A statistical design of experiments, or DOX, is the preferred method to determine an optimum blow molding process and processing window. Initial mold/process evaluations should be completed on a small prototype mold when possible. The mold is normally shipped from the tool builder to the molder's development laboratory. The molder's development laboratory should have the time and resources available to evaluate the molding process for that mold and to perform any initial Design of Experiments.
The objective of a design of experiments is to not only identify an optimum process and processing window, but also to identify the main effects from process variables, interactions between them, and possible curvature effects. Initial experiments are run to determine which combination of process control parameters yields the lowest dimensional variability. Once this is determined, the tool steel can be adjusted to meet the molded container dimensional specifications. It is important NOT to choose process control parameters solely based on the part specifications. If your tool builder constructs the mold using a "Steel Safe" approach, you will be able to easily cut and ‘adjust’ the steel for the most stable and efficient process.
It is recommended that the Design of Experiments be performed on the blow molding machine on which the mold will run on during production or sampling. Only a skilled molding company can accurately translate the process control parameters from one blow molding machine to another. Even if the machines are similar, there are always differences that will affect the attributes of the molded part. The amount of wear on the extruder, the mold clamp system design, hydraulic valves, screw, barrel and timers can differ and can have large effects on process control parameter adjustments. In addition, the molder’s development laboratory often utilizes a different (mold) cooling system than the system used in a manufacturing area. Differences in development laboratory (individual mold temperature control systems) and manufacturing plant (central) coolant system performance can be significant. Careful attention must be given to the actual cooling (heat transfer) variables in the mold.