RPA troubleshooting for eliminating injection molding non-fills with ASTM D6204.

Troubleshooting injection molding problems can be frustrating and time consuming. When an injection molded job is first started up on the production plant floor, many process parameters have to be optimized for good part molding. These parameters include mold temperature, cure time, injection pressures, injection speed and injection profiles. The goal is a consistently well-molded product with minimal scrap.

Hillsdale Tool & Mfg. launched a product consisting of two metal inserts bonded to a SBR/NR compound. The compound (Cpd.A) was a high viscosity material due to its specific dynamic property requirements. The uncured compound was supplied in strip form by a custom mixer to the production plant. Four injection presses were required to meet customer orders on this product. During production runs, occasional non-fills would occur. The problem was batch specific, since trials on the other injection molding presses produced the same results.

Corrective action to solve problem

The custom mixer tested each batch of material to specific cure limits. The cure test had [+ or -] 3 Sigma limits established for minimum torque, maximum torque, ts2 and tc90 cure parameters. Batches of Cpd.A that were producing non-fills were all within the [+ or -] 3 Sigma cure limits. There was no way to pre-screen potential problem batches. Discussions with the custom mixer led to an interesting discovery. Our production plant was sending the questionable batches of Cpd.A back to the custom mixer. The custom mixer was putting a short remill pass on them and sending the batches back to our plant, where these remilled batches of Cpd.A were injection molded without non-fills. Viscosity variations under the high shear of injection were now a potential cause. Possible viscous heating leading to scorch could also be the issue.

To evaluate the viscoelastic performance of a compound under high shear rates (1/s), the ASTM D6204 test was chosen. Method B of this test (for high strain) was chosen for the S' elastic torque component as an indicator of compound flow characteristics under increasing shear. Method B was modified by adding another frequency condition at 285 cycles per minute (29.86 rad/sec.). This is at the limit of the RPA 2000 for shear at a set strain of 100%. Higher shear would give additional data for injection molding.

Samples were collected from Cpd.A batches that did not produce non-fills. These samples were run on the RPA under the modified Method B test, and [+ or -] 3 Sigma limits were established for the S' elastic torque (figure 1). These limits were used for comparative purposes for bad processing batches. Eventually, batches (A1, A2) that produced non-fills were obtained and tested. These batches gave S' values above the upper +3 Sigma limits at the higher shear rates (figure 2). Variation of the processing characteristics for Cpd.A viscosity under high shear rates was a prime suspect. It should be pointed out that at low shear rate (6 CPM), a statistical difference was not evident. In a compression molding application, these bad batches (A1, A2) may not have been any different than the regular batches of Cpd.A, because compression molding is a low shear rate process.

[FIGURE 1 & 2 OMITTED]

A review of these data with the custom mixer led to the discovery that after mixing the first pass masterbatch, the second pass curative addition was immediately started with no aging time in between. With a high viscosity compound, being mixed in two passes, a sufficient aging time between pass 1 and pass 2 is needed to promote better mechanical breakdown and better uniformity. A 24 hour age time was established between the first and second pass. The results of the new mixing procedure are seen in figure 3.

[FIGURE 3 OMITTED]

We have now added ASTM D6204 with the modified Method B as a second RPA test for each batch of Cpd.A. This test allows one to see potential molding problems due to variations in uncured viscoelasticity under high shear rates. It also can be a good tool to use with custom mixers in developing better and more efficient mixing procedures for hard-to-process compounds.

Conclusion

The RPA 2000 is a very useful tool for doing troubleshooting and quality monitoring of both cured and uncured compound properties. The rubber compounder has a wide range of configurations for temperature, strain and frequency settings from which to select. Careful modeling can allow one to develop a specific RPA test configuration to evaluate the mixing, processing and dynamic characteristics of a compound.

Gary Veselica and Mickey Condon,

EaglePicher Automotive