Hello Mumus
Inspection criteria, and Gage R&R, in injection molding usually is performed for dimensional and visual characteristics, but for destructive measurements can be performed as well.
In my opinion the molding process, having good molds, machines, set up, and stable process, is fairly capable, with small variation.
If all the above is in place, usually the following mistakes are made during gage R&R:
- samples are not collected correctly. Assuming that the process variation is small, let say 30% of tolerance, assuming a tolerance for a part of 100mm +/- 0,2mm, then the stable process variation itself will be 0,12 mm(usually less, in order of hundrets of mm.) The samples then are collected from this range, while during set up, or mass production , you will produce parts out of tolerance or at least close to upper or lower tolerance, which in my opinion are part of the process, or at least you have to measure them and give indications, in which direction the settings should go. If the process under controll, does not produce parts to cover the whole tolerance range, usually I will take parts from start ups, and try to select them in such a way, to have at least covered the tolerance range, and have at least one or two samples out of tolerance, close to upper or lower limit.
As far as I know the R&R study results are reported to tolerance range and process variation range. If you chose samples from a process with a very small variation compared to your tolerance, then you will have poor results, while your measurement system in real life is capable.
Shortly:
- collect samples for study from set up parts as well
- chose samples in such a way that there is a significant part to part variation, for example having the sample above:99.75, 99.8, 99.85, 99.9, 99.95, 100.00, 100.05 and so on
If all the above are in place then we can go to measuring device, and fixtures.
First of all, your measuring device for the sample above should have a resolution capability of at least 10% of the tolerance range, ex. 0.04mm, to be able to indetify the part to part variation.
Fixture should be accurate enough to correct, eliminate, minimize the placement erors made by operators.
Everybody should be trained, and use correctly the measuring instrument.
If you have a measuring program that should be build up correctly, datumes or reference points corrected in each routine.
That's all what comes in my mind by first, attribute and destructive gage R&R, is an other story.
For process validation, adding to the good advice alrady given:
- mold qualifications or prequalifications should be run at mold manufacturer, and at least a half capability run done there, before takin it to your shop
- Injection molding machine technical specs and it's capability of meeting these, adequate screw for some resins, and raw material drying are some of the kes issues
- proper set up, and processing window, I have posted good sites and a set up sheet in several threads, see some:
http://elsmar.com/Forums/showthread.php?p=204687#post204687
http://elsmar.com/Forums/showthread.php?p=198126#post198126
http://elsmar.com/Forums/attachment.php?attachmentid=7063&d=1184139212
http://elsmar.com/Forums/attachment.php?attachmentid=6491&d=1172333050
There is a good guidance on the the ghtf.org site, document posted in other threds, you can find here:
http://elsmar.com/Forums/fileslist.php?mode=allfiles&sortby=filename&pageamt=2&criteria=ghtf
http://elsmar.com/Forums/showthread.php?p=204687#post204687
Most probably I may have made mistakes in my advice, compared to the theorethical R&R study rules, but this works in real life, and the gurus in the cove will correct me.

Good luck and let us know the results of your interview.
BR
György