EXPERIMENTAL DETERMINATION OF THE AESTHETIC, DIMENSIONAL AND

CONTROL PROCESS WINDOWS USING THE TECHNIQUE OF

DESIGN OF EXPERIMENTS IN INJECTION MOLDING.

Suhas Kulkarni, FIMMTECH Inc., Vista, CA

Abstract

Traditionally, the mold qualification procedure involves

sampling the mold and establishing a process to make

acceptable parts. Process engineering studies such as in-

mold rheology, gate seal tests and generating process

windows are most commonly performed. Design of

Experiments (DOE) are conducted to study the effect of

the molding factors on the quality of the part and sometimes to find the process to mold a part within the

quality specifications. The process engineering studies

and the DOEs are conducted independent of each other.

This paper combines these techniques and introduces the

concepts of the Aesthetic Process Window, the

Dimensional Process Window and the Control Process

Window.

Introduction

The quality of the molded part is dependent on a

number of processing variables. Each part has a specified

quality requirement such as dimensions or the aesthetics

of the part. These are mentioned on the part drawing

together with upper and lower tolerance limits. The part

quality must be maintained within these limits for it to be

acceptable.

Process Engineering Studies are often performed to

optimize the process. In-mold rheology studies1 help

optimize the injection speeds. A gate seal test will

determine the optimum holding time. Process window

studies, also known as molding area diagram2 are used to

determine the extent of the molding areas. Wider the area,

more robust is the mold.

Design of Experiments are techniques where one can

study the effect of multiple variables on the quality of the

part. The analysis will provide the most significant factor

or factors that affect the part quality. The production team receives the information about the parameters that need to

be changed in case there is an issue with part quality.

DOE results will also give information about the factors

that have no effect on part quality and should therefore be

left unchanged. They also provide information about the

optimum combinations of the parameter values to produce

a part with certain dimensions or quality requirement.

This is done through the generation of prediction

equations.

Although the two procedures are widely used and generate useful results, the correlation between the two is

very important. In this paper we will study the

combination of these two techniques. Three terms are

introduced. They are the Aesthetic Process Window

(APW), the Dimensional Process Window (DPW) and the

Control Process Window (CPW). The concept will

demonstrate the acceptable area where one can mold

acceptable parts.

Experimental

Materials: High Density Polyethylene (HDPE)

Equipment: Arburg Molding Machine, 77 Ton, 75 gm shot size,

30 mm screw dia.

Injection Mold

NAUTILUS Process Engineering Software3

WISDOM DOE Software4

Procedures

The mold was mounted on the molding machine and

the melt and the mold temperatures were set within the

manufacturer’s recommended parameters. The injection

speed was optimized using the in-mold rheology

technique and was set to 101.6 mm/sec (4 in/sec). The

holding time was optimized via the gate seal test and was

set to 4 seconds. It was determines that the part had sink

below 2757 kPa (400 psi) and had flash above 5515 kPa

(800 psi). Therefore these were considered as the low and

the high limits of the holding pressure. Based on this information and the manufacturer’s recommendations, a

three factor DOE was designed using the WISDOM

software. The experimental matrix is shown in Table 1.

Table 1: DOE Matrix (Units: deg C, sec, kPa)

The experiments were run and the data was collected. A

length and diameter specified on the part print were

measured. Table 2 shows the quality specifications for the

part.

Length (mm) Diameter (mm)

Nominal 144.65 (5.695 in) 40.08 (1.578 in)

+ Tol 0.13 (0.005 in) 0.08 (0.003 in) - Tol 0.13 (0.005 in) 0.08 (0.003 in)

USL 144.78 (5.700 in) 40.16 (0.003 in)

LSL 144.52 (5.690 in) 40.00 (1.575 in) Table 2: Quality specifications for the part.

The data was analyzed using the DOE software and it was

found that the holding pressure was the most significant

factor that affected both the length and diameter of the

part. The second most significant factor that affected the

length and diameter of the part was the cooling time. This

is shown in the Pareto Chart in Figure 1 for the length of

the part. A contour plot for the both these responses were

generated with the Y-axis being the holding pressure and the X-axis being the cooling time. These are shown in the

subsequent figures and are used for the rest of the

discussions below.

Results and Discussions

The following terms will first be discussed with respect to one dimension only, the length of the part. The concept

will then be extended to multiple dimensions.

The Aesthetic Process Window (APW): The process

window that was generated during the process engineering

studies only looked for the aesthetics of the part; parts that

were free from defects such as shorts, sinks or flash.

Dimensions were not considered. This window where

only part aesthetics are observed is called as the Aesthetic

Process Window or the APW. The window determines the

processing limits that cannot be crossed to avoid molding an aesthetically defective product and/or molding below

or above a recommended process parameter value. Figure

2 shows the APW.

To mold parts that are dimensionally acceptable, one

must now look inside the cosmetic window. A DOE with

these two parameters as the main factors must be

performed and then the data must be analyzed. The part

we will be discussing is a part that had a length and a

diameter as critical dimensions shown in Table 2 above.

The Dimensional Process Window (DPW): Refer to

Figure 3. This is a contour plot that was generated for the

length of the part using the same limits for the DOE. The

data has been extrapolated to show the window and avoid

any overlap. On this contour plot, the LSL, the NOM and

the USL are also marked. The green shaded area is the

area where the parts that are molded are dimensionally

acceptable. Outside this area there are reasons that the

parts will be defective. Above 5515 kPa the parts will

flash and below the red LSL line, the parts will be out of

spec, This window that is now a subset of the aesthetic

process window is called the Dimensional Process Window or the DPW. In this case, the window is not a

uniform quadrilateral and the molding parameters must be

selected inside this area.

The Control Process Window (CPW): We shall now take

this concept further. If the dimension under consideration

is a dimension that needs to be statistically capable, then a

Control Process Window can be defined. Once the

molding process is started, the process capability can be

calculated and the control limits can be established. We

can now apply this to the DPW. Refer to Fig 4. Based on

the statistical process capability, the Lower Control Limit (LCL) and the Upper Control Limit (UCL) were

calculated and are shown in the figure. These contours can

now be plotted on the contour plot and the CPW can be

determined. This new window where the parts are now

not only acceptable but also under statistical quality

control is called the Control Process Window (CPW).

The internal blue shaded area is the CPW.

Multiple Dimensions: We shall now extend the above

concepts to incorporate the second dimension, the

diameter of the part. When there are multiple dimensions involved, the contour plots get complex. The contours of

the nominals and the specification limits of each

dimension will rarely overlap and even rarer would be the

slopes of the contours of the same value. The effective

process window will therefore be the intersection of the

two individual process windows. Refer to the Figures 5

below which shows the overlap of the two windows, one

for the length and the other for the diameter. It also shows

the control process window. Naturally, the overlap is

smaller than the individual process windows. This is the

Composite Dimensional Process Window.

Multiple Cavities: The concept can again be extended to

multiple cavities. As one can imagine, the above concepts

can get very confusing when applied to multiple cavities.

In theory, all the cavities should be producing parts with

the same quality and it is a matter of applying the results

of one cavity to the rest of the cavities. For parts to be

identical in all the cavities, geometric and rheological

flow balance is required in the mold.

The following terms can now be defined:

Aesthetic Process Window (APW): The limits between which an aesthetically or cosmetically acceptable part can

be molded. Dimensions are of no concern.

Dimensional Process Window (DPW): The limits between

which a dimensionally acceptable part can be molded.

Control Process Window (CPW): When the statistical

control limits are applied to the dimensions on the

Dimensional Process Window, the window of operation is

called the Control Process Window.

Composite Process Window: (CoPW): When multiple

dimensions and multiple cavities are analyzed, the results are displayed in a Composite Process Window.

Conclusions and Remarks

Although mold qualifications using process

engineering studies and DOEs are routine, there is an

absence of a method of correlating the data from the two techniques. The two techniques provide important data

but the combining the two provides a very clear

understanding of the capabilities of the mold and the

molding process. Wide process windows provide an

opportunity to run successful molding operations. Some

would argue that the window gets smaller and smaller as

the number of dimensions and cavities increase.

Unfortunately that is the reality, and the fact of the matter

is that this has been a hidden truth. This is the primary

reason why most companies have ‘one’ approved process

sheet with some acceptable tolerances. The tolerances are usually placed based on past experiences in molding. The

tolerances must be set based on the type of analysis

performed above and the results obtained from them.

Understanding the sensitivity of the part quality to these

factors gives an insight to setting the process parameter

tolerances during molding.

Having multiple dimensions and multiple cavities

makes the windows smaller and smaller. This further

pleads the case for a well built mold and parts with

practically defined tolerances. Concurrent engineering principles and practices become important. A robust

process requires less human intervention. The frequency

of inspection can also be reduced. Upfront planning and a

disciplined approach to process development are therefore

imperative to the efficiency of the production run and the

profitability of the operation.

References

1. John Bozzelli, Injection Molding Solutions, Midland,

Michigan.

2. Injection Molding Handbook, 2nd Edition, Dominic

Rosato.

3. FIMMTECH, www.fimmtech.com, Vista, CA

4. Launsby Consulting, www.launsby.com, Colorado

Springs, CO.

Acknowledgements

The author thanks Distinctive Plastics, Vista, CA and its management for the help with this study.

Figures

Figure 1: Pareto Chart obtained from the DOE analysis for the length of the part.

Figure 2: The Aesthetic Process Window

Figure 3: The Dimensional Process Window for the Length of the part.

Figure 4: The Control Process Window for the Length of the part.

Figure 5: The Composite Control Process Window for the Length and the Diameter of the part.