the actual production of the foam patterns, if the fill air supply pressure was reduced, the number of scrap patterns due to voids increased substantially (Fig. 4).
[FIGURE 4 OMITTED]
Stagnation zones and vortexes were generated where the bead fronts converged. In addition, the beads flowed out of the fill guns and tended to stop where they met an opposing bead front. From that point on, the bead filling process progressed back toward the fill guns until the cavity was full, and beads could no longer be blown into it.
Since void formation was not observed in any of the high speed videos of the filling process, the first step of the pattern fusion process was examined. For these experiments, the cross steaming method of fusion was used. For cross steaming, the stationary side steam valve was opened with the stationary drain valve closed. This process forced the steam to penetrate through the loose beads that were blown into the tooling on the previous step. The steam valve remained energized for three seconds.
Videos indicated that the only path for the steam to flow into the molding cavity was through the tooling vents. The first steam to enter the molding cavity condensed on the beads and the tooling. Shortly thereafter, the steam flow began to fluidize and rearrange the beads in the vicinity of the tooling vent. Since the beads were rearranged and packing had been enhanced by the action of the steam, the accumulated space created a void at the steam inlet location (the vent). If steaming were continued, the pentane within the beads would boil and the bead walls would soften, allowing the beads to expand. In some cases, this expansion was sufficient to fill the void created from the rearrangement of the beads in the vent region.
The high speed videos also demonstrated that the motion of the beads was centered on the vent locations in the center bar, and no motion was observed in the other bars. This result likely was due to the limited cross section of the 0.197-in. bar. In the thicker bars, the beads may have fluidized below the first bead layers and therefore could not be seen (Fig. 5). As a result, no motion was observed on the top surface. In addition, the thicker bars contained more beads backing up the layer along the tooling surface, which created resistance to the bead motion.
[FIGURE 5 OMITTED]
Based on the results of the high speed videos, it was apparent that the steam was moving the poorly packed beads around during the fusion steps. One solution would be to move the bead fill guns closer to the problem region, which would cause the beads in the problem area to experience a higher level of packing. Therefore, they would have a higher resistance to the steam flow than the poorer packed regions made with the original fill gun configuration. However, in production, fill guns often cannot be moved or added due to tooling geometry and piping constraints. It was apparent from the grate pattern that the 0.197-in. bar was susceptible to void formation, while the 0.394- and 0.787-in. bars were not. This behavior is typical of what occurs in production patterns. Again, geometry changes in a production tool can be difficult to implement. As a result, venting modifications were tested to understand their influences on void formation.
Altering Venting for Void Preventing
Several vent types were made to attempt to control the flow of steam into the sensitive region of the grate pattern. One new style of vent was made with 32 holes with diameters of 0.008 in. (0.2 mm). Vents also were made containing nine holes with diameters of 0.01 in. (0.25 mm). Finally, vent plugs containing no holes were used. Patterns originally were made with the conventional slotted vents installed in the inserts that are used to make the center bar. The new vents were installed in the inserts to understand the influence reducing the venting would have on pattern quality.
Decreasing the amount of venting in the thin section of the pattern had a dramatic influence on the occurrence of voids in that area. By decreasing the steam flow through the vents, the occurrence of voids also decreased. When the vents were plugged, the voids were eliminated because the steam flow was no longer able to rearrange the beads and create a void. Several tests were run by swapping the plugged and vented inserts in the tooling, and each time voids were found when the conventional vents were used and eliminated when the inserts with the plugged venting area were used.
In addition, longer term runs were made where the grate tooling was run daily with the conventional vented inserts installed. After the sixth day, the inserts were replaced with the plugs. The scrap level immediately decreased (Fig. 6). Patterns continued to be made daily under the new conditions, and a continued reduction in scrap further validated the theory that the flow of steam through the vents was able to fluidize the beads and generate a void in locations where the pattern was thin. By eliminating this steam flow path, the beads remained undisturbed but still were allowed to fuse, yielding a defect-free surface.
[FIGURE 6 OMITTED]
This article was adapted from a 2006 AFS Transactions paper (06-034).
Scott Biederman is a senior project engineer and Qi Zhao is a senior development engineer for GM Powertrain at Metal Casting Technology Inc., Milford, N.H. Douglas Matson is an associate professor at Tufts Univ., Medford, Mass.
For More Information
"Flow-Induced Bead Density Variations Observed During Expanded Polystyrene Bead Fill Operations," D.S. Samet, D.C. Mencel and D.M. Matson, 2006 AFS Transactions (06-042).
Scott Biederman and Qi Zhao, GM Powertrain at Metal Casting Technology Inc., Milford, New Hampshire Douglas Matson, Tufts University, Medford, Massachusetts
|Printer friendly Cite/link Email Feedback|
|Date:||Dec 1, 2006|
|Previous Article:||Avoiding gas defects through mold and core package design: results from a study on the gas evolution of certain mold and binder formulations can aid...|
|Next Article:||Crunching the numbers of void formation.|