Prevent overheating of the mold valve pin
We had a clear nylon vessel that exhibited a “splay like” condition that emanated from the gate region. In our case, the mold had valve gates and what was happening is that some of the melt was being left on the face a sides of the valve pin. The next shot these now solidified pieces of nylon would break loose as the incoming melt stream re-melted the prior shot but because of the viscosity differences would result in a splay like appearance to the naked eye. Only under 30X microscopic exam could you actually see the residual debris and the flaring trails behind them like comets.
Our solution was to reconfigure the hot tip drop seating to prevent the overheating of the valve pin.
I suggest that this can happen anywhere in the melt stream where the melt flow can tear off crystalized resin that is just below the glass transition temperature, such as dead spots in the sprue or cold runner due to rough machining or dead spots in the internal runner channels of a hot runner system.
Probably the last thing you will look at/for. In our case it was 30+ days of solid troubleshooting before we hit upon this as the cause of our problem.
Also, I may have missed it, but if the injection can be profiled to vary the speed (reduce it) just before cavity full i.e. 99% full this can sometimes eliminate some shear related splay type problems. It is quick to do on machines so equipped and if it works, everyone smiles.