Selection of mould design variables in direct stereolithography injection mould tooling

Stereolithography (SL) can be used rapidly to produce injection moulding tools. The disadvantage of the technique is that it is capable of producing only a small number of parts before failure. Stereolithography tools may break under the force exerted by part ejection when the friction between a moulding and a feature of the tool is greater than the tensile strength of the tool, resulting in tensile failure. Very few justified recommendations exist concerning the choice of mould design variables that can lower the part ejection force experienced and reduce the risk of SL tool failure. This research investigates the ejection forces resulting from the injection moulding of polypropylene (PP), acrylonitrile-butadiene-styrene (ABS) and polyamide 66 (PA66) parts from SL tools that are identical in all respects except for their build layer thickness (a process variable when generating the SL tooling cavities) and incorporated draft angles (a tooling design variable). This work attempts to identify appropriate evidence for recommendations with respect to these variables and SL injection moulding. The results show that linear adjustment of draft angle results in a fairly minor linear change in part ejection force according to the moulding material. A linear adjustment of the build layer thickness results in a greater change in part ejection force as a more non-linear relationship. In both cases the greatest ejection forces were experienced by PA66, then ABS and then the PP parts. The results also show that the surface roughness of all tools remains unchanged after moulding a number of parts in all polymers. A mathematical model was used in an attempt to predict ejection forces according to the moulding material used. This model did reflect the experimental results in terms of relative values but not in absolute values, which may be due to the limitations imposed by the development of the expressions and uncertainty about some specific values.