SPC and Injection Moulding

[Howard Atkins]

After the subject arose in the cove in the last few weeks (Measuring Injection Molding Quality using Weight of final assembly I found this article that I wrote in 1992 as part of a project that I did for a practical plastic engineering course that I took.

 

[razorsedge]

Can’t thank you enough

Howard,

I have been working in injection molding for over 20 years both in manufacturing and quality. Your paper is much appreciated. We did some studies years ago with two “identical molds and identical machines” sitting side by side. All the equipment and the molds were new. We were attempting to control the process with weight, even went as far as using gram scales that measured down to .001 grams in an attempt to detect significant variation. To summarize we found the weights from the two machines had a very strong correlation as long as we were pulling the material from the same bin. As soon as we changed lots of material or changed the regrind blend (a whole other subject) there would be significant changes in weight. If we fed the machines from separate batches of material the weight correlation stopped. When we went back to feeding the machines from the same container of material the correlation came right back. Our conclusions were, set the machines and leave them alone. We found the same thing your study showed, SPC did not work well in a precision molding environment. Everything appeared to be out of control and the most significant factor was minor changes in the consistency of the raw material. The machines compensated for these changes better than we could. Our attempts to compensate just induced over correction.

The tough thing has been trying to explain this to various customers over the years. That’s were my gratitude comes in. Your paper is well written and precise. With your permission I will be handing it out to customers that insist we must use SPC, when in fact our efforts are better use to ensure the basics like consistent setup are in place. Thanks so much.

 

[Jim Wynne]

After the subject arose in the cove in the last few weeks(Measuring Injection Molding Quality using Weight of final assembly I found this article that I wrote in 1992 as part of a project that I did for a practical plastic engineering course that I took.

Thanks for sharing the paper, Howard. I think it will provide food for thought for a lot of our friends who deal with injection molding and the struggle to give customers the statistical data they want. One small quibble: you wrote,

A further problem is that SPC cannot deal with no change, this
means that a parameter that does not change will be by definition out of control, an absurd situation.

If there's no change, you've proven that SPC isn't necessary . In point of fact, though, just because one doesn't detect change doesn't mean that there isn't any. The situation you describe is evidence of a measurement system that isn't sensitive enough to detect changes in the process. Of course, in terms of specifications, the changes might be totally insignificant and further attempts to detect them might be unnecessary, but the statement "SPC cannot deal with no change" is tantamount to blaming SPC for lack of measurement discrimination.​

 

[pabloquintana]

Incredibly good. Thanks for sharing this information. I was pushing in that direction but without results. Now I know why.



Pablo

 

[gszekely]

SPC in Injection Molding

Thank Howard for sharing.
Pablo !
If somebody intrested more in SPC in IM, may take a loke on the folloving sites.
https://www.immnet.com
****DEAD LINK REMOVED****
If you run a serch on these sites on SPC, or Statistical process controll, you may find some more inputs. we had some good debates.
All of today's molding machines are equiped with SPC features for molding parameters. If you set controll limits parts can be sorted by robots, if you are out of controll. I like this more, is somekind of prerntion. I found it more efficient. even my dimensional measurements by a CMM, is too late sometimes.
On the other hand, there are out on the market a lot of good publications on IM, and each has a good section on SPC.
I would reccomend, Injection Molding Handbook, 3rd edition by Rosato, Quality Molding, By John Goff and Tony whelan, SPC in Injection Molding and extrusion.
If you open a machine handbook (Engel, KM, or Fanuc) you can find good info as well.
Hope this helps. Let me know if you need more info, as molding is one of my favorites. I just wanted to be a good QA Engineer, in molding, now I feel that I undertand a bit.
Gyorgy

 

[pabloquintana]

Thanks Gyorgy.

Yes, our machines are equipped with automatic SPC monitoring for process parameters. That is some homework I still have pending.

Regards.

Pablo

 

[Howard Atkins]

Thanks for sharing the paper, Howard. I think it will provide food for thought for a lot of our friends who deal with injection molding and the struggle to give customers the statistical data they want. One small quibble: you wrote,

A further problem is that SPC cannot deal with no change, this
means that a parameter that does not change will be by definition out of control, an absurd situation.

If there's no change, you've proven that SPC isn't necessary . In point of fact, though, just because one doesn't detect change doesn't mean that there isn't any. The situation you describe is evidence of a measurement system that isn't sensitive enough to detect changes in the process. Of course, in terms of specifications, the changes might be totally insignificant and further attempts to detect them might be unnecessary, but the statement "SPC cannot deal with no change" is tantamount to blaming SPC for lack of measurement discrimination.​

Yes you are right I did not really understand this in 1992 and I did not want to change the paper. In fact the resoloution become so fine as to be totally irrelevant to normal working conditions.

I have had a number of arguments with customers as to the use of SPC and have always believed and in fact found true that if you argue in a professional and intelligent manner then normally you can find a compromise.
Nowadays when I am in a different position one of my expectations from moulders is for them to argue about requirements for SPC by customers that In represent and I use this as a barometer of their professionalism.
If they accept CC's on the drawings with out comment then I know that they are not really serious and ask questions about shrinkage etc which they suddenly don't know the answers.

 

[Nehal]

thanx for info.

 

[ymselim]

Thank you for your information and ur file.

 

[KidPaddy]

Sir Howard,
many thanks for your generosity in sharing this