Plastics undergo a reduction in molecular weight with every molding cycle. The heat and shearing result in chain scission which reduces molecular weight, which results in a loss of ductility and impact resistance. For this reason, regrind is not usually a good idea for parts subjected to impact. The test I suggested above would help determine the level of regrind at which mechanical properties are degraded beyond an acceptable level. If your part has a flat straight section, you could machine an impact specimen straight from a molded part. This is a good idea anyway as you get the exact processing you part will see in production. The specimen does not need to be to exact ASTM dimensions, as long as it is the same for every test. Find a good lab and do impact tests on specimens with different levels of regrind. Melt flow testing would be beneficial if you have a rheometer in house and would allow you to do some preventative QC.
An issue with regrind is contamination. Any foreign matter that gets into the plastic can have a bad effect on impact resistance. If you know you are going to use reground runners, take care to keep them clean. Store them in covered containers, make sure the grinder is clean and free of dirt debris and other plastic, and keep the hopper covered. Only use runners from properly molded molded parts. Make sure that the regrind particles stay mixed with your virgin pellets. If they segregate because of different shape or size, you may get some parts made with very high regring content.
It may be that the parts a marginally acceptable in impact even without regrind, and the addition of regrind tips them over the edge. Make sure your process is optimum. There is one school of thought that says use a melt temp near the bottom of the recommended range and inject fast. The idea here is to get your reduction in melt viscosity from shear thinning rather than temperature. I saw data where two samples of I think PP were processed by molding parts ad grinding them up, checking melt flow, remolding, rechecking, and comparing results. One sample was molded fast and cool, the other slow and hot. Slow and hot had the same melt flow after one cycle as the fast and cool did after five molding cycles. Heat is the killer here. I think the suggestion about testing for antioxident levels is a good one if ABS is used. If the butidiene oxidizes, is will cross link and get brittle and you loose impact resistance.
Another issue with glass filled material is damage to the fiber from excessive shearing. I worked on a problem once where a glass filled polyester part was breaking at some hardened steel pins held captive in the assembly. The molder was inexperienced with glass filled part and was using a very fast screw speed. This was chopping the glass fiber into glass particles. We set the screw speed to a level recommneded by the resin supplier, and the part was so strong that the failure mode was now fracture of the steel pins. It may be that grinding the runners and the additional molding cycle are breaking the fibers.
If the failure is at a sharp corner, do what you can to remove the corner. PC is notch sentitve, meaning its a normally ductile material that will undergo brittle fracture in the presence of a stress concentration.
Rick Fischer
Principal Engineer
Argonne National Laboratory