SPC and Injection Moulding

[joansun]

Thanks very much for sharing!

Our company carry our strict quality-control , but still problem occur when the frist trial or samples deliver to our customer, will introduce to our factory.

 

[gszekely]

Hello Prezmek !
I'm not an on hand molder, just an engineer working some time ago in molding and frustrated by some department leaders telling me that molding is a black magic. So I started reading.
First of all I would highly reccomend a few books to have in your library:
Injection Molding handbook 3rd edition by Rosato (byble)
or Quality Molding by John Ghoff and Thony Wheelan
there are some more here:
https://www.amiplastics.com/ami/pidsubject.asp?dept_id=209
As for every other procees, improvement and controll without knowledge will lead to continuous firefighting, with problems, but also having heroes solving the problems. In my opinion there is no need for heroes in manufacturing at least.
Injection molding as every other process will have the basic variables you have to keep under controll:
material, man, machine, method, measurement,milieu, management, motivation,money.
Material:
start to know, there are technical data sheets available with the reccomended processing parameters available on manufacturer sites, the basics for processing.
You have to keep under controll if the material supplier will supply you consistent material, at least the moisture content of the material before processing. For this you must know to what moisture content you should dry your material, how long at what temperature. Also some knowledge about dryers, dew points, dessicant beds, volume of air flow and temperature, troughput and so on. If you know all these and keep under controll, but you have a central drying sistem with a small curvature in the conveying pipes where the material can seattle, or you have too much residance time in a non heated hooper dryer all you effort is useless.
Some info on dryers to start, and check manufacturer sites as well:
(broken link removed)
Some basics about your materials and handling you should know:
(broken link removed)
(broken link removed)
Moisture content:
(broken link removed)
there are som out on market based on loss of weight which usually are siuitable for monitoring and laboratory category on Carl Fischer method for "absolute" values.
There is some more to know about the materials, as rheology, with viscosity, shear rates, Melt flow rate or index, structures as cristaline semicristaline, amorphus, fillers and additives.
https://www.ides.com/
Check articles and browse the internet.

Mold: in my opinion this is the most critical to keep under controll, with part geometry designed, mold construction, with hot or cold runners, gates and types, cooling channels and so on. I don't have to much knowledge you can find some info on material manufacturer sites with some basic reccomandattion on tool design and on internet doing a serch.
Some sites:
https://www.moldmakingtechnology.com/
https://www.efunda.com/DesignStandards/plastic_design/plastic_intro.cfm
https://plastics.bayer.com/plastics/emea/en/technology/article.jsp?docId=10305
Machine:
how reliable, repeatable, stable machines you have.
What is your screw size, geometry, and condition of your plasticating unit (wear, ram size/shot size, do you inspect them regulary for wear, and tier.)
Some glas filled materials can wear out your unit in a few months, and not proper maintenance can lead to scrathes, dents on your front end components, screw and barrel. )
Set ups: do you have standard work, master set up sheets, processing windows and standard rules for troubleshooting. I have seen too many times people turning the knobs even with a proper set up sheet available, instead of trying to find out what has changed in processing conditions.
Some basics you can find on manufacturer sites, searching the net and here:

https://www.immnet.com/
Go to article arhieve, there are many usefoul info since 1996, and visit the networking forum, we have disscussed most of your questions, so you will find the answers: For example I have run a search on the forum for "monitoring parameters" (please check the box running the search for archieved threads as well) and I got the following:
(broken link removed)
Some more:
https://www.scientificmolding.com/
https://www.wjtassociates.com/page/page/1168073.htm
I would reccomend to purchasse at least Qualifications, startups and tryouts 2nd edition.
Check the links we have posted in this forum for students to check first:
(broken link removed)

I don't have more time at the moment, but I will come back if you say that you are interested and need more info.
I didn't answer your question, but that would have been only one sentence, you will find out more from my post.
BR
György

 

[Caster]

I've come back to this thread several times, but now I think I know what has my goat.

It seems the original question is whether SPC works/does not work with weight and or dimensions in plastic injection molding?

Aren't weight and dimensions are product features?

You can run control charts on them just fine. But it is called S - PRODUCT - C or S - QUALITY - C.

S - PROCESS - C for plastic injection molding would be temperatures, pressures, etc.

And I agree there is no need for S - PROCESS - C on new equipment which likely has real time process feedback/adjustment controllers built in.

S - PROCESS - C would be redundant if these controllers are set up and working.

I think what you are looking for is designed experiments to tie the product features (weight, dimensions) to the process settings (pressure, temperature, time) that make them?

I'm a foundry guy. Our melters have PID controllers on them. So I have mechanical/electronic process control, no need for S - PROCESS - C. Its fun to make control charts anyway and learn about autocorrelation and ndc.....fun, but not necessary. Lots of opportunity for error with controllers as well. Fun to play with them.

In my world we do "SPC" on machined casting dimensions, we measure the product feature but adjust the machining process (offsets, speeds, feeds, etc) to correct it. Works just fine.

For casting process parameters to "significantly" affect an as cast dimension (mold too cold, metal too hot, etc) we usually have already made a scrap casting (cold shot, misrun, run out, etc).

Maybe it's completely different in plastics world, but I think my main point is "was this discussion mixing SQC (product) vs SPC (process)?" and something about DOE.

I'm not sure this made any sense, but at least I was able to use lots of TLA (three letter acronyms)

Cheers

Caster

 

[Jim Wynne]

I've come back to this thread several times, but now I think I know what has my goat.

It seems the original question is whether SPC works/does not work with weight and or dimensions in plastic injection molding?

Aren't weight and dimensions are product features?

You can run control charts on them just fine. But it is called S - PRODUCT - C or S - QUALITY - C.

S - PROCESS - C for plastic injection molding would be temperatures, pressures, etc.

And I agree there is no need for S - PROCESS - C on new equipment which likely has real time process feedback/adjustment controllers built in.

S - PROCESS - C would be redundant if these controllers are set up and working.

I think what you are looking for is designed experiments to tie the product features (weight, dimensions) to the process settings (pressure, temperature, time) that make them?

I'm a foundry guy. Our melters have PID controllers on them. So I have mechanical/electronic process control, no need for S - PROCESS - C. Its fun to make control charts anyway and learn about autocorrelation and ndc.....fun, but not necessary. Lots of opportunity for error with controllers as well. Fun to play with them.

In my world we do "SPC" on machined casting dimensions, we measure the product feature but adjust the machining process (offsets, speeds, feeds, etc) to correct it. Works just fine.

For casting process parameters to "significantly" affect an as cast dimension (mold too cold, metal too hot, etc) we usually have already made a scrap casting (cold shot, misrun, run out, etc).

Maybe it's completely different in plastics world, but I think my main point is "was this discussion mixing SQC (product) vs SPC (process)?" and something about DOE.

Without going back over the entire thread, I think I might understand the point you're trying to make. First, I don't think your definitions of SQC and SPC are accurate. SPC is a form of SQC. The whole idea of there being a meaningful distinction between quality control and quality assurance (with the former concentrating on parts and the latter on processes) doesn't add value to anything. In order to assure that processes are properly designed and controlled, we have to measure output.

The idea of DOE comes into play only if it's not clear how the interaction of variables will affect output. In most cases, experience will obviate the need for controlled experimentation.

Part weight might be helpful in determining whether conditions exist that aren't readily apparent to visual inspection. In molding, wall thickness is a good example. We generally don't need to weigh parts in order to determine whether or not there's been a short shot, for example. Those are usually evident by eye. Nonetheless, weighing a part could be faster than looking for defects visually. How do we know how to identify "good" weight? It usually involves weighing a number of known conforming parts, taking the mean and standard deviation and setting some operating "contol limits." I put control limits in quotes because this method doesn't necessarily follow orthodox SPC methods. The idea is simply to segregate parts that don't fit the weight "profile" and examine them for defects. As time goes by, it should become evident whether or not this is a cost-effective method of output verification. It might turn out that weighing isn't helpful at all, or it might also reveal that weighing is more efficient than 100% visual inspection. It may also lead to process adjustments that bring the output closer to the target in general.
]

 

[Howard Atkins]

I've come back to this thread several times, but now I think I know what has my goat.

It seems the original question is whether SPC works/does not work with weight and or dimensions in plastic injection molding?

Aren't weight and dimensions are product features?



Caster

I started on the parameters but

"These control limits in some cases were so narrow as to be meaningless, for example, Injection time in one chart was within control limits only 0.006 s wide. The machine can only measure in hundredths of a second, this value is therefore of no practical significance" (page 3)

The paper showed that in injection molding the control of parameters of production were not a viable concept.

Most injection machines now are closed loop which makes the use of SPC moot.

 

[fresh]

Hi all, this is my first message on the forum, I must say that the knowledge on my subjects (Quality engineering and injection molding) of interest is incriduble, thanks.

I have gone through the article "SPC and injection molding" and the question that I have is wheather its necessary or not to measure variables on injection molded parts (made of engineering polymer's like nylon). The reason I am asking this is because a cavity is somewhat fixed, and there should not be significant changes on dimensions (besides short molds etc).
For an example the tolerances on our part are typically +/-0.1mm. We do a complete first off check and measure in process every one hour. My question is should we even be measuring throughout production or is first off measurements sufficient enough for a run. Thanks for any response.

 

[Howard Atkins]

This article was written some 15 years ago.

I agree with you about the "fixed" state of the mould. In my experience 95% of problems are due to "visual" problems- flash, short shots etc.
In my plant we only used to measure some 4-5 dimensions at set up and then every 24 hours.

From your history how many times have you had discovered a problem at your hourly inspections?
Use this information to build your own sampling plan.

 

[fresh]

Thanks a lot Howard for your response, this puts some clarity on my question. I would like to implemement a more effective attribute inspection method in place and reduce variables check to a minimum (i.e. once a shift). Do you by any chance have a good method of doing attributes inspection, or should I look at SPC for attributes. Thanks a lot.
Fresh

 

[Howard Atkins]

The problem is that I do not think that SPC is relevant in this case or in plastic at all!
The attributes "flash", "splay" "short shots" are defects and not acceptable. The situation is usually not a matter of gradual change which SPC can show as a trend which can be identified and corrected, but rather a problem with the mould or material or machine which suddenly appears.
Tolerances for the various parameters on the machine are important and use these to give yourself alarms.
Have you got data on your rejects over the years? This could be due to inserts breaking, contaminated material etc. Find the reasons for rejects and act on these systematically in the system as they probably are not only effecting 1 mould..
Apart from checking a set every period of time, the use of weight can hep to find flash or short shots, but remember that if you are working with nylon the weight will change with the time from injection.

 

[Palt88]

Dear all,

First I would like to thanks all the one(s) involved in this interesting post.

I'm not a specialist of SPC, I would say that I had never built or monitored real case. But now I'm starting to go deeper as our customer starts to speak about the introduction of the SPC in our molding lines.

I would just say to set the context that we are producing since 5 years the product without any SPC on, and we have 5 years of datas recorded behind us, and now the customer wants to introduce the SPC as part of their 6sigma project.

So I have 2 basics questions:

• Which criteria gives you the decision to put in place a SPC on a molding process or not?
• 2nd thing is that personnaly I would say that since a specification criteria (for example dimensional specification) got historically a very good CpK there is no interest to make any SPC as this means that the process is well capable!

Does someone could just tell me if I'm right or not?

Thanks