Multi Cavity Plastic Injection Mould SPC

[Bezerra]

HI there.

We work with plastic injection, and each mould as 2 to 8 cavitys, i was reviewing our Control charts, to make them more easy for the operators, and i run with a problem.

Literature says that for our case we should use Xbar & R charts, my doubt is:

Should i plot the limits (UCL and LCL) for each cavity (and then the chart will be very confusing)
or is there a way for me to find the limits for the group of cavitys?

Thanks.

 

[NumberCruncher]

Hi Bezerra

You might find the following article useful.

http ://www .minitab. com /uploadedFiles/Shared_Resources/Documents/Articles/being_in_control.pdf - DEAD (404) LINK UNLINKED

NC

 

[bobdoering]

I suggest that the first thing you should consider is what decisions do you expect the operators to make from the information the control chart will give them? Are you doing "report card" charting, where all you want to know is the product coming out of the process is OK? Or, is there an adjustment in the process you expect the charting to direct the operator to do?

Rubber stamping X-bar R charts is generally a waste, and - you are correct - confuses the operators. You need a "process control" decision to make statistical process control to be effective.

You also need to know your process capability. You may find that one cavity is more sensitive to variation, so tracking that one may be better than tracking all of them. You may also want to track a "within" variation chart - but rather within mold, rather than within part.

 

[Bezerra]

we whant the operator to be hable to read the chart and if necessary, make the necessary corrections on the machine, or in the worst case call the supervisor to stop the production.
we seach some kind of signal that can be understod to make immediate ajustements.

 

[bobdoering]

At this point in preparing an SPC chart you many want to consider a "Total Variance Equation" and/or CNX chart. What variables can you measure? Weight? Thickness? Length, etc.? Which variable will be affected by process variation first? Do you have a mold flow analysis? Do you know where the high and low pressure zones are? Your clues will likely occur in that area. Some high density parts may allow weight to be effective, but more often than not it does not enough sensitivity.