Maybe someone can help
We are trying to improve the performance an injection moulding process.
At the moment we are seeing flashing, sink marks, knit lines and other
cosmetic problems.
One of the first problems we see is: how do we decide on a response
variable. Here are my suggestions:
-Simultaneously minimise 2 variables, one for flash and one for
underfill/sink marks. To calculate the flash rating of a piece look at
each surface that can have a flash mark and for each one give a rating
of none (0), some (1) or lots (2). Add these ratings up, perhaps run a
batch of 10-20 pieces for each treatment or test. Similarly for
sink/underfill 0=none 1=small sink/knit line 2=large sink. I reckon one
problem here would be measurement R&R.
-An alternative would be to establish an optimum weight for a piece and
measure deviations from it. Problems with this approach would be in
establishing an optimum, measuring accurately and establishing whether
cosmetic defects always cause changes in weight.
Any advice, references?
Thanks
Willo
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|