When one process output is inconsistent, check the process data in case there is not correlating data which can help pinpoint the actual cause of the problem.
Problem: Screw recovery time variation in PBT material.
notes: The questioner sent a screenshot of the SPC data for the past 200+ shots. In this data, it was shown that the shots with a low recovery time had a high cushion, very often above the transfer or near position. Along with the inconsistent cushion and recovery time, the transfer pressure and revolutions per recovery were also unstable. Additionally the average cushion is almost 40% of the shot size.
My Response: With only recovery data, one would initially infer a feeding problem, but when you combine all the available data, it is clear that the whole process is inconsistent. The cause could be due to many issues, but the core concern here is that the inconsistent recovery was a symptom of the problem, and not the problem itself.