Kent 2016

Chapter 10
Processing quality
management
Processing is where the difficult stuff moulding. In any case, many of the issues,
starts and quality processing is a key to such as shrinkage, are unique to the
success in plastics processing. In the plastics industry but are shared by all of
preceding chapters we have covered the the common processes.
basics of quality management but this We will also cover Engineering Process
section goes into the detail of managing Control (EPC) a development in the
plastics processing to produce high quality control of processing equipment to deliver
products to the customer specification. quality automatically and the process for
This is what plastics processors do every initial sample approval (PPAP).
day and is at the heart of why they exist.
In many ways, this chapter brings
If you are not producing high-quality
together the threads that have been
products then no matter how low your
developed earlier in this Workbook and
price, you will not be in business for long.
readers are advised not to jump in here
In this chapter we will cover most of the without reading much of the preceding
main plastics processing methods and sections - still, you won’t listen to me
consider some of the unique issues that anyway.
need to be considered to consistently
deliver quality products cost-effectively
from each process.
Initially, we will deal with some problems
that are common to many industries, e.g.
setting process standards, creating process
FMEAs, establishing control plans,
deciding between product and process
control and testing, the best ways to set
machines and the basics of off-line quality
control. These are all vital issues
whichever process you are using and help
to set the scene for controlling quality in
processing.
We then deal with the unique issues for
several of the major processes used with
plastics. The main processes used by the
industry are extrusion and injection
moulding (by volume) and naturally we
will spend longer in the discussion of these
processes. The other, smaller-volume, “If you have always done it that way, it is
processes will be dealt with but simply not probably wrong.”
to the extent of extrusion and injection Charles Kettering
Chapter 10 - Processing quality management 293

10.1 Process design and development
Taking product design to • Consider ‘push/pull’ production If you are inspecting

process design decisions. lots of products
then you have got it
• Identify where any reworked products (if
As the product is being designed (see wrong further up the
Chapter 8) the production process itself any) will be inspected before re-entering chain.
needs to be designed to ensure that the the production process.
system is capable of delivering to the • Identify and correct any potential quality
customer requirements. Process design issues with handling or external
must be carried out in parallel with processes.
product design.
The product design is validated at the
Process FMEA
prototype stage and the outputs of the A process FMEA is essential before
product design and development review production trials are started. A process
are shown in the left column of the table FMEA is similar to a design FMEA but
on the lower right. All of these outputs are covers the possible failure modes during
necessary to complete an effective process production rather than design failure
design. However, process design and modes. The methodology is similar but the
development will always continue for some boundaries are very different.
time after pilot production to The process FMEA is covered in more
accommodate changes in product design. detail in Section 10.2.
During process design and development
the team develops the process to ensure Pre-production control plan
that this is both comprehensive and A pre-production control plan is essential
effective, i.e. that the process is ‘capable’ of before production is started. This is
meeting the customer’s requirements. similar to the site-wide control plan
Process design and development in the discussed in Section 4.18 but is much
context of the AIAG APQP process is
shown on the far right and the outputs of
the process design and development phase Inputs from product Outputs to product and
are shown in right column of the table on design and development process validation
the lower right. (see Section 8.9) (see Section 10.5)
Process flow chart Design FMEA. Process flow chart.
The process flow chart is one of the most DFM and DFA. Process FMEA.
important outputs from process design
Design verification. Pre-production control plan.
and development. This is a complete flow
chart of the process flow and can be used Finite element analysis (FEA). Process instructions.
to identify sources of variation using cause Engineering drawings and Measurement systems
and effect charts) for the complete process. specifications. analysis plan.
A process flow chart means that the team
can look at the complete process rather Materials specifications. Preliminary process capability
than on the individual steps in the process study plan
Tooling specifications.
and this is invaluable when completing Floor plan layout.
the Process FMEA (see Section 10.2) and New equipment
the Control Plan (see Section 10.3). specifications. Packaging standards and
specifications.
The process flow chart is completed as per Draft control plan.
a standard flow chart (see Section 7.2). It Prototype build.
must:
Design review.
• Show the sequence and linkage between
production and inspection/control areas.
The inputs and outputs of process design and
• Include links to the Process FMEA and development
the Control Plan.
Initial process design is carried out in parallel with the product
• Describe, in detail, how the product will
design but after the product design is complete, process design
move during the process.
and development continues.
294 Chapter 10 - Processing quality management

more product-specific. layout must be linked to the process flow A site cannot
The pre-production control plan is covered chart and to the pre-launch control plan. operate at more
than one quality
in more detail in Section 10.3. • Tip - Check that all process and
level.
• Tip - A pre-production control plan will inspection areas are of the correct size
and adequately lit. The quality level for
inevitably contain more product and all of your products
process controls than the final • Tip - Check that inspection areas are will effectively be
production control plan. located to prevent the shipment of non- the same whether
conforming products. you want it to be or
Process instructions • Tip - Check that the layout and/or
not.
The process instructions are at the heart controls prevent the mixing of finished
of successful plastics processing but many goods with other similar products.
companies neither set machines in a • Tip - Check that the layout and/or
logical manner, record these settings controls prevent environmental
accurately nor enforce the use of these contamination from external sources.
settings.
Getting the process instructions correct is Packaging standards and
covered in more detail in Section 10.4. specifications
It is usual that the customer will have
Measurement systems analysis
specified the packaging for delivery and
The pre-production control plan gives use but if this is not available then
details of the controls and measurements packaging standards and specifications
to be made during production but analysis must be established and documented so
of the measurements to ensure that they that the product characteristics are
are valid is rarely carried out in the maintained during storage and transport
plastics industry. In fact, it appears to be to the customer.
almost unknown except for those
companies dealing with the automotive Product and process validation
industry.
After completion of the process design and
The use of measurement systems analysis development, the next step is validation of
and a discussion of this is given in the product and process via pilot
Appendices 6 and 7. production and a production part approval
process (PPAP). This is covered in Section
Preliminary process capability 10.5.
study plan
The characteristics identified in the pre- •1. AIAG. 2008. Advanced Product Quality
launch control plan (see above and Section Planning and Control Plan. AIAG.
10.3) will need assessment for process
capability during product and process
validation. This will necessarily be a
machine capability assessment (see
Section 5.4) because full process capability
(see Section 5.22) can only be carried out
after long-term production to allow other
external factors to be included.
At this stage the output is the plan for
machine capability assessment and not
the actual assessment.
Floor plan layout

The floor plan layout describes, in detail,
the layout of the production area. The
floor plan layout must show the location of
all machinery, tools, process instructions,
inspection areas, control charts, visual Process design and development in APQP
aids and storage areas for materials (raw Process design is essentially complete after pilot production
materials, work-in-progress, suspect when the product and the process are both validated to ensure
materials, rejected materials and finished that they produce a product that meets the customer’s needs
goods). Material flow in the floor plan (see Section 10.5).

10.2 Process and machinery FMEA
It is all coming together now Risk Priority Number (RPN) Process and
machinery risk
A process and machinery FMEA is very For each potential cause, the individual assessment can
similar to a design FMEA (see Sections factors are rated and multiplied together remove many of the
8.11 and 8.12) and uses the same basic (RPN = S x P x C) to produce the relevant small errors in
risk assessment methodology. In this case, RPN rating. process design that
the assessment is concerned with the cause poor quality
Action can then be taken and recorded on
production process and machinery. as it is perceived by
the significant RPNs to: the customer.
A process and machinery FMEA should be • Reduce the severity.
part of the process design to assess
• Reduce the probability.
potential risks to quality and the action
that can be taken to reduce these risks. • Improve the process or temporarily
This will lead to more fault-tolerant improve the controls. There is no ‘magic’
RPN number above
production processes and a reduction in The RPN is the reassessed after which you must
production-related quality issues. completing the actions. take action. This is
The process and machinery are • Tip - Action should be mandatory for simply a method of
investigated to assess potential areas any severity rating of 9 or above. highlighting and
prioritising risks.
where poor performance or failure could • Tip - A process and machine FMEA
impact on product quality or safety and should be reviewed and revised with You can take action
each potential risk is assigned a rating on whatever you
time (it is a ‘live document’) to reflect like but it does
after considering the following: new equipment, processes and makes sense to
• Severity (S) - 1 is low severity and 10 is procedures. This will allow the control reduce the highest
high severity. plan to be reviewed with the experience risks first.
• Probability (P) - 1 is low probability and gained.
10 is high probability. • Tip - A process and machinery FMEA
• Current controls (C) - 1 is high current should be linked to each part of the
controls and 10 is low current controls. process flow chart. The best FMEAs
• Tip - As noted in Section 8.11, are often the
Values for S, P and C for each risk/event shortest. Get the
are given in the tables below. For SnapSheets XL from SigmaZone can do basics of the
consistency, these are the same tables that most of the hard work in generating a process right and
are used for a design FMEA. process FMEA. the rest follows.
Severity of event Probability of event Current controls

Ranking Ranking Ranking
(S) (P) (C)
Hazardous: Very High: Absolute uncertainty of
10 10 10
without warning event is inevitable detection
Hazardous Very remote chance of
9 9 9
with warning detection
Very High 8 High: Repeated events 8 Remote 8
High 7 7 Very low 7
Moderate Moderate:
6 6 Low 6
Occasional events
Low 5 5 Moderate 5
Very low 4 4 Moderately high 4
Minor Low:
3 3 High 3
Relatively few events
Very minor 2 2 Very high 2
None Remote:
1 1 Almost certain to detect 1
Event is unlikely

Process Sash Profile - Compact 121 Date 05/10/2015
Prepared By Robin Kent
Notes
Current process controls Recommended actions Action results

Potential effect of
Function / Requirement Potential failure mode Recommended Target
failure Potential cause Prevention Detection R.P.N. Responsible Actions Taken
Class
(P)
(C)
(S)
(P)
(C)
Sev
Sev (S)
Prob
Prob
actions completion date
Class
R.P.N.
Controls
Controls
Extruder and upstream see previous sheet
Regular Start-up check by Production control

maintenance operator/setter. for checklist
Controller/thermostat failure 2 QC on output 5 80 21/03/2016 Added to checklist 8 2 2 32
check for Add visual check Operator/setter for
operation start-up checklist regular checks
Start-up check by
Extrusion difficult Regular Production control
operator/settter.
Die heat failure Profile out of tolerance maintenance for checklist
8 Wiring failure 2 QC on output 5 80 Add visual check 21/03/2016 Added to checklist 8 2 2 32
(complete) Potential damage to check for Operator/setter for
to start-up
thrust bearing operation regular checks
checklist
Start-up check by
Die heating Regular Production control
operator/settter.
maintenance for checklist
Heater failure 2 QC on output 5 80 Add visual check 21/03/2016 Added to checklist 8 2 2 32
check for Operator/setter for
to start-up
operation regular checks
checklist
Add separate
Maintenance
Die heat inaccuracy thermostat and
Profile out of tolerance 6 Controller/thermostat failure 3 check for QC on output 5 90 Maintenance 21/03/2016 To be completed 6 3 3 54
(too high) warning light to
calibration
system
Chapter 10 - Processing quality management

Extrusion difficult Regular Add separate
Die heat inaccuarcy Profile out of tolerance maintenance thermostat and
8 Controller/thermostat failure 3 QC on output 5 120 Maintenance 21/03/2016 To be completed 6 3 3 54
(too low) Potential damage to check for warning light to
thrust bearing calibration system
LED indicator of
Regular
Chiller settings for complete chiller water flow
2 maintenance 1 12 No action 6 2 1 12
system temperature visible
check for settings
in production area
LED indicator of
Regular
Chiller failure water flow
1 maintenance 1 6 No action 6 1 1 6 WFF
(general system) temperature visible
Profile not fully cooled check for settings
Water bath temperature in production area
and not calibrated 6
(too high)
properly LED indicator of
Regular
Profile calibration Pump failure water flow
2 maintenance 1 12 No action 6 1 1 6 WFF
(general system) temperature visible
check for settings
in production area
Start-up check by
Regular Production control
operator/settter.
Hose restrictions maintenance for checklist
4 4 96 Add visual check 21/03/2016 Added to checklist 6 4 2 48 WFF
(for specific machine) check for hose Operator/setter for
to start-up
condition regular checks
checklist.
LED indicator of
Profile cooled too quickly Regular
Water bath temperature Chiller settings for complete chiller water flow
before calibration 2 2 maintenance 1 4 No action. 2 2 1 4 WFF
(too low) system temperature visible
effective check for settings
in production area
Continued on next sheet
Example of a Process FMEA using Snap using SnapSheets XL to generate Risk Priority Numbers
This is not a complete example and the actual process FMEA is much longer and detailed. It is shown only to illustrate the method and the output.
297
10.3 Control planning and control plans
A documented description of Stages of control planning Every part must

have a control plan,
systems and processes for A control plan should be produced and but family control
controlling the product revised at 3 separate stages during the plans can be used,
Control planning and the control plan are journey to full production: e.g. an extrusion
• The prototype control plan focuses site producing many
essential elements of advanced quality
different profiles for
planning and are some of the building mainly on the product design and
similar uses could
blocks by which quality is planned into the conformance to requirements rather devise a ‘family’
product. than on the process. control plan’ for the
• The pre-production control plan is a full all production.
Control planning for production is similar
to the global quality plan concept shown in description of the dimensional, material
Section 4.18. In the global case, the plan is and functional checks that will take
used to gain an overview of the quality place after prototyping and before full
system for the complete company. In the production. This will include the
production case, control plans are product- additional checks needed before full
specific and deal with the checks and production validation. These may
balances used during production using include:
specific processes. Increased inspection levels (SPC or
The control plan: otherwise).
• Identifies the controls required to ensure Increased points for in-process checks.
product quality with a focus on any Assessment of machine and/or process
special characteristics defined by the capability.
customer. • The production control plan will often be
• Defines the reaction plans to be the shortest of all the control plans. This
implemented where nonconformities are is because the process will have been
identified, e.g. product containment and validated (see Section 10.5) and the
100% inspection to ensure that the checks that remain will be to ensure
process becomes stable and capable. control rather than capability.
• Is an output from the FMEA process. • Tip - All the stages of control planning
The control plan provides information on use the same basic format. The only
the process and product controls during difference is in the detail and the type of
production and for finished products. It control plan should always be noted.
also provides the essential guidance on
defect prevention at the operator level. Completing the control plan
Overall, the control plan is a The control plan should use the process
systematically compiled document which flow chart and the FMEA data (design,
summarises the quality planning and the process and machinery) to provide the
requirements for all significant basic information for the control plan and
characteristics of the product. It can be should cover:
used not only for control of internal • Incoming raw materials inspection
production but also as a reference points and requirements.
document by sub-contract moulders for the • Production inspection points (both in-
quality standards to be applied. process and final) and requirements.
• Tip - If external suppliers are being used • Off-line inspection points, e.g. laboratory
then the control plan must be supplied tests, and requirements.
as part of the supplier package (see The control plan is a
• Fit and function testing for mating parts. ‘living document’.
Section 9.3).
• Assembly inspection points and It should be used to
The control plan is a ‘live’ document and
requirements. identify and
must reflect any significant changes made communicate any
in the sourcing, processing or • Packaging requirements and inspection
changes to the
requirements of the product. points. process or the
• Tip - Control plans can be made at the • Final inspection points and product and the
requirements. associated changes
system, sub-system or component level.
to the controls.

• Tip - Standards requirements (both measurement of the process capability, Any inspection or
customer and regulatory/legal) should be then it is appropriate to reduce control activity costs
clearly identified on the control plan. inspection and monitoring. Equally, if money. The control
the process capability decreases then plan allows an
• Tip - If specific customer issues have assessment of the
been identified these should be identified increased inspection and monitoring may
overall cost of
in the control plan and be signed off by be needed. control (and equally
the customer. the risk and cost of
Reaction planning failure to control).
• Tip - Any special characteristics
(product or process) should be identified One of the key benefits of creating a
using the following key: control plan is that it forces consideration
of reaction planning, i.e. what you are
CTQ: Critical to Quality.
going to do when something goes wrong.
KPOV: Key Process Output Variable. Reaction planning should include all of the
KPIV: Key Process Input Variable. following cases:
USL: Upper Specification Limit. • Action when incoming materials are not
LSL: Lower Specification Limit. to specification.
• Tip - Where gauge testing, e.g. go/no-go • Action when processes are out of control
or similar, then gauges must be either through special or common causes.
Error proofing is
compatible with any used by the • Action when products are out of
more effective (and
customer. specification but the cause is not known. cheaper in the long
• Tip - The control plan is the ideal The actions to be taken must be planned term) than any
opportunity to review the control and documented for all the staff so that other type of control
suspect or nonconforming product is method. The control
methods for both process and product so
plan provides the
that they are robust and defensible. quarantined to prevent further processing opportunity to look
• Tip - If process improvements are made, or delivery. The reaction plan is the for opportunities to
and shown to be so through essential reference for risk-based error-proof the
thinking. process.
Control Plan
Prototype Pre-Launch Production Page: 2 of 5
Control Plan Number: Key Contact/Phone: Date (Orig.): Date (Latest Rev.):
CP 010 153 Robin Kent (034 146 457 697) 12/03/2015 28/10/2015
Part Number/Latest Change Level: Core Team: Customer Engineering Approval/Date:
C60-1101, Drawing Issue 5 (12/03/2015) MA, TM, MMcM, GL, DS. Not required. Internal production for stock supply.
Part Name/Description: Supplier/Plant Approval/Date: Customer Quality Approval/Date:
Main Frame: Compact 60 Michael Adams, 30/10/2015 Not required. Internal production for stock supply.
Supplier/Plant: Supplier Code: Other Approval/Date: Other Approval/Date:
Shuttleworth Extrusion N/A N/A To meet EN 12608.
Characteristics Methods
Part/ Process Name/ Machine, Special
Process Operation device, jig or Char. Specification or Evaluation or Sample Reaction plan
Number Description tool No. Product Process Class measurement Control method
tolerance Size Freq.
technique
Cease production and
investigate. Check
Calibration Automated
36 Calibration Temperature calibration bath water
Calibration of bath alarm for
3 (Process - bath water CTQ 12°C ± 1°C sensor and N/A N/A temperature.
extrusion (water temperature
FMEA) temperature alarm Contact maintenance
temperature) variation
to resolve.
Calibrators 1 to Cease production and

3: 25 -kPa ± 5 - Automated investigate. Check
Calibration 37 Calibration calibration bath
Calibration of kPA Vacuum alarm for
4 bath (Process - bath vacuum CTQ N/A N/A vacuum levels.
extrusion Calibrators 4 to sensor (6x) vacuum
(vacuum level) FMEA) levels Contact maintenance
6: 20 -kPa ± 5 - variation
kPA to resolve.
If chart indicates that

USL Micrometer Mean and the process is out of
Operator Profile Depth 60 ± 0.3 mm
45 - and and/or 5 Hourly range control control then take local
measurement (D) (EN 12608)
LSL Shadowgraph chart action to resolve.
Final extrudate Check production
7
dimensions since previous check
USL Micrometer Mean and
Operator Profile width 75 ± 0.5 mm for conformance to
46 - and and/or 5 Hourly range control
measurement (W) (EN 12608) specification.
LSL Shadowgraph chart
Control planning gives all the details necessary to show compliance with the customer
requirements
The control plan is a method of getting all the product and process requirements in one place to review the level
and extent of control in one place. It must include both the product and process controls as well as the full
reaction plan in the event that either of these indicate the potential for the production of nonconforming product.

10.4 Process setting
Getting the best from the start

The initial process settings selected at
T- T+
process design and development are key to
getting the best results from the process
but many companies do not realise the
effect that variability has on process
setting and the need to explore the
complete ‘process space’ when deciding on
these settings.
The effect of process variability Offset 1

The classic experiment to illustrate the
effect of process variability and
‘tampering’ on running production is the The process is set with a single reading. If this is located at the top of
Deming ‘Funnel Experiment’. This the distribution then it will appear that the process is ‘high’, even though
describes what happens when processes the process is actually centred and in control.
are constantly adjusted in response to the
natural process variations.
T- T+
The same type of effect is seen when Adjustment 1
initial process setting during process
design and development is based on very
few samples and, in the worst case, on a
single sample. Setters are often desperate
to get a process approved (the project
started late and they are trying to make
up time) and they may base the initial
process settings on very few samples. Offset 2
Consider the following series of events:
• The setter selects one sample (which
happens to be at the top of the process The process is adjusted down by the amount that it was ‘high’ - pushing
the process so that it produces outside specification. The process is
distribution) and finds that it is ‘high’.
measured and is ‘low’.
This is despite the fact that the process
is actually totally fine (see upper right).
• The process is adjusted downwards by T- T+
the same amount that it was running Adjustment 2
high - therefore pushing the process
outside the set limits so that it produces
parts outside specification. The process
is measured again but this time the
selected sample is at the bottom of the
process distribution. The setter obviously
finds that it is ‘low’ (see middle right).
• The process is again adjusted and this
time overshoots so that it is running
high and parts are again produced
outside of specification (see lower right). The process is again adjusted and this time overshoots so that it is
running ‘high’ and parts are again produced out of specification.
• The cycle continues and the process see-
saws in and out of specification whilst
the setter continues to adjust the process Do not set processes from a single sample!
in a futile attempt to find settings that
Process variability is similar to tampering and introduces
actually work.
excessive variation into the system. Setting processes from a
This situation happens day after day in single sample will always result in excessive variation and
industry as setters try to find settings that difficulty in finding stable process settings.

will allow them to sign off the process. such as evolution where the attempt to Setting sheets are
Even worse (perhaps) is when the setter achieve a ‘local maximum’ can prevent a valuable tool to
further progress to the ‘global maximum’ reduce process
selects a single sample that shows the
because any change results in a lowering variation. Failing to
process to be OK when it is mostly outside use them is a
the tolerance limits - the process is then of performance (for a more complete
mortal sin in terms
approved with settings that will inevitably discussion of local maxima in evolution see of quality
result in nonconforming parts being ‘Climbing Mount Improbable’ by Richard management.
produced when the process goes into full Dawkins, 1996, Penguin, 978-
production. 0141026176).
• Tip - To set a process you should never • Tip - Processes must be set using
rely on a single sample result. Always correctly designed experiments and
make at least 5 measurements and set studies - Taguchi methods and DOE and
the process on the basis of the average of scientific moulding studies are the best
the 5 measurements. for this (see Section 8.15).
• Tip - If you really are in a hurry then • Tip - Always ensure that the material
you can always try using ‘pre-control’ to used is the same as the production
account for process variability (see material and that the machine used is
Section 10.6). the same model that will be used for
subsequent production.
Searching for the local • Tip - The initial parameters should be
maximum used to calculate the control chart
parameters to allow SPC charts to be set
At the process design stage, processes are
up for future runs.
often set using a first ‘guess’, i.e. some
traditional settings that worked for a
similar product. The setters then play
with the settings for a while until they get
the ‘best’ settings for the process. This
first ‘best’ setting is assumed to be the
overall best setting and this is recorded as B
the full production settings. This may
work if the ‘best’ setting is actually the
overall ‘best’ setting but, in most cases, it A
will be a local maximum rather than the
system maximum (which can be very
different).
The figure on the right shows an
imaginary response surface with a local
maximum (A), and a system maximum
(B). If the first ‘guess’ based on traditional
settings is in the area of ‘A’ then changing
the settings, (often via the ‘blind man’s
walk’ approach) to increase the response
will result in the optimum settings being
those at which ‘peak’ performance is
encountered, i.e. A.
Any changes from these settings lead to a
decreased response and the process is
regarded as optimised at the ‘A’ settings
and the process is set to run under these
conditions.
This method both ignores and fails to
discover the higher ‘global’ maximum in
the area of ‘B’ where the settings could be
better for consistency and speed. Process setting by the ‘blind man’s walk’ method
Setting for local maxima on a complex
Having a process set on the parameters for the local maximum
response surface, e.g. most plastics
‘A’ means that any small process change will involve a decrease
processing methods, is always misleading.
in response and a subsequent return to ‘A’ as the ‘best’ setting.
It can prevent progress not only in plastics The global maximum of ‘B’ will remain undiscovered.
processing but also in many other areas

10.5 Product and process validation
Checking that you got it right validation and should be a ‘significant PPAP provides a
production run’. This is a production run logical and
Product and process validation is to thorough process
of between 1 and 8 hours (longer is better)
ensure that products made from the full for assessing the
with a minimum of 300 consecutive parts.
production tools and processes meet the readiness of a
This will obviously increase for multi-
relevant engineering requirements or process and
cavity tooling. product for full
specifications, that the processes are
capable (using Cp, Cpk and other SPC data) The significant production run should: production.
and can produce acceptable product • Be at the final production site.
consistently over time. • Use the proposed production rates/cycle
• Tip - Product and process validation times.
should also check that all the production • Use full production tooling.
documentation (control plan and process
• Use production gauges and test methods
flow chart) is valid and being followed.
from the pre-production control plan.
The product and process validation
• Use the proposed production materials.
procedure is called PPAP (Production Part
Approval Process) in the automotive • Use the production operators.
industry but in the medical or Trial production should be carried out on Some companies
pharmaceutical industries it is often called every unique production process in the (mainly German
IQ-OQ-PQ (Installation Qualification- and working to VDA
process flow chart. This includes all 2) require an ISIR
Operational Qualification-Process downstream operations, e.g. assembly and (Initial Sample
Qualification). Despite the different decoration operations. Only by looking at Inspection Report)
names, these are designed to give much the complete process can the effectiveness for acceptance of
the same output, i.e. process and process and efficiency be assessed. production.
validation. We will use ‘PPAP’ to • Tip - There is often pressure to make All of the data for an
represent the complete validation process. trial production short to make up for ISIR will already
The format of the output will vary with time lost in other parts of the process. have been
completed as a
the industry and is often closely defined by Resist this temptation at all costs. This result of the PPAP
the industry, e.g. the AIAG PPAP process. is the time to find out if the process is process (and more).
Whatever format is used, the end result
should be a documented process that can
reliably deliver the final product to meet Inputs from process Outputs to production and
the customer’s requirements. design and development feedback
Successful completion and sign-off of the (see Section 10.1)
PPAP process is a key event in the journey
to full production and can act as the signal Process flow chart. Trial production.
for several events: Process FMEA. Measurement systems
• Approval for full production and to ship evaluation.
Pre-production control plan.
production to meet customer orders.
Preliminary process capability
• Payments from the customer for tooling Process instructions.
study.
and/or machinery. Measurement systems
Production validation.
• Payments to suppliers for tooling and/or analysis plan.
machinery. Packaging evaluation.
Preliminary process capability
• Revision of supplier assessments by the study plan Production control plan.
customer.
Floor plan layout. Full-quality plan sign-off.
The financial implications of PPAP can be
very significant and getting it right is a Packaging standards and Production part approval.
specifications.
crucial milestone in process and product
development. The main actions that are
needed for PPAP are listed below.
The inputs and outputs of process and product
validation
Trial production After completion of the product and process design, the complete
Trial production is obviously one of the system needs to be validated to ensure that it delivers the
main parts of product and process required level of customer satisfaction.

robust. After PPAP is signed off you are Packaging evaluation This is the last
into production orders and fixing chance to get it
Most products will be packaged for right.
something up is much more difficult.
delivery to the customer, i.e. in ‘bulk’
• Tip - Pay attention to the ability of packaging or ‘display’ packaging and the Do not rush the
linked processes to match production product and
trial production should be used to evaluate
rates. If decoration cannot keep up with process validation
the ‘fit and function’ of the packaging that stage. After this,
production there is pain in the future. will be used to protect the product. you are into series
production and any
Measurement systems Production control plan changes are not
evaluation The results of the initial process capability
only difficult but
painful.
Trial production should be used to carry studies and production validation are used
out measurement systems evaluation (see to produce the production control plan.
Appendices 6 and 7). • Tip - The initial production control plan
will typically contain less detail than the
Initial process capability study pre-production control plan.
Machine capability data (see Section 5.4)
may have been carried out earlier in the Full quality plan sign-off
process development but trial production After the control plans and process flow
provides the extended production run that chart have been validated then the overall
is necessary to assess the overall process quality plan can be signed-off and issued.
capability (see Section 5.22). Initial
process capability studies should be Production part approval
carried out on all of the characteristics
Final production part approval will be
identified on the pre-production control
given by the customer (if the product is an
plan.
external development) or by internal
• Tip - These characteristics will have also release (if the product is an internal
been identified on the control drawings development). The status of production
(see Section 8.13). part approval can be one of:
The initial process capability study will • Approved for production.
reveal areas where the process is capable The output of PPAP
• Interim approval for production (subject should be complete
but may also reveal areas where changes
to further studies). documentation of
may need to be made to drawings, tooling, the process and
mating parts or downstream process • Rejected for production.
product. Use a
equipment. checklist to make
sure that all the
• Tip - In many cases, process capability •1. AIAG. 2008. Advanced Product Quality components have
studies are carried out on too many Planning and Control Plan. AIAG. been completed.
dimensions. To reduce the workload,
make sure that the control plan is
focused on those dimensions that are
important and ‘control’ the product.
• Tip - There will inevitably be non-
Production Part Approval
critical dimensions that do not require a

capability study. These values can be
simply reported (without capability
studies).
Production validation
In addition to the process capability
studies, there will be other tests, e.g.
material tests, mechanical tests and ‘fit
and function’ tests that must be carried
out away from the process area. This
testing must all be carried out on the trial
production to fully validate the process Where production part approval fits into APQP
and product.
Production part approval is the final link in the APQP process. It
• Tip - PPAP is concerned with the whole is a significant production run at the production site using the
product and not simply with the production facilities (including operators). This process is
dimensions and appearance. designed to validate the process at volume production levels.

10.6 Pre-control (zone charts)
Start-up control The pre-control rules Pre-control was

developed by
Pre-control (sometimes called ‘zone Start-up rules Satterthwaite at
charting’ or ‘stop-light’ control) is a Rath and Strong in
statistical process control technique that At start-up, the operator checks five
1954 and has been
differs from the conventional SPC consecutive samples for the required controversial ever
methods discussed in Chapter 5. In tolerance (dimension, weight, etc.) and since.
conventional SPC, the control limits are then applies the following rules:
It is with some
based on the performance of the process • If the five consecutive samples are in the trepidation that I
(the voice of the process). In pre-control, green zone, then the set-up is OK to run include it in this
the control limits are set based on the and the operator can move to the Workbook (and risk
‘running rules’. the wrath of the
specified tolerances (the voice of the
control chart
customer). There is a great deal of • If one of the five samples is in the yellow purists) but this is a
controversy over the use of pre-control (see zone then restart counting and sampling. Workbook for
Section 10.7) but at this stage we will • If two consecutive samples are in the industry.
simply describe the method and the yellow zone then adjust the process (if As a technique it is
significant uses. trained) or call for help to adjust the regarded as
Pre-control is very useful for working with process. statistically valid
short run processes where there may and can be a
• If one sample is in the red zone then useful tool.
never be time to collect all the relevant adjust the process (if trained) or call for
SPC data to set control limits and for help to adjust the process.
starting up for a process that has already
been qualified. It gives immediate • Tip - There are several slight variations
to these rules but those listed above are “Talk to the parts;
feedback on the state of the process and they are smarter
guidance on the next steps, i.e. when to reasonably well accepted.
than the
adjust or seek help, when it is safe to use This means that a start-up cannot be engineers.”
sampling, when the process needs validated and cannot take place until five
Dorian Shainin
examination and when to do nothing.
Pre-control limits ¼ ¼
tolerance ½ tolerance
The limits for pre-control (for bilateral band tolerance band band
tolerances) are set by dividing the total (7%) (86%) (7%)
tolerance zone into three zones as shown
on the lower right. These are:
• The middle half of the tolerance band is
the ‘green’ zone. For a process with a Cpk
of 1.00, i.e. the tolerance band is equal to
the process spread and the process mean
is the same as the nominal value, this Red Yellow Green Yellow Red
green zone will contain ≈ 86% of the
population.
• The two outer quarters of the tolerance
band between the green zone and the
upper and lower set limits are the
‘yellow’ zones. For a process with a Cpk of Nominal
Lower Upper
1.00, i.e. the tolerance band is equal to Set Set
the process spread and the process mean Limit Lower Upper Limit
is the same as the nominal value, these Pre-control Pre-control
Limit Limit
two yellow zones will contain ≈ 14% of
the population, i.e. 7% each.
On this basis the probability of getting The basics of pre-control
two consecutive readings in the yellow The complete tolerance band is divided into three areas: The
zone will be 0.0049 or less than 1 in 200. green zone (OK), the yellow zone (caution) and the red zone
• The areas outside the upper and lower (nonconformity). These zones are used to assess the process
set limits are the ‘red’ zones. status at start-up and during running.

samples in a row are in the green zone. If • Upper tolerance is maximum, e.g. tensile Pre-control charts
this is not the case, then the process is strength, burst pressure, then: are useful to
incapable of producing parts reliably quickly centre and
‘Green’ is from the best result
within specification (for whatever reason) qualify a process. It
recorded down to 3/4 of the tolerance should not be used
and process improvements are required to band. for significant
improve the process capability (Cp). process adjustment
‘Yellow’ is from 3/4 of the tolerance
If the operator attempts to continue without a full
band down to the lower tolerance
production without meeting the ‘five process
limit. investigation.
greens in a row’ then every part will need
• ‘Red’ is less than the lower tolerance
to be checked.
limit.
Running rules
Running under pre-control is equally
Product not process
simple. The operator checks two Pre-control differs significantly from
consecutive samples for the required conventional SPC in that it looks more at
tolerance (dimension, weight, etc.) and the product than at the process and sets
then applies the following rules: the process limits based on the ‘voice of
• If both samples are in the green zone or
the customer’ (the tolerances) rather than
one is in the yellow zone (upper or lower) the ‘voice of the process’ (the actual
and the other is in the green zone then process capability). This makes it more I think that most
continue production. closely aligned to conventional product readers will have
inspection (with all the defects that this understood the
• If both samples are in the yellow zone has) than to SPC but it is sometimes more concept a lot
(on the same side) then adjust the compatible with operations in most quicker than the
process (if trained) or call for help to concept of control
plastics processors. charts.
investigate and adjust the process.
• If both samples are in the yellow zone That is why it is in
(on different sides) then ‘stop’ and call this Workbook.
for help immediately. This indicates a
potential change in the process
capability and requires more
investigations and action than can be
taken by an operator.
• If any sample is in the red zone then stop
and call for help to investigate and
adjust the process (if trained) or call for
help to investigate and adjust the
process. Inspect all parts since the last
check.
• Tip - The original pre-control rules call
for adjustment but there is legitimate
concern that this may lead to excessive
tampering with the process. Plastics
processing rarely has a single ‘dial’ that
you can change to adjust the process and
in most cases there will be a need for
more process investigation rather than a
simple adjustment.
One-sided tolerances
If the tolerance is not bilateral then the
rules for one-sided tolerances are:
• Lower tolerance is zero, e.g. ovality,
flatness, then:
‘Green’ is from zero to 1/2 of the The start-up and running rules for pre-control
tolerance band.
The rules of start-up and running using pre-control are very
‘Yellow’ is from 1/2 of the tolerance simple and easily understood. It operates by checking that the
band to the full tolerance. product is being produced in the green zone and meeting the
‘Red’ is greater than the full tolerance. requirements of the customer.

10.7 Inspection, pre-control or control charts?
What do we want to control? Pre-control (product and

In the introduction to SPC (see Section process control)
5.1) we discussed the difference between
simple product inspection (defect detection
The advantages
and after the event) and process control • Pre-control is useful in ‘short-run’
(defect prevention and before the event). processes where the batch size is small
The conclusion was that defect prevention and control charts are difficult to set up.
was a far preferable strategy and this • Pre-control can be used to rapidly check
remains the case. start-up for the correct location of the
Pre-control is somewhere between these process.
two extremes of simple product inspection • Pre-control gives clear guidance on when
and the more complex control charting of to adjust a process or make tool changes.
SPC. It compares individual product • Pre-control is easily understood and used
values against the specification limits but by operators.
uses these to make a judgement on the
• Pre-control is simple to operate, i.e. the
capability and setting of the process
rather than simply accepting or rejecting
the inspected products.
The traditional process of inspecting
either single or multiple samples (without
a sampling system - see Chapter 6) for
simple conformance to specification is very
limited in effectiveness and efficiency.
Inspection (product control)

The advantages
• Inspection is easy to carry out.
• Inspection is initially inexpensive to
carry out.
• Inspection is easy to understand. The inspection and pre-control view of loss
• Inspection is easy to schedule. Inspection and pre-control treat all parts that meet specification
• Inspection does not need calculations or as 100% ‘good’ and parts outside the specification as 100%
software. ‘bad’. This is the conventional view of loss and is useful for
production and accounts but is not reality (see Section 8.15).
The disadvantages
• Inspection tells you about the inspected
product but not about the other products
being produced or about the process.
• Inspection gives no information about
the process spread or location.
• Inspection does not separate common
and special causes.
• Inspection provides no reliable guidance
for process adjustment. In fact, process
adjustment based on limited inspection
can be detrimental to outgoing product
quality (see Section 10.4).
• Inspection can easily lead to excessive
The control chart view of loss
tampering with the process. Conventional control charts are similar to the Taguchi Loss
• Inspection does not provide any function (see Section 8.15), i.e. the further away from the target
information or driving force for process then the ‘worse’ the product is and the greater the loss. The aim
improvement. is to get on the target with minimum variation.

target is to stay in the middle of the • Control charts are accepted by all types As with anything
specification. of customers. else, you need to
choose your
• Pre-control is inexpensive to operate, i.e. • Control charts provide a written record
weapons and tools
it does not require expensive calculations or chart. carefully.
or charting software. • Control charts provide a common
• Pre-control can be used with attributes language for all levels of staff.
(gauging) to start a process.
The disadvantages
• Pre-control has a low risk of rejecting
• Control charts are based on the process
parts or processes, i.e. <2% in the worst
conditions and 0% if Cpk > 1.66. limits (‘the voice of the process’) rather
than the specification limits (‘the voice of
• Pre-control works from the specification the customer’) but the process should
limits (‘the voice of the customer’) rather never take precedence over the customer.
than the process limits (‘the voice of the
• Control charts work best when the
process’).
process capability is high.
The disadvantages • Companies can suffer from control chart
• Pre-control needs good and realistic ‘overload’ where control charts multiply
specification limits. and take over the production area.
• Pre-control needs a stable process that • Control charts need trained operators.
can operate well within the specification • Control charts often use expensive
limits, i.e. a process with good capability. software or time-consuming calculations
• Pre-control does not distinguish between to operate.
common and special causes and is • Control charts are best in long-runs
therefore not useful for process where the process has a chance to
improvement, i.e. it provides no become stable for short runs they can be
incentive to improve the process because difficult to operate.
all good parts are treated as equally
good. Summary
• Pre-control is not useful where the Having described and discussed
process has a poor capability and pre- inspection, pre-control and control charts,
control can signal process changes when it should be evident that inspection is not
these are not really needed. adequate for modern plastics processing
• Pre-control is best used for processes but that both pre-control and control
with low part-to-part variation and can charts have strengths and weaknesses.
increase process adjustments if part-to- Unfortunately, many plastics processors
part variation is high, i.e. there is some have not got beyond product control
danger of false signals. (inspection) - the least useful technique of
• Pre-control does not provide a record of all and need to move forward significantly
the process, i.e. there is no requirement and quickly.
for a written record or chart. Those plastics processors who have taken
• Pre-control is not accepted by some the initiative and moved into process
major automotive companies as being control have generally done so as a result
acceptable for process control. of customer pressure and have been forced
into control charting by the same forces
Control chart (process control) that forced them into ISO 9001, i.e. ‘if you
don’t comply then we won’t buy’. They
The advantages have rarely critically examined control
• Control charts allow process variation to charting or even considered pre-control.
be separated into common and special Despite this, pre-control is a technique
causes and the causes to be treated that can offer, almost, the best of both
separately. worlds and a technique that should be
• Control charts use common causes to set seen as complementary to control charts
the control limits. rather than an enemy.
• Control charts give clear guidance on • Tip - If you have never used pre-control
when to adjust a process and when to then try it on some products that have
leave it alone. good capability and see if it works for
• Control charts drive process
you.
improvement (if used properly).

10.8 Engineering process control (EPC)
Dynamic changes result and the desired target (or set SPC can detect
point). This type of control uses a special causes
To date we have focused on a variety of quickly but EPC can
‘summing junction’ to calculate the
SPC techniques to reduce process output keep a process on
difference between the output result and
variability by detecting and eliminating target. They are
the target and make the required change
assignable causes of variation (special very different
to the input variable, e.g. a thermocouple things.
causes). In SPC, unless the process shows
measuring barrel temperatures can be
a warning then it is left alone to avoid
used to control barrel heaters. Closed
‘tampering’.
loop control with feedback works very
Engineering Process Control (EPC) takes effectively for processes where there is
an alternative view and seeks to reduce
process output variability by regularly
measuring a variable and then
continuously adjusting one or more input
variables to compensate for any variation.
In EPC the process is being continuously
changed in response to the measured
process output.
SPC was developed largely by statisticians
and quality engineers but EPC was
developed largely by process and control
engineers and for many years these two
different groups approached variability
very differently.
The fundamental difference between the
two approaches is shown in the flow
diagrams on the right.
This section introduces EPC as a concept
and Section 10.9 shows how the two
concepts can potentially be integrated to
reduce process variability and improve
quality at the same time.
The basics of EPC

In the simplest form, EPC makes dynamic
changes to an input variable to control the
resulting measured output variable. EPC
therefore needs a model of the dynamic
behaviour of the process, i.e. which knob
to turn to adjust the process.
The control used by EPC can be:
• Open-loop control - this is where the
adjustment to the input variable is by an
operator taking action based on the
difference between the output result and
the desired target. If the operator takes
no action then the system can run out-of-
control. This is the least used of the EPC
control systems for obvious reasons.
• Closed-loop control with feedback - this The difference between SPC and EPC
is where there is no operator SPC and EPC are different in their approach to process control
intervention and the process control because they come from different developers. The downside of
system automatically takes action based EPC is that it does not look for special causes (it adjusts the
on the difference between the output process) and therefore does not fit the concept of improvement.

only a small time delay for the change in simple cases, there is rarely a single ‘knob’ EPC has been
the input variable to take effect to turn to change the process output, i.e. described by some
changing any single process input variable (harsh) critics as a
• Closed loop control with feedforward -
can often have multiple and sometimes ‘bandage to cover a
there is, again, no operator intervention wound not to cure
and the process control system unpredictable effects on the output result.
it’.
automatically takes action. However, in This is because most plastics-forming
this case the action is based not on the processes are ‘multivariate’, i.e. there is
output result but on the disturbances to more than one variable involved in the
the system. These disturbances are process. Multivariate systems are much
measured (before the process output) more difficult to control, even with SPC
and changes are made to the input where it is the output that is being
before the output is affected or even measured and manually controlled. With
measured, e.g. measuring the moisture EPC it is the system model, however
content of a plastic and adjusting the established, that is trying to make
processing conditions before processing multivariate adjustments and this is a
One of the major
to prevent moulding concerns. Closed more complex task. disadvantages of
loop with feedforward is most useful Previously, this was extremely difficult EPC is that it
where there is a long time delay for the but advances in both computing power corrects the
process to reduce
change in the input variable to take and in polymer rheology now make EPC a
the effect rather
effect and where it is possible that faulty serious contender for controlling plastics- than seeking to
product could be produced for some time forming processes to ensure that processes remove the
before the correction was effective. remain in control to produce high-quality underlying cause of
• Cascade control - this is where multiple products. the problem.
control loops, either feedback or
feedforward, are used to control complex Open loop control
processes.
Disturbance to system
These control options are shown in the
diagram on the right. In practice, most
current systems for plastics processing Input System Output
will use closed-loop control with feedback
because of the difficulty in setting up the
necessary model for effective control using Manual input
feedforward systems. adjustment
The controller used in EPC will affect the
response of the system and controllers can
use a variety of techniques such as: Closed loop control with feedback
• Proportional control (P) - where the Disturbance to system

applied correction is proportional to the
measured difference between the output Input System Output
result and the desired target.
• Integral control (I) - where the applied
correction is proportional to the time Automatic input
integral of the measured difference adjustment from
between the output result and the Summing feedback
desired target. junction
• Derivative control (D) - where the

applied correction is proportional to the Closed loop control with feedforward
rate of change of the measured difference Disturbance to system
between the output result and the
desired target.
In practice, a combination of these control
Input System Output
methods is generally used and one of the
most common is the PID (Proportional/
Integral/Derivative) control. The control options for EPC
Problems in plastics Control under EPC can be open loop, closed loop with feedback
or closed loop with feedforward. Closed loop with feedback is the
One of the major problems with EPC in most common system due to the difficulties of establishing the
plastics processing is that, except in model necessary for closed loop with feedforward.

10.9 Integrating SPC and EPC
Can we use both? Conventional SPC EPC is rapidly

developing into
SPC and EPC were developed separately The simplest approach is to use SPC to
APC (Automatic
and in separate industries and the main verify that the process/product is in Process Control)
differences between the two approaches control and, if any of the traditional SPC where the system
are shown in the table on the lower right. ‘out of control’ signals are seen, then to controls itself to
Despite this, they are largely adjust the process based on a set of EPC monitor and control
complementary and look at different rules. Successful corrective actions are the process.
aspects of controlling the quality of the stored as the rules for action in the future
output. In the past, the lack of computing if SPC signals a special cause. This is only
power, the lack of reliable sensors and a a slight variation on standard SPC where
poor understanding of polymer rheology corrective action is not explicitly stored to
restricted the development of EPC for provide a basis for future action.
plastics processing but these issues have
largely been removed. The concept of
integrating SPC and EPC to improve SPC EPC
quality is therefore rapidly gaining
Developers Statisticians and Process and control
ground. Integration of the two techniques
quality engineers. engineers.
offers the potential to use SPC to monitor
the process for special causes and EPC to Industry basis Product industry. Process industry.
reduce the effect of measurable process
variations. This idea has gathered much
Function Process monitoring Process adjustment
support and research with very good
to control product. to control process.
results but implementation in industry is
currently very poor, i.e. most plastics View of process Process is either in Process is
processors are still at the inspection stage control or not. continuously
let alone contemplating the integration of adjusted to stay at
SPC and EPC. target value.
The approaches to integration Process changes None unless output Continuous process
EPC was first introduced to statisticians warning signal seen. changes made in
and quality engineers by MacGregor1 in response to output
1988. There was initially considerable signal.
controversy about the idea of ‘continuous
tampering’ with a process that was in Process model Not necessary Good process
except when model needed to
control even if this approach had been
process is out of make adjustments.
used by process and control engineers for
control.
many years and had a long history of
success. EPC has historically been used Expects Process to be Continuous process
extensively to control processes, e.g. a stationary (centred). drift from centre line.
simple temperature controller on a barrel
heater typically uses closed loop control Measures Quality features. Process features.
with feedback and a PID controller and
the EPC ideas should not come as a Timescale Medium- to long- Short-term control.
surprise to any engineer. The issue was term control.
really how to convince the statisticians
and quality engineers that these ideas Technology Low. High.
could be used to control the complete
process and product and be successfully Results Process Process
improvement. optimisation.
integrated with SPC. Apart from simple
process control, e.g. barrel heaters,
progress has been slow in the plastics The differences between SPC and EPC
industry where this approach has rarely SPC and EPC attack the ‘quality problem’ from different but
been trialled. complementary angles and there are obvious benefits to both
There are several basic ways to approach approaches. This has lead to attempts to integrate the two
integration and these are: approaches to give both improved quality and control.

Algorithmic SPC (ASPC) conventional SPC methods. The process Integrating SPC
This approach combines an algorithm changes are made by EPC controllers. and EPC is still at
the very early
(model)-based controller and SPC This type of approach minimises the stages in plastics
techniques. It uses EPC to reduce effects of ‘continuous tampering’ but still processing but
predictable variations and then monitors needs a good process model to define the offers the ‘best of
the system using SPC. Instead of using control actions. both worlds’ with a
SPC to monitor and making adjustments future of self-
adjusting machines
only when necessary, ASPC turns this Integrating SPC and EPC in automatically
around by automatically making plastics processing producing quality
adjustments and then monitoring to products.
Integrating these complementary
ensure that they are effective.
approaches in plastics processing needs: That is why it is
There are some concerns about ASPC and important.
• A good process algorithm to control the
the main issues are:
process - these are now available for
• What are the variables that are most processing methods.
monitored to assess control? This is not
• A good correlation between the input
only about the actual variables but about
variables and the output variables -
whether they should be output variables
multivariate approaches to plastics
(quality features) or input variables
processing now give much better
(process features).
correlations for plastics processing. “All models are
• How accurate does the process model
This progress in understanding the basics wrong but some are
need to be for effective control? useful.”
of plastics processing means that the time
• Should the controller be automatic or has come to start this integration to George Box
manual, i.e. open loop or closed loop? improve process control and quality.
This type of approach has been used This is shown by the fact that in
successfully for multivariate systems September 2015 KraussMaffei released
using multivariate SPC and shows a great Adaptive Process Control (APC) to control
deal of promise for plastics processing. IMMs (www.kraussmaffei.com/en/new-
Note: When SPC is applied in conjunction adaptive-process-control-apc.html). This
with EPC it is most common to use an uses data from a series of sensors to
Exponentially Weighted Moving Average continuously monitor and adjust the Machine and
chart (EWMA chart). EWMA charts are machine changeover point and holding plastic
beyond the scope of this Workbook. pressure (even during the shot) to variables?
dramatically reduce the shot weight John Bozelli, in an
Active SPC excellent article
spread and improve part quality. This is
SPC is traditionally an ‘off-line’ process, just the beginning. (Plastics
i.e. samples are taken and measured off- Technology,
September, 2015),
line. Active SPC takes SPC ‘on-line’ and How effective is integration? makes the vital
uses a process model to define the control
There is limited reporting of integrating point that we should
limits and the necessary control actions. consider the plastic
SPC and EPC in plastics processing but
The on-line SPC results are used to decide variables and not
Montgomery et al.3 showed that the
if action is required, i.e. if any of the the machine
proper use of an integrated SPC and EPC
traditional SPC ‘out of control’ signals are variables for
control system ‘can always outperform the successful
seen, and then takes action following the
use of either alone’. This was a purely moulding.
rules of the process model.
theoretical paper using simulation results
This type of approach minimises the Machine variables
but validates the use of EPC for control will depend on
effects of ‘continuous tampering’ and has and the use of SPC to search for special valves, voltages,
been used for many continuous processes causes. temperatures and
although applications in plastics set points. These
processing are rare. •1. MacGregor, J.F. 1988, On-Line Statistical will also suffer from
Process Control, Chemical Engineering Proc- variability and this
Run-to-run (or run-by-run) ess, Vol. 84 (No. 10), pp 21‑31. will vary the
experience of the
This approach has mainly been used in •2. Box, G.E.P. All of Box’s work in this area is plastic in the
the semi-conductor industry where a ‘run’ valuable and worth reading, even if it is very process and the
is a group of products, e.g. a batch or a high level. See www.stat.wisc.edu/people/ results.
group of products produced under the george_box/publications for a list of these.
EPC mainly works
same conditions. In this approach, •3. Montgomery, D.C.; Keats, J.B.; Runger, on the machine
changes to the process conditions can only G.C. and Messina, W.S. 1994, Integrating variables so maybe
be made at the end of a run and the need SPC and EPC, Journal of Quality Technology, there is some extra
for changes is detected/signalled by Vol. 26 (No. 2), pp 79-87. work to do here?

10.10 Injection moulding - scientific moulding
Taking the art out and putting control valve gates for multi-cavity tools. The theoretical and
the science in practical setting of
Pressure-controlled packing IMMs has
Design of experiments is one of the key After changeover, the packing pressure is advanced hugely in
tools for finding the system maximum (see used to fully compress the plastic as it the past 20 years.
Sections 8.15 and 10.4) and there has been ‘Scientific
cools and shrinks in the mould. Changes
some excellent work carried out in this moulding’ is a
to the packing pressure will control the general term for
field but it is not the only way to improve compression of the melt along the flow using a defined set
quality in injection moulding. One of the path. Insufficient packing pressure can of tests to fully
greatest advances in injection moulding in result in voids, short shots, excessive determine the
the past 20 years has been the use of what shrinkage on ejection, warping and other moulding settings
is termed by some as ‘scientific’ moulding in a logical manner.
filling defects.
where the injection moulding process is
effectively broken down into three stages: Cooling
• The filling stage of ≈ 95% of the volume The cooling stage, whilst one of the longest
stages, is simply waiting for the material Scientific moulding
of the mould which is controlled by the is sometimes
ram/screw velocity. to cool sufficiently for successful ejection referred to as
without significant post-ejection warping. ‘decoupled
• The packing stage of the remaining ≈ 5%
of the volume of the mould which is • Tip - If the cavity is fully filled then the moulding’ - this
experience of the melt during cavity term is a service
controlled by the packing pressure. mark of RJG Inc.
filling (temperature and pressure) is a
• The cooling stage after gate seal where The company set
major factor in product quality. up by Rod Groleau
no more material can enter the mould
and the packing pressure can be released • Tip - If a cavity pressure sensor shows a in 1985 and now a
significant drop at ejection then it is world leader in
as the product is simply cooling before
sensor and
ejection. likely that the gate has not yet fully
moulding
sealed. technology.
The recognition that these were separate
stages allows separate optimisation of the
stages to give better process control. An
idealised graph of the cavity pressure and
the hydraulic pressure (for a hydraulic
machine) is shown on the lower right. The
Cavity pressure / Hydraulic pressure
curves are not aligned because of time

delays in signalling and valve switching.
Note: The curve will be different for
amorphous and semi-crystalline plastics.
The stages of moulding

Velocity-controlled filling
This controls the majority of the cavity
filling and sets the basic shape of the
product. Changes to the fill speed can be
used to control orientation, crystallinity,
appearance and other aspects at the
surface of the product. Excessive speeds
can result in jetting (in the early stages),
flash formation and even mould damage.
After the mould is nearly filled, control is
changed from velocity control to pressure
control. This changeover point will be set An idealised cavity and hydraulic pressure curve for
using a closed-loop system based on a
injection moulding (not to scale)
signal from screw position, hydraulic
pressure, cavity pressure or cavity The cavity and hydraulic pressures will change throughout the
temperature. cycle, most noticeably near the changeover point and as the gate
seals. After gate seal, the hydraulic pressure can be removed.
• Tip - Cavity pressure is also used to

The scientific setting process and mould for all runs. Scientific moulding
• Never be an optional practice - failure to is a complex
Scientific moulding uses the separate subject and the
stages to set the moulding parameters to follow the setting sheet should be a
reader is referred to
fall within the Moulding Area Diagram disciplinary offence (it is money). one of the
(see right). This is also sometimes termed • Not stop at the settings of the main references given for
the ‘process window’5 as it defines the machine but should include all checks further details.
basic process conditions for successful and settings for all ancillary equipment. This Workbook
processing. Scientific moulding looks at • Tip - If settings can be recorded directly concentrates on the
variations in process and material quality aspects of
on the machine then they should be set
properties by studying: plastics processing
(and locked out) to prevent unauthorised and scientific
• The effect of processing speed (shear changes. moulding is a
rate) on material viscosity - this is to • Tip - Setting sheets should include a valuable tool to
select an injection speed which gives improve injection
‘Special Notes’ section for details of
consistency of the melt viscosity with moulding quality.
safety concerns or notes, start-up and
changes in injection speed. shut-down procedures and any details of
• The effect of pressure drops on the purging to be carried out.
process - this is to avoid a ‘pressure- • Tip - Setting sheets reduce set-up times
limited process’ where the machine is not Machine and
by allowing machine parameters to be plastic variables
able to deliver the material at the set rapidly.
required injection speed.
(again)?
• The time to gate seal - product weight An article by John
Bozelli (see Section
versus time shows when the product •1. Goff, J.R. 2013. ‘Moulding Masterclass’. 10.9) shows the
weight stops increasing and if the PID. effect of both the
packing time is long enough. machine and plastic
•2. Routsis, A. 2015. ‘Injection Molding variables on a part.
Additional studies can also be carried out Reference Guide’. An excellent general guide
on: that is available free from Any setup sheet will
• The cavity balance of the specific mould - www.traininteractive.com. list the machine
variables, e.g.
this is to validate cavity-to-cavity fill •3. Bozelli, J. 2016. Various publications free injection speed.
consistency in multi-cavity moulds. from www.scientificmolding.com. These are machine
• The effect of processing conditions on the •4. Kulkani, S. 2010. ‘Robust process parameters and not
product cosmetics - this is to set the development and scientific molding’. Hanser. the plastic
parameters which
conditions to maximise the process
•5. Kazmer, D. 2009. ‘Plastics Manufacturing are the ones that
robustness in terms of cosmetics. Systems Engineering’. Hanser. really count.
• The cooling time - this is to determine
the shortest cooling time to produce an
acceptable part.
Scientific moulding has many quality
advantages for processors:
• It provides a robust process that is more
Injection speed (mm/s)
consistent in production - this improves

product consistency and therefore
quality.
• It provides a disciplined and
scientifically valid approach to solving
injection moulding quality issues.
• It introduces science to what was, for
many, a ‘black art’ based on experience.
Setting sheets
The output of a scientific moulding process
is a setting sheet which fully defines a
robust set of moulding parameters but
they have to be used to be successful.
A Moulding Area Diagram (MAD) or process window
Settings should: The MAD summarises the conditions that will give acceptable
• Be recorded on machine and mould mouldings.
setting sheets. After Kazmer5
• Always be used to set up the machine

10.11 Injection moulding - multivariate analysis
A multivariate process -3σ +3σ

Whether you use scientific moulding or
not (and we would always recommend that
you do), at some stage you have to
determine and fix the settings of the IMM.
This also means that you have to
determine the critical parameters to
control during the process.
One of the main issues with controlling
injection moulding is that the process is a
multivariate process, i.e. there is no single
‘knob’ to turn to manage the process and -3σ +3σ
the variables are related. This issue can be
reduced by using simple DOE (see Section
8.15) to improve the robustness of the
process, i.e. to make it less sensitive to
changes in the variables. However, simple
DOE often does not fully consider
interactions between the factors being
considered and the issue of a multivariate
process still remains.
+3σ
The best work in multivariate analysis

(MVA) for injection moulding is being
carried out by Kazmer and his group at
-3σ
UMass Lowell and they have published

many excellent papers on this subject in
the last 20 years.
• Tip - Search the internet for any of the
papers on injection moulding by Kazmer.
There is some excellent information on -3σ +3σ
the moulding process and particularly
the work that has been carried out with
Hazen at MKS Instruments.
Dependent or independent
variables?
The simple case of measuring a single
variable using SPC is shown on the top
right. In this case, the control limits are
+3σ
set at X ± 3σ and these limits will contain

C
on
99.73% of the population, i.e. there is a

tro
ls
single set of control limits for the process.

pa
ce
If the variable is within these limits then

-3 σ
the process is regarded as being in control.

If two independent variables are involved
then this can be considered as shown on
the middle right. In this case, the control
limits for the independent variables are The effect of dependent variables on control limits
both set at X ± 3σ and these limits
With a single variable there is a single set of control limits. For 2
effectively define a square or rectangular
independent variables there is a rectangular control space but for
control space. If both variables are within
dependent variables the control space is an ellipse that is smaller
these limits then the process is regarded than the rectangle.
as being in control.
After Hazen1

If two dependent variables are involved • Tip - Find out about using MVA with Validation of
then this can be considered as shown on DOE to set and run moulding machines processes using
the lower left. In this case, the control to get the best out of the process. The MVA and DOE
limits for the two dependent variables are information is out there and waiting to works but the
investment in staff,
both set at X ± 3σ but, because of the be used.
training and
interdependency of the variables, the equipment has
control space is no longer a square or resulted in very
rectangle but an ellipse. This reduces the limited take-up by
allowable control space and it is possible industry.
for both variables to be nominally ‘in
control’ but for the actual process to be •1. Hazen, D. 2009. ‘How better fault detection
‘out-of-control’. methods improve molding quality’. Injection
Moulding magazine, UBM, December 2009.
This is because the control space is
defined not simply by the values of the two •2. Hazen, D. and Kazmer, D. 2007. ‘Using
variables but also by the interaction and multivariate process analysis to predict
injection molded part quality’. Injection
correlation between the two variables. Moulding magazine, UBM, October 2007.
A moulder is then left with three choices: Available from MKS Instruments,
www.mksinst.com.
• Set the process window so that it lies
entirely within the elliptical multivariate •3. Hazen, D. 2012. ‘The Optimization of
control space. This is the cautious option Injection Molding Processes Using Design of
and will reduce the control limits but Experiments’. MKS Instruments Application
Note 06/12-7/12,USA), www.mksinst.com.
will also cause multiple false alarms
showing that the process is ‘out-of-
control’. Taking time to save time - traditional but it works.
• Set the process window as for the
Whilst working with an in-house moulder some years ago we had
independent variables control space.
a ‘problem’ tool with an average scrap rate in the region of 25%.
This is the standard option for many The mould was never capable of running for more than 2 hours
companies but will allow the process to without having to be seen by the setters and the output was
continue producing when it is actually always inconsistent and of poor quality. An ideal candidate for
‘out-of-control’. improvement!
• Use multivariate analysis (MVA) with
We spent some time on the basics:
DOE to set the process limits to those
which reduce false alarms, reliably Finding out the best settings (using capability studies and a set
indicate if the process is ‘out-of-control’ of properly designed experiments).
and produce good products. Recording the settings on the setting sheet.
This will typically use Principal
Training the operators in SPC and ‘when to do nothing’.
Components Analysis (PCA) to reduce
the size of the data set and to develop a Establishing the SPC limits and checking protocols.
multivariate model of the process.
The simple control was to measure the weight of the output and
In reality, for injection moulding, there use this for a control chart. Operators were warned that
are more than two dependent variables tampering with the controls whilst the chart showed the process
and traditional variables such as injection to be in control was a disciplinary offence.
speed, packing pressure and melt
The product ran for around 56 hours with absolutely everything
temperature will all depend on each and
running fine and no setting changes by anybody. This was
interact, i.e. it is even more complicated regarded as a true miracle by all concerned but we knew it
than shown in the diagram on the left. couldn’t last.
The use of MVA, PCA and DOE to
Finally the control chart indicated ‘out-of-control’ and we all
optimise the injection moulding process is
panicked. We checked the settings, the material, the control
outside the scope of this Workbook and is
charts and everything we could think of. We were just about to
best described in publications that are
give up and modify the settings when the Senior Tool Setter
available from MKS Instruments 2,3 which found the problem. The cleaner had moved a rubbish bin at the
describe this process in detail. back of the machine and accidentally kinked a cooling water
This process has shown excellent results hose. This restricted the flow and the cooling of the mould. We
in practice, e.g. decreasing cycle times by straightened the hose and waited. After 20 minutes the chart
up to 10% and reducing rejects by up to indicated that the process was ‘in control’ once more and the run
5%, but the disappointing aspect is the continued.
lack of real utilisation by industry of a The process does work but sometimes thought is needed before
proven process. action.

10.12 Injection moulding - managing variables
Too many variables • The changeover point from speed to Critical process
pressure control - cavity pressure is a parameters such as
Most injection moulders will not use MVA the integral of the
sensitive indicator of when changeover
with DOE for process control and will use cavity pressure
should take place (or has taken place).
some other control system, e.g. SPC or curve and the melt
simple random inspection. SPC whilst • The peak pressure in the cavity. temperature curve
valuable for many applications, measures • The gate freeze point. are very sensitive in
and monitors a single variable and for detecting process
Temperature sensors can be mounted at variations.
multivariate processes can be misleading the end of the flow path to show and
as it does not consider dependent Quality measures
control:
variables (see Section 10.11). The other from the finished
• The presence of the flow front - this can moulding still need
disadvantage of SPC is that the variable
be used to control changeover, i.e. to to be correlated to
to be monitored must be specified and in the process
indicate filling of the cavity.
injection moulding there are simply too parameters, e.g. it
many variables. We need a way to make • The cooling of the cavity during the needs to be shown
this simpler. cooling phase. that the integral of
Real-time cavity pressure and the cavity pressure
The cavity is the thing does indicate an
temperature measurements provide the optimised process.
The major controller of product quality in essential information for process control However, after this,
injection moulding is the mould itself. The because these measures are very sensitive the effort should be
mould not only defines the product to changes in either the machine focused on
dimensions but acts as a ‘heat exchanger’ conditions or the material. Consistent monitoring the
cavity pressure and temperature curves process and not on
to control cooling. Fill the mould under
monitoring the
consistent conditions with a consistent through the cycle, as measured by the
moulding.
amount of plastic and you are more than integral of the curves, will allow consistent
halfway towards a quality moulding. moulding and part production from:
• Tip - Machine conditions may appear • Run to run.
constant but it is the cavity conditions, • Cavity to cavity.
Changing materials,
i.e. what the plastic experiences, that • Machine to machine. e.g. increasing the
counts. regrind percentage,
If good products are produced for defined,
In the past, it was difficult to control the should be treated
maintained and consistent cavity pressure as a change in the
process using real-time variables due to and temperature curves then moulders machine settings.
both their rapid changes and their inter- are assured that the final product will also Varying the MFI will
relationships but it is now possible to be consistent, even if the mould is change the cavity
measure real-time cavity pressure and transferred to another machine or if the pressure and
temperature data. This can be used to material changes, e.g. an increase in change the
both validate each moulding and to control moulding
regrind percentage. These curves show parameters needed
machine parameters for consistent mould what the plastic is experiencing rather for a good
filling based on the p-V-T relationship for than simply the machine variables. moulding.
the plastic being moulded.
• Tip - Cavity pressure alarm limits can
Cavity pressure and temperature are both be set to alarm when not in tolerance to
sensitive to and proportional to the indicate poor-quality parts.
resulting part weight and dimensions
(assuming shrinkage is allowed for) and • Tip - Cavity pressure and temperature Unlike extrusion,
are much more sensitive than screw measuring the
provide the information necessary for
position for determining shot size and current (amps)
process and thus part control. drawn by an IMM
cushion.
Cavity pressure sensors, e.g. Kistler, does not provide
Priamus and RJG, can be mounted near • Tip - In the past, thermocouples were much useable
the gate to show and control: relatively slow in indicating temperature information.
changes but new thermocouples are The current drawn
• When the flow front has passed the
much faster in registering temperature varies widely during
sensor. changes. the cycle and much
• The rate of fill of the cavity - this should of the energy used
• Tip - Fast-acting servo valves allow
be as uniform as possible to avoid is due to machine
better control of changeover and machine losses and not due
warping and the cavity pressure should control.
increase linearly during fill. to processing loads.

• Tip - For basic quality control, the
maximum values of cavity pressure and We don’t have all that stuff!
temperature can be monitored instead of Although we are strongly in favour of using cavity pressure and
the integral of the curves. temperature to control the injection moulding process, many
processors simply do not have the time, money or staff to invest
Temperatures in and operate process control tools such as cavity pressure or
The mould temperature and melt temperature measurement or to set up and use MVA/DOE for
temperature are two other measures that establishing robust process settings. In cases such as these then
have a large effect on moulding quality. what are the options?
The first option is to manually measure the key dimensions
Mould temperature
identified on the control drawings (see Section 8.13) and to use
The mould temperature is probably the standard SPC variables charts (see Section 5.18) based on
most important temperature in terms of these dimensions.
quality and this affects quality in many This approach needs a lot of operator time, relies on the
ways. The mould temperature is not the operators measuring dimensions correctly and can end up with
same thing as the cooling water a lot of charts to control the process. In cases such as this, it is
temperature, it is the temperature of the normal to assign one or two dimensions as the ‘control
actual mould cavity after it has stabilised. dimensions’, i.e. they are the dimensions that effectively control
The mould temperature determines: or measure the other dimensions of the product. The other
• How fast the product will cool and how dimensions are not measured as they are assumed to be
much stress is locked into the product. dependent on the control dimension.
This approach also relies on manual intervention and
• The surface finish of the moulded adjustment if the chart indicates a need for process changes
product. A higher mould temperature and often the necessary process changes to be made are not
will increase gloss levels. clearly defined.
• How much crystallisation take place in However, for plastics processing the main difficulty with
the mould. measuring product dimensions is that they do not stabilise for
some time after moulding. For most polymers this can be some
• Tip - Getting a stable and controlled
hours due to thermal shrinkage, i.e. until the moulding has fully
mould temperature is essential.
cooled, but for polyolefins this can be some days due to
• Tip - The gate should always be the first crystallisation shrinkage, i.e. until crystallisation is complete.
place to freeze. This means that measurements must be made using ‘proxy’
dimensions or ‘hot’ gauges to allow for thermal and
Melt temperature crystallisation shrinkage and it is not always easy to predict how
Good control of and the correct melt the shrinkage will change with moulding conditions and time.
temperature is vital for any plastics
process but this is not often well controlled The second option is to use a single surrogate measurement
for all of the process parameters. The most common surrogate
or measured. The main controller of the
measurement is the product weight. If the tooling has been
melt temperature is the shear heating
dimensionally validated and the process has been set correctly
that takes place in the screw/barrel area
before series production then the product weight will be a
and the barrel heaters are primarily to measure of continued process control, i.e. if the cavities are not
adjust and maintain the temperature. A fully filled and packed then this will show as a weight variation.
complication for injection moulding is that This is again a manual approach that requires operator
the screw changes purpose during the measurement but weighing the product needs less time and
cycle, initially it is used to move and operator skill than measuring a dimension. Weighing the
condition the melt and then it is used as a product also relies on manual intervention and adjustment if the
ram to fill the mould. control chart indicates a need for process changes and as for
These dual functions make controlling the dimension measurement, the necessary process changes to be
melt temperature complex. Barrel made are often not clearly defined .
temperature settings, which measure the It is reasonably simple to set up and use a variables control
barrel temperature and not the melt chart based on product weight (see Section 5.18) and then to
temperature, are not the same as the melt use the product weight as a surrogate measure for process
temperature because they do not allow for control. I have used this with good results in the past but it must
shear heating. Other factors that affect be recognised that product weight is not as sensitive to the
inevitable process and material variations as the cavity
the melt temperature are injection rate,
pressure and melt temperature curve.
i.e. increased shear heating, back
pressure, cycle time and residence time. Note: If using product weight as a surrogate measurement then
• Tip - Getting a stable and controlled the control limits may need to be reduced to be sure of reliably
melt temperature is essential. identifying process changes, i.e. try using X ± 2σ instead of
X ± 3σ to set the control limits for the variables chart.

10.13 Injection moulding - managing attributes
Visual properties Follow a process Attributes are

generally the visual
Good process setting and cavity pressure If the rules for attribute charts indicate a aspects of the
control allows moulding within the process non-conforming attribute then the process moulding but can
window and this should also allow is ‘out-of-control’ and action needs to be include dimensions
consistent processing within the ‘cosmetic taken. The process is as follows: if these are
window’, i.e. the processing conditions monitored by ’go/no
necessary for good part cosmetics. What is the concern? -go’ gauges. A fail
There are many visual non-conformities on a ‘go/no-go
However, variation is always present in gauge’ is then
any process, and processes (material, that can arise in injection moulding and treated as an
mould or machine) can change during a the terminology varies not only from attribute defect.
run. Attributes (visual features) are not company to company but also from
only extremely important to customers but country to country. A list of some of the
are often an early indicator of process terms for the most common defects is
shown below. If the mould is the
changes or excessive variability in the
cause of attribute
process. The first step in solving attributes issues failure then fix the
• Tip - Define a ‘cosmetic window’ as well is to define what the actual concern is and mould.
as defining a ‘process window’. They may to get everybody to agree on what you are Do not try to adjust
not be the same thing! dealing with. Without agreement on what the process. You
Attributes in injection moulding are also the non-conformity actually is it will be might get the mould
impossible to resolve the issue. working but this will
multivariate, i.e. they depend on more
only push the ‘good’
than one variable, but in the case of • Tip - The range of terms used to describe
cavities close to the
attributes we are concerned with what we visual defects is very wide. Visual edge of the
see rather than what we measure and this samples or photographs of each type of processing window
does make it simpler to measure attribute concern will help people to agree on what and reduce the
data (although resolving any non- the issue is. Start a collection for the robustness of the
future now - I have my own samples. complete process.
conformities can be just as difficult).
Have we got a concern? Surface and visual (attribute) concerns

The easiest way to monitor attributes in
There are many different types of concerns that can be tracked
injection moulding is to use SPC charts for
via attributes. The terminology varies throughout the industry and
attributes (see Sections 5.19 and 5.20).
one man’s ‘silver streak’ is another man’s ’splay’, some of the
Attributes control charts are under-used
more common terms are:
in plastics processing but provide an
excellent early warning method to Black specks or streaks. Gate blush.
determine if there is a real issue with the
visual properties and if action needs to be Blisters. Gas marks.
taken. Burn marks. Gloss marks.
Attributes control charts are easy to set up
Brittleness Jetting.
(see Section 5.19 ) and analysis is easier
than for variables control charts as there Bubbles. Knit lines.
is only one chart to look at. The control Colour variations. Matt marks.
limits for attribute charts will give
warnings of any out-of-control conditions Contamination. Short shots.
(the rules are the same as for variables Cracking. Shrinkage.
control charts) and the need to take
action. Crazing. Silver streak.
• Tip - Accurate assessment of attributes Delamination. Sink marks.
needs consistent application of standards
Discolouration. Splay/splash.
and judgements (see Section 5.20).
• Tip - Use a constant sample size to
Ejector pin marks. Voids.
remove any issues with changing sample Flash. Warping/bowing.
size. This makes sampling easier for
Flow lines. Weld lines.
operators (and everybody else) to
understand. Flow marks.

When is the concern? •1. Routsis, A. 2015. ‘Injection Molding Analyse everything
Reference Guide’. Pg 46-47. Available free and use the Quality
When an attribute concern appears is vital
from www.traininteractive.com. Tools (see Chapter
to finding a solution and the options are 7) to isolate the root
shown on the upper right. Concerns can •2. 4Plas. 2014. ‘Injection Moulding cause and then
be: Troubleshooting Guide’, www.4plas.com/ solve it
download.php?id=111. permanently.
• Mould and machine trial related - this
type of concern must be resolved before •3. Kenplas. 2016. ‘Trouble shooting for the
the process is accepted for production. injection molding process’, www.kenplas.com/
service/imtroubleshooting.aspx.
• Start-up related - this type of concern is
at start-up with a mould, machine and
material that previously produced good
product. In this case the focus is on When does the
determining what has changed between concern appear?
the runs. Start-up-related concerns must
be corrected by the setter before the
machine is handed over to production.
Mould and Production start- Running process
• Tip - The main reason for start-up machine trials up (previously OK
concerns is generally changes in machine (previously OK with proven
settings made on the previous run. with proven material, mould
material, mould and machine)
Restore the machine to the ‘good’
and machine)
settings for production.
• Tip - If the mould is being used on a
different machine then settings will Adjust process Determine what Determine what
obviously need to be modified. setting for robust has changed has changed
process between runs: during the run:
• Tip - Check that the mould has not been (see Sections Material. Material.
modified since the last good run. 10.10 to 10.12). Machine. Machine.
• Tip - Check if the material is from a Mould. Mould.
Reset to Reset to
batch that previously produced good previous good previous good
products. If a new batch/delivery has settings. settings.
been used then try a different batch to
eliminate material issues.
When is the concern?
• Tip - Retain a sample of ‘good’ material
that can be used to quickly check for QC is mainly concerned with when a process fails to start-up
material issues if start-up is a problem. correctly or when a running process starts to produce poor-
quality parts. The first step is to return to the previous good
• Quality control-related - this type of
settings to see if this resolves the concern.
concern occurs on a running process
when the chart shows that the process is
‘out-of-control’. In this case the focus is
on determining what has changed during
Where is the concern?
the run and any changes in material,
machine or mould should be checked.
• Tip - For QC-related concerns look at
the material and machine first. Every shot Random
Every shot Random
on single on single
Where is the concern? on every
(same)
on random
(same)
cavity cavity
The location and frequency of the concern, cavity cavity
especially in multi-cavity moulds, can give
excellent information on the source of an
Look first Look first Look first Look first
attribute concern and the options are at: at: at: at:
shown on the lower right Material. Mould. Material. Mould.
• Tip - The important thing is to use all of Machine. Machine.
the available information.
How do we solve the concern? Where is the concern?
The Internet has many ‘troubleshooting Multi-cavity moulds give good information on the possible cause
guides’ for visual defects but my favourites of the attribute concerns. The important thing is to use all of the
are from Andy Routsis1, 4Plas2, and available information to determine the cause of the concern and
Kenplas3. resolve it.

10.14 Extrusion - managing variables
The die is the thing Key parameters Most extruded

products will be cut
The basic shape or outline of an extrusion Extrusion is, in some ways, simpler than to length whilst still
is set at the exit from the die lips whether injection moulding because there are warm and are likely
it is a profile, pipe, sheet or film. The fewer process variables. The motor driving to contract after
extrudate wall thickness or film thickness the screw is set at a nominally constant cutting. There will
is set by a combination of the die lip rotational speed to deliver a constant often be a cutting
opening and the haul-off speed. Depending volume of well-conditioned plastic to the tolerance applied to
ensure that the
on the product, these can be adjusted die. The consistency of the melt delivered customer gets the
whilst running by changing the die lip to the die is a major factor in process correct length of
opening or by changing the haul-off speed control for extrusion and process product.
to draw down the extrudate to the measurements can be either direct, i.e.
Try monitoring the
required wall or film thickness. they measure the plastic properties, or shrinkage and using
• Tip - Haul-off speed is the main method indirect, i.e. they measure the machine SPC with variables
of adjusting wall thickness for profiles properties. to minimise the
but can be limited by the need for cooling excess length
Direct process measurements needed and
of the extrudate before handling. increase quality and
Direct process measurements relate to the
• Tip - For profiles, the shape can be condition of the melt and are: profits.
adjusted slightly during calibration but
• Melt pressure - despite the nominal
the main shape is still set at the die lips.
‘steady state’ of the extrusion process,
Film and sheet are the simplest extrusion the melt pressure will always fluctuate
processes in terms of dimensional We once worked
and can lead to surging and dimensional with a BOPP film
tolerances and product control focuses variations. Melt pressure measured with producer whose
mainly on the film or sheet thickness. Pipe pressure sensors at the die is a sensitive process showed a
and tube is slightly more complicated and measure of the overall process stability. Cp = 1.407 with
has two critical product tolerances (wall Zupper = 0.712 and
• Tip - The best method of providing a
thickness and ovality). Profile extrusion Zlower = 2.103, i.e.
can be much more complicated and has consistent melt pressure is to use a gear Cpk = 0.712.
many product tolerances to control. It is pump. This will remove any pressure
This was a process
possible to assign one or two dimensions variations and deliver a constant
that was capable in
as the ‘control dimensions’ (see Section material flow at a constant pressure. statistical terms but
10.12), i.e. they are the dimensions that • Tip - Melt pressure can be measured by the location was
effectively control or measure the other pressure transducers at the end of the ‘high’ and the
barrel, at barrel vents or in the die. product thickness
dimensions of the product. The other would often exceed
dimensions are not measured as they will • Melt temperature - melt temperature is the upper tolerance.
be dependent on the control dimensions a vital part of a well-conditioned melt. It
We suggested
(assuming a properly built and sized die can be measured by thermocouples at moving the process
and consistent calibration and draw various points in the melt flow but to reduce the
down). control is more difficult. This is because thickness and still
• Tip - Control dimensions can be used melt temperature is mainly affected by stay within the
with standard SPC variables charts (see shear (frictional) heating in the interior tolerance. This
of the melt and heater bands or cooling would reduce the
Section 5.18) for basic control. amount of material
will only control heating or cooling of the used for the same
• Tip - Always note the control dimensions
on control drawings (see Section 8.13). melt at the barrel wall (these can be very m 2 value.
different). Shear heating is related to the
Manual measurement of ‘control However the
square of the shear rate and even small
dimensions’ can provide a basic quality company did not
increases in the motor speed can lead to sell their film by the
control strategy but this needs a lot of large increases in the melt temperature m 2 (what their
operator time, relies on the dimensions that cannot be quickly or easily customer wanted to
being correctly measured and often needs controlled by adjusting the heating or buy) but by the kg
several charts to control the process. This cooling at the barrel wall as this can take (what the company
also needs manual intervention and wanted to sell).
some time to affect the overall melt
adjustment if a chart indicates a need for temperature. Nothing was
process changes. There is a clear need for changed: There is
• Tip - Temperature sensors on the barrel
process measurements and controls no accounting for a
similar to those for injection moulding. do not indicate the melt temperature. company’s strategy.

• Tip - Adjustments to heater band • Tip - An SPC variables control chart of Cooling control
controls will suffer from a time delay in thickness can provide an early indication
For profiles, pipes
terms of adjusting the melt temperature. of potential process problems such as
and tubes the ability
blocked melt filters, etc. to cool the product
Indirect process measurements
• Tip - Select the control dimensions defines the
Indirect process measurements mainly carefully to control the complete product. production rate, i.e.
relate to the screw speed. The extruder an extrudate cannot
motor is set for a constant speed to deliver • Tip - If the extruder is a twin-screw be removed from
a constant flow of melt to the die but this processing powder, e.g. PVC, then the haul-off until it is
material feed control is essential for good sufficiently cooled
will vary with any variations in the melt.
mixing and process control. to be dimensionally
Common measures are: stable. Cooling and
• Motor speed - the effect of shear means • Tip - If the product is an extruded foam heat transfer
that drive systems must maintain a then it is not generally possible to make therefore define the
constant screw speed to maintain a any significant wall thickness production rate, i.e.
adjustment via haul-off control. slow the line down
constant melt temperature. Accurate or make it longer.
control of the motor speed is therefore
essential for melt temperature control.
Pipe and tube extrusion For blown film and
Thickness, diameter and ovality are sheet the
• Tip - Good control of motor speed will production rate is
result in good melt temperature control. generally measured using ultrasonic, laser
also limited by heat
or other measurement systems. transfer and most
• Tip - Overheating of the melt will reduce Depending on the system, the gauges can systems use PLC-
the viscosity and, if the motor speed is either rotate to measure the complete based line
poorly controlled, this can lead to an product or be stationary using two-, three- controllers to
increase in motor speed, an increase in or four-axis thickness, diameter and manage not only
shear heating and a ‘runaway’ process. the extruder and all
ovality measurement.
This is an issue because of the difficulty downstream
of cooling the melt close to the screw via Adjustment of thickness and ovality can equipment but also
be carried out by adjusting the position of the material feed
fans on the exterior of the barrel. and blending
the mandrel in the die either
• Amps drawn by the main motor - this is automatically or manually. systems.
a measure of the power drawn by the
motor (and therefore indirectly of the • Tip - Unlike four-axis systems,
power input to the melt in the form of stationary two- or three-axis systems do
not provide full coverage for pipes. Close control of
shear heating). This makes it a good
thickness in film
measure of the melt condition. extrusion allows
Film
• Tip - A change in the amps drawn is a down-gauging
Film thickness can be measured using a whilst still meeting
good indicator of a change in the melt
variety of non-contact methods, e.g. lasers, the specification.
viscosity (for whatever reason).
X-rays, beta, gamma or infrared, to
• Tip - The melt pressure at the die will measure the overall film thickness and
also change with any change in screw also to measure the layer thickness in
speed and this will cause dimensional multi-layer films. It is also possible to use Shrinkage will affect
variations (see above). extruded products
mass measurement as an indicator of film just as it does any
Measurement and control of the direct and thickness. plastic product and
indirect process measures through closed- Adjustment of film thickness can be needs to be
loop controllers (see Section 10.8) will give carried out by adjusting either the haul-off considered in any
stable process conditions and good process product dimensional
speed or the die lips and this can be
and product control. measurement.
automatic or manual.
• Tip - Minor fluctuations in amps and • Tip - Film requires not only control of
motor speed are inevitable but using an variations in the machine direction but
SPC variables chart can provide an early also in the transverse direction. Changes in factory
indication of potential process problems conditions can have
such as material variations, etc. a large effect on
Sheet extruded products,
Sheet thickness can be measured either e.g. temperature
Profile extrusion changes due to
manually (see Section 10.15) or by using a
Product thickness is generally the main laser callipers. draughts from open
measurement used for process control. doors, because it is
Adjustment of the die bolts to control the an ‘open’ process.
Assuming a stable process, product
sheet thickness can be carried out by
thickness can be controlled by adjusting Extrusion needs
adjusting the die bolts to change their good environmental
the haul-off speed relative the die exit
length and this can be automatic or controls for
speed to control the dimensions of the
manual. consistent products.
finished product.

10.15 Extrusion - using variables data
A case study tolerance band width. This means that The data are out
defects will always be produced at the top there, but they are
An extrusion site was producing extruded rarely converted
or bottom of the tolerance band and it will
sheet for later thermoforming into food into information that
not be possible for the process to produce
trays. The site regranulators were stacked can be used to
parts within tolerance. For this process
with ‘reject’ product from the extruders on drive process
(Cp = 0.96), out-of-tolerance products will improvement,
the basis of thickness variations.
be produced even if the process is exactly improve quality and
Capability study centred and does not vary. reduce the cost of
quality.
A rapid process capability study (see Process location
Section 5.22) was carried out on 50 Cp describes the process spread but does
consecutive historical data points not describe the process location. This is
extracted from the QC records for sheet described by Cpk (see Section 5.22). For the
thickness. These data were taken over process studied, Cpk = 0.93.
both the day and the night shifts.
When Cpk is < 1.00, as in this case, then it
These data were used to produce the will not be possible for the process to
capability study histogram shown on the produce parts within tolerance. The
lower right. This does not look exactly like process will always be located such that
a normal distribution but this is explained out-of-tolerance products are produced.
when it is considered that the operators
This is a process which is not capable of
are making manual measurements with
reliably producing products within
micrometers. The operators are reluctant
specification, i.e. nonconformities will
to measure in ‘5s’ and prefer to measure in
always be produced.
‘10s’. This is easy to do with a manual
micrometer and the operators will tend to In cases such as this, where the process is
‘fudge’ the results. This is normal human not capable, then the process:
Simple techniques
behaviour and is not a criticism of the • Should be examined for improvement to and tools can be
operators. This behaviour is shown quite reduce the variation. used to examine
clearly in the histogram. AND/OR processes to gain
an insight into how
The specification for the product is 575 µm • The tolerances adjusted to more realistic good they are
± 6% (± 34.5 µ, ). This gives: levels, in consultation with the customer, compared to the
• USL = 609.5 µm (Upper Set Limit or to increase the process capability. specifications.
Upper Acceptable Tolerance).
Capability study
• LSL = 540.5 µm (Lower Set Limit or 80
Lower Acceptable Tolerance).
70
The nominal thickness, the USL and the
LSL are all marked on the capability 60
study histogram.
Frequency
50
The distribution was considered to be
approximately a normal distribution 40
(subject to the operator adjustment) for
30
the purpose of analysis.
The results of the capability study showed 20
that the average ( X ) of the results was
10
574.02 µm and the standard deviation ( σ̂ )
of the results was 11.96 µm. 0
535 540 545 550 555 560 565 570 575 580 585 590 595 600 605 610 615 620 625
Process spread Thickness (µm)
The process spread (relative to the
specified tolerances) is described by the An initial process capability study for sheet extrusion
process capability index, Cp (see Section
The process capability study shows that Cp = 0.96 and Cpk =
5.22). For the process studied, Cp = 0.96. 0.93. This process is not capable in terms of either spread or
When Cp is < 1.00 then the normal location. Out-of-tolerance products will always be produced
distribution width is greater than the whatever the process settings.

If this is not done then the process will Using the numbers This is not high-
continue to produce excessive out-of- tech, it is simple
The available QC records show that this is statistics based on
tolerance products and reject material will
a process that is fundamentally not the natural
continue to be sent for regranulation.
capable of producing to the set variation of
• Tip - In reality, the situation will specification. The management were not processes.
actually be worse than this because this aware of this and this information
is more of a machine capability study prompted a review and improvement of
than a process capability study and will the complete process (and a reduction in
not show the full variation of the process. the scrap product being produced).
• Tip - This type of data is available in
Process control
many companies but is rarely used for
The QC records were further analysed to process management.
produce average and range ( X and R )
charts for the running process. The 5
values measured at each inspection were Range Chart
35
taken as the five samples for SPC by
variables (see Section 5.8). 30
The control limits were calculated using
standard SPC formulae and are: 25
Range (µm)
Range chart: 20
UCLR = 24.9 µm.
15
R = 11.65 µm.
LCLR = 0 µm. 10
Average chart: 5
UCL X = 567.27 µm.
0
X = 574.02 µm. 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49
LCL X Sample number
= 580.78 µm.
The control limits and process data are A range chart for sheet extrusion
shown on the right for the range (upper
The range chart (the first chart to look at) shows striation due to
right) and the average (lower right).
the measuring instrument resolution and also several times
The range chart shows that the range where alarms were indicated for process control. These were not
chart was broadly in control for the whole seen by the operators and scrap was produced.
period with the exception of two short
periods between 9 and 42 (the alarm
Average Chart
signals for range charts are given in 610
Sections 5.10 to 5.13). The rapid changes
in the later section of the range chart (37 600
to 50) may indicate that more than one
person was making the measurements or 590
Average (µm)
‘tampering’ with the machine when it

should not have been adjusted. 580
• Tip - The resolution of the measuring 570

equipment for this case is suspect due to
the operator tendency to measure in 10s 560
and not in 5s.
The average chart shows that the process 550
was out of control almost all of the time

540
(the alarm signals for average charts are
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49
given in Sections 5.14 to 5.16). This is Sample number
thought to be due to the operators
‘hunting’ for a stable process to meet a An average chart for sheet extrusion
specification that it simply could not meet
(see above). This behaviour explains the The average chart (the second chart to look at) shows poor
control of the process due to either a moving process, i.e. the
excessive rejects and the excessive use of
average appears to be increasing, or to poor process control.
the regranulators in the extrusion process.
There is also evidence of ‘hunting’ or tampering with the process.

10.16 Extrusion - managing attributes
A range of processes Quality concerns in extrusion Manage the

attributes (surface
Extrusion is used for a broad range of Whichever extrusion process is being concerns) found in
process but the basic principle of screw managed, it is essential to follow a process any of the extrusion
extrusion remains: a rotating screw (or to identify and remove the concern (see processes using
screws) feeds, pressurises, conditions and Section 10.13). The steps are: control charts for
moves the polymer melt through a die attributes (see
(profile, sheet, film or wire coating) to Have we got a concern? Sections 5.19 and
Attributes control charts are under-used 5.20).
form the final product. This seemingly
simple process is the basis for most in plastics processing but provide an
plastics forming methods, e.g. injection excellent early warning method to see if
moulding uses a rotating screw to there is a real issue with the visual
condition and move the melt into the properties and if action needs to be taken.
cavity, but the simplicity should not be Attributes control charts are easy to set up
underestimated, this is a complex process. (see Section 5.19 ) and the control limits
We will not cover the basics of extrusion for attribute charts will give warning of
process as these are covered in many any out-of-control conditions and the need
standard texts. By far the best general to take action.
text for all the necessary theory and • Tip - Do not over-react. Unless the
practice is Rauwendaal1 and the best for control chart shows a need to adjust the
troubleshooting the process is Noriega and process then leave it alone.
Rauwendaal2. They are essential.
• Tip - Accurate assessment of attributes
Screw selection needs consistent application of standards
and judgements (see Section 5.20).
Getting the correct screw for the material
being processed is a basic requirement. • Tip - Use a constant sample size to
‘General-purpose’ screws are sometimes remove any issues with changing sample
used but these are more accurately size. This makes sampling easier for
defined as ’no-purpose’ screws because operators (and everybody else) to
they requires so many compromises in understand.
Attributes charting
terms of screw design that they become What is the concern? and the control
unsuitable for most applications. limits tell you when
The range of visual non-conformities to take action and
• Tip - Get the right screw for the found in extrusion is as wide as those when to leave the
material and application. found in injection moulding and the process alone
Die care
The die is the heart of extrusion and it is
the die that determines the final shape of
the product as well as many of the visual
concerns that occur in extrusion. Despite
the seeming robustness of extrusion dies,
they are delicate, susceptible to damage
and must be handled with care (as with
any tooling). Damage to any die will show
in the final product and not only create
dimensional concerns but also visual or
attribute concerns.
• Tip - Always use soft tools when
cleaning dies and never use steel tools to
dig plastic out of a die.
• Tip - Disassembled dies are very The key screw dimensions in extrusion
vulnerable. Take extra care and always Good screw selection is important in getting good quality
use wooden surfaces. extrusion. The screw should always be specific for the material.
• Tip - If you have ‘rusty fingers’ that rust As the saying has it: ‘A general purpose extrusion screw is
dies then use gloves when handling dies. actually a no-purpose extrusion screw’.

terminology also varies. Without a clear Where is the concern? Never
definition and agreement on what the The location of the concern in extrusion underestimate the
concern is then it will be impossible to power of the Mark 1
does not give as much information as in ‘Human Eyeball’ in
resolve the concern. injection moulding but it is always good assessing attributes
• Tip - The range of terms used to describe practice to look for regularity in location to information. Be
visual defects is very wide. Visual find patterns and possible solutions. warned that it can
samples or photographs of each type of also be misleading,
concern will help people to agree on what How do we solve the concern? particularly in the
Cause and effect charts (see Section 7.4) area of colour
the issue is. Start a collection today.
assessment.
are invaluable in resolving extrusion
When is the concern? concerns and can be used in conjunction
When an attribute concern appears is vital with troubleshooting guides to provide a
in troubleshooting. Concerns can be: guide to the possible root cause of the
• Die and machine trial related - this type concern.
of concern should be resolved before the • Tip - For an excellent trouble-shooting
process is accepted for production. guide to film extrusion see the Qenos
• Start-up related - this type of concern is publication3.
at start-up with a die, machine and Extrusion attribute concerns will
material that previously produced good generally relate to one of the following:
product. In this case the focus is on • Material quality and preparation - these
determining what has changed between can be due to either poor incoming
the runs. Start-up related concerns must material , e.g. some gels in film
be corrected by the setter before the Improve the power
extrusion, or to excessive gas in the of the Mark 1
machine is handed over to production. material which is not vented sufficiently. ‘Human Eyeball’ by
• Tip - Start-up concerns are generally investing in a
• Process instabilities at the extruder -
due to changes in machine settings made simple portable
these can be high frequency, e.g. shark microscope. I have
on the previous run. Restore the machine skin, melt fracture or draw resonance, or several, including
to the ‘good’ settings for production. low frequency, e.g. surging, pressure one that plugs into
• Tip - If the die is being used on a fluctuations. the USB port of my
different machine then settings will laptop and displays
• Tip - Looking for a relationship between the view on screen
obviously need to be modified. the extruder screw frequency and the and can also
• Tip - Check that the die has not been concern frequency can sometimes reveal capture this as a
modified since the last good run. the cause of the concern. file.
• Tip - Check if the material is from a • Die concerns - these can be due to poor Invaluable tool for
batch that previously produced good die maintenance or a build up of detailed cosmetic
products. If a new batch/delivery has material in the die, e.g. continuous die and general product
been used then try a different batch to assessment.
marks along the length of a profile or
eliminate material issues. lines across the profile.
• Tip - Retain a sample of ‘good’ material • Process instabilities downstream from
that can be used to quickly check for the extruder - these are generally
material issues if start-up is a problem. product specific, e.g. wrinkles in the
• Quality control related - this type of machine or transverse direction for film
concern occurs on a running process production that are caused by the
when the chart shows that the process is downstream handling of the product.
‘out-of-control’. In this case the focus is If attributes charting indicates a need for
on determining what has changed during process adjustment then it is essential to
the run and any changes in material, follow a process to define the root cause
machine or mould should be checked. and not to jump to conclusions.
• Tip - For QC related concerns look at the You will never be
material and machine first. able to carry out
•1. Rauwendaal, C. 2014. ‘Polymer Extrusion’.
Although extrusion is seen as a continuous Hanser. effective 100%
process, it is actually a cyclic process inspection of your
•2. Noriega, M and Rauwendaal, C. 2010. production.
where many concerns can be identified by ‘Troubleshooting the Extrusion Process’.
relating their frequency to the crew Hanser. The end user
rotation frequency. Rauwendaal1 gives an (potentially your
•3. Qenos Pty. Ltd. 2015. ‘Film extrusion and customer) will
excellent chart relating the frequency of
conversion’. Free from www.qenos.com/ always carry out
concerns to the potential cause. internet/home.nsf/web/ 100% inspection of
PolyethyleneTechGuides. the final product.

10.17 Blow moulding
Even more processes • Tip - Venting is important. Manage dimensions

in any of the blow
Blow moulding is actually not a single Cooling the parison moulding processes
process. Although the principles are much The formed parison is finally cooled by the using control charts
the same, i.e. a parison or pre-form is chilled mould to form the finished product. for variables (see
heated and blown with compressed air to Section 5.9).
form the final product, there are actually 3 Concerns Manage the
distinct processes and these will be EBM can suffer from a range of concerns attributes (surface
discussed separately. and the main concerns are: concerns) found in
any of the blow
• Gels and black specks from extrusion -
Extrusion blow moulding moulding processes
treated as for standard extrusion. using control charts
(EBM) for attributes (see
• Extrusion die lines - treated as for
EBM is widely used for the production of Sections 5.19 and
standard extrusion.
containers in PE-HD or PE-LD. As with 5.20).
• Die swell - treated as for standard
any processing method, quality starts at
the design stage and an important but extrusion and corrected by parison
often neglected issue is the location of control.
vents in the mould. Vent areas are • Axial thickness variations in the parison
generally the last areas of the mould to - unless the parison is accurately centred
take the shape of the mould and if venting then the thickness will vary around the
is inadequate then this can prevent full extrusion axis and create weak areas. The crystallinity,
formation of surface details. • Parison pleating (parison folds in on density of the
material and the
itself as the length increases) - this will
Parison formation melt strength define
be more pronounced if there are weak which material can
Parison control is used to control and vary areas around the extrusion axis. be used for which
the wall thickness of the parison and blow moulding
• Parison droop (elongation of the parison
hence the wall thickness of the final process, e.g. PE-
due to own weight) - this can be reduced
product. Parison control varies the flow HD and PE-LD are
by reducing the parison length or by used for EBM
from the die head through the length of
using a polymer with a higher zero shear because of their
the parison (see diagram on the right) to
viscosity. high melt strength
provide a consistent parison and
whereas PP is not
consistent production. The important • Mottled surface due to water or uneven
often used because
points in achieving a good parison are: cooling. of the lower melt
• Consistent pressure at the gap head - • Pinch scarring from poor mould closing. strength.
this is needed to allow parison control to
be effective.
• Adequate land length - as with any
extrusion process, die swell at the die
exit can be minimised by providing an
adequate land length.
• Consistent polymer viscosity - parison
control works by varying the die annulus
and this is only effective if the polymer
viscosity is consistent.
Parison control also allows parison droop
to be controlled.
Blowing the parison
After the parison is formed it is trapped
and compressed air is blown into the
parison to form it into the shape of the The principle of parison control in EBM
mould. Adequate compressed air pressure Good parison control is the key to reducing product weight by
and flow is needed to form the parison to changing material distribution in the part. Not all areas are blown
the mould shape before it cools too much to the same extent and the wall thickness can vary widely across
to take up the surface details of the mould. the part. Parison control puts the material where it is needed.

• Trimming marks from poor trimming. process produces a highly biaxial oriented Injection stretch
• Tops and tails folding - where excessive material which greatly increases the blow moulding of
mechanical properties. PET bottles is one
tops or tails fold back onto the main
of the leading users
product and stick to it. Concerns of vision systems
for control of
Testing • PET moisture content before moulding -
dimensions. These
As most EBM products are designed to the presence of moisture will reduce allow the inspection
contain liquids, it is common to carry out a product strength and give a poor surface of up to 100,000
pressure test on every product. finish to the pre-form (and product). bottles/hour for a
range of
• Tip - For an excellent trouble-shooting • Heating distribution during blowing -
dimensions and all
guide to extrusion blow moulding see the uneven heat distribution will result in automatically.
Qenos publication1. preferential blowing and uneven wall
thickness. Vision systems and
automatic
Injection blow moulding (IBM) • Pearlescence (stress-whitening) on the inspection are a
IBM is an integrated process (see diagram inside of the bottle is due to over- new frontier in
on the right) which does not suffer from stretching in the blowing stage. inspection.
many of the issues of EBM but is limited • Haze or whitening of the outside of the
to smaller products. In IBM the pre-form bottle is due over-heating and
is injection moulded and, whilst still hot, crystallisation of the PET at any stage.
transferred to the blow stage before being • Low load strength is due to low wall
transferred again for cooling and ejection. thickness and general bottle design.
IBM gives better thickness control than • Impact failure is due to inadequate
EBM, has no tops and tails, i.e. greater orientation strengthening.
process efficiency, and has a reduced
tendency for surface defects due to the Testing
closed nature of the process. It is also Almost all ISBM products are designed to
capable of using PP as there is no concern contain pressurised liquids and it is
with the tensile viscosity of the melt. common to carry out a pressure test on Infrared cameras
every product. can be used to see
Testing the temperature
As many IBM products are designed to distribution in the
contain liquids, it is common to carry out a •1. Qenos Pty. Ltd. 2015. ‘Polyethylene Blow parison or pre-form
Moulding’. Free from www.qenos.com/ to check for uniform
pressure test on every product except internet/home.nsf/web/ temperature
where it is carrying solids, e.g. pills. PolyethyleneTechGuides. distribution.
• Tip - For an excellent trouble-shooting
guide to injection blow moulding see the
Qenos publication1.
Injection stretch blow

moulding (ISBM)
ISBM is largely used to produce beverage
bottles and one great advantage is that
the process creates biaxial orientation
which greatly increases the mechanical
Pre-form blowing Part removal
properties of the product.
Moulding the pre-form
Pre-forms are moulded using standard
injection moulding techniques (see Section
10.10) and stored before blowing.
• Tip - Good control of the neck area in Pre-form injection
the pre-form is needed as this is the final
form for the blown bottle.
Blowing the pre-form The IBM process
After storage, the pre-forms are clamped IBM is often carried out on a 3-station rotary turntable. The
at the neck area and the body area is turntable can have multiple stations (only one is shown for
reheated with IR heaters for blowing with clarity). After pre-form injection, the turntable is rotated for the
high-pressure air (35 bar). The blowing blow operation and then rotated again for part removal.

10.18 Thermoforming
A secondary process
Original
Thermoforming can be considered to be a mark
secondary process because it uses the
output of the extrusion process to produce Transverse
products. However, it is such an important direction
process that it will be considered in it’s
own right.
Machine
The most common type of thermoforming direction
is roll fed, plug-assisted thermoforming
and this is the process that is considered Results after heating
but many of the points here are common
Orientation in Orientation in
to other thermoforming processes. machine direction transverse direction
Sheet orientation and quality

Variations in the orientation of the
extruded sheet may show as poor or
uneven sheet performance during
thermoforming and a simple check (see
diagram on the upper right) can provide
an early detection and recording method Orientation in both
for uneven sheet orientation. machine and
transverse
Sheet heating (reheat) directions
Sheet heating is a critical phase of

thermoforming and affects all the
following phases. For most methods of The ‘quick and easy’ test for sheet orientation
sheet heating, and particularly for thick
sheet, the temperature of the sheet will be Sheet orientation can affect thermoforming quality. The ’quick
uneven through the thickness due to the and easy’ test is to mark the sheet (see ‘Original mark’ above),
poor thermal conductivity of the sheet. heat the sheet with unconstrained edges and check the resulting
Thermoforming has a ‘processing shape of the mark. The test will reveal the sheet orientation.
window’ (see diagram on the lower right)
for good production and the centre-line
temperature must be inside the processing
window before the surface temperature
goes outside the process window. If the
surface sheet temperature is higher than
Temperature oC
the processing window then degradation

will occur and the product will show
surface blisters and bubbles. If the sheet
temperature is lower than the processing
window then the sheet will show
incomplete forming or tearing.
• Tip - Understanding and controlling the
processing window is essential for good
thermoforming.
• Tip - The processing window defines the
process speed as well as quality.
• Tip - Using infrared heating can
Sheet heating and temperature are critical parameters
minimize temperature variations
through the sheet by matching the Close control of sheet temperature is essential for good
infrared heater emissivity to the thermoforming but the poor thermal conductivity of plastics
absorption characteristics of the material means that the temperature of the sheet will vary through the
being processed. Infrared heaters also thickness.

have a better response time and can into the final shape of the product. Manage dimensions
reduce energy costs. • Vacuum assist - this can be applied to in thermoforming
using control charts
Directional effects can also give uneven assist in forming the final shape.
for variables (see
sheet heating across the sheet. If standard Monitoring the plug location and force, the Section 5.9).
radiation heating is used then the centre air pressure and the vacuum through the
areas will ‘see’ more radiation than the Manage the
cycle provides excellent process control attributes (surface
edges and this can complicate getting the information and allows quality monitoring concerns) found in
sheet temperature correct even more by setting acceptable limits for each of thermoforming
although in this case the temperature these parameters. using control charts
variation will be across the sheet rather for attributes (see
• Tip - Toolvision (www.toolvision.de) uses Sections 5.19 and
than through the sheet.
this type of curve to provide excellent 5.20).
• Tip - Sheet temperature is a key process information for process control, process
and quality variable. Get the sheet development, quality control (to detect
temperature inside the processing quality issues in real time) and
window through and across the sheet for maintenance.
consistent processing and quality
• Tip - As with the cavity pressure curve
products.
for injection moulding (see Section
• Tip - Temperature control is more 10.10), the plug force versus time curve
important for semi-crystalline materials through the cycle is a sensitive measure
such as PP than for amorphous of process consistency.
materials such as PS because the
processing window is smaller. Quality controls At the design stage
the draw ratio, i.e.
• Tip - As a simple process control it is In the absence of good process control, total surface area of
possible to use IR temperature sensors to simple attribute control charts for the the formed part/the
measure the sheet surface temperature most common concerns (see below) can footprint of the part
and alert the operator if this goes outside provide excellent process control before forming, is a
specified limits. information for the heating phase - link key value.
The sheet heating phase of thermoforming the identified attribute concerns to the Given the desired
is ripe for improved process control (see possible causes and then solve the final wall thickness
Section 10.8) and whilst some attempts underlying cause. then the draw ratio
will provisionally
have been made, these do not appear to
define the starting
have made it into production machines to thickness of sheet
improve quality and the majority of needed for the part.
machines are still poorly controlled.
This assumes a
• Tip - As with all plastics forming perfect material
processes, the thermoforming process distribution and can
can take a little time to ‘settle down’. be affected by local
features such as
Sheet forming sharp corners.
Sheet heating is not the only process

variable in plug-assisted thermoforming Surface and visual (attribute) concerns
and other variables are:
Thermoforming has a range of concerns that can be tracked via
• Plug speed - this controls the strain rate attributes. As with injection moulding, the terminology varies
that the material experiences during throughout the industry. Some of the more common terms are:
forming. Increasing the plug speed will
increase the force on the plug and an Blisters/bubbles. Chill marks/striations.
excessive plug speed can lead to sheet Incomplete forming Shiny streaks.
tearing due to the strain rate sensitivity
Poor details. Post-forming shrinkage.
of polymers. Decreasing the plug speed
should decrease the plug force but the Blushing. Distortion.
sheet will be cooling whilst it is clamped
Color intensity change. Poor material allocation.
and the plug force can actually increase.
• Plug shape - this controls how the Webbing. Thin corners.
material deforms and is stretched and Bridging Part sticks in mould.
consequently thickness variations in the
Wrinkling. Sheet sticks to plug.
product.
• Air pressure - this will provide the final Excessive sag. Sheet tearing.
driver to deform the plug assisted sheet Sag variations.

10.19 Rotational moulding
Leaving the ‘Dark Ages’. ‘check weighing’ at the machine to Manage raw
validate the shot weight. materials (mainly
As with most plastics processing methods, particle size) in
the control methods used for rotational • Tip - Control material and shot weight
rotational moulding
moulding have improved dramatically in though the process. using control charts
the past 30 years. The start of this was the for individual
Avoid contamination measurements (see
pioneering work of Crawford1 and the
team at Queens University in Belfast who The fine powders used for rotational Section 5.18).
first introduced science to the process and moulding and the ‘open’ process makes Manage dimensions
developed entirely new process control material contamination a higher in rotational
systems based on ‘what the polymer saw’ probability than with other processes and moulding using
good contamination controls are needed. control charts for
as opposed to ‘what the machine saw’. As
variables (see
we saw in injection moulding, this is what • Tip - Keep the production area clean.
Section 5.9).
really counts. Don’t forget to look up and down as well
as around for potential contamination Manage the
These developments in both the theory attributes (surface
and process control provide the basis for sources.
concerns) found in
huge improvements in process control and Protect the parting line rotational moulding
part quality. using control charts
Parting lines on rotational moulds are for attributes (see
The basics sensitive to damage from mould opening Section 5.19).
and from cleaning.
Despite the developments in process
• Tip - Protect the parting line area by
control, there are still some basics that Standard process
building in features for mould opening.
need to be correct to produce quality control tests for
products in rotational moulding. • Tip - Always use the correct tools for rotational moulding
mould opening and parting line scraping. are:
Get the material right
• Tip - Never, ever use compressed air to • Impact strength.
Raw materials quality management in open a stuck mould - it can kill!
rotational moulding is potentially more • Bubble content.
complex than in other methods. In Process control • Internal surface
addition to all the other controls that are appearance.
Internal air temperature is the most
required (see Chapter 9), rotational • External surface
important process control parameter for appearance.
moulding needs the correct size of powder.
rotational moulding and a typical trace of
This is because rotational moulding uses • Dart impact.
oven temperature and internal air
fundamentally different materials to other • Bubble content -
temperature is shown on the lower right.
plastics processing. All other processes use the bubble content
either granules or mixed powder (twin- This shows that recording or controlling is inversely related
screw extrusion) but rotational moulding the oven temperature tells us very little to cure, i.e. high
about what the bulk of the polymer is bubble content
generally uses a mechanically milled indicates low cure.
powder of between 74 µm and 2,000 µm. seeing, i.e. the internal air temperature.
As well as checking the material and
grade, controlling the polymer powder size Surface and visual (attribute) concerns
is a key input variable. Rotational moulding has a range of concerns that can be tracked
• Tip - A powder that is too coarse will not via attributes. As with injection moulding, the terminology varies
only increase cycle times but will throughout the industry. Some of the more common terms are:
potentially give ‘under-fused parts’. Warped parts.
Under-fused parts.
Get the right amount of material Over-cured parts. Powder bridging.
Although rotational moulding is capable of
‘Ringworm’ on outer surface. Inside surface discolouration.
controlling wall thickness within ± 10%
this is only true if the correct amount of Mould sticking. ‘Lightning’ colours.
material is both weighed out and put into Non-uniform wall thickness. Low impact resistance.
the mould.
Part and parting line bubbles. Low stiffness.
• Tip - Use auto-weighing systems if
possible. Blow holes. Long-term part failure.
• Tip - Control initial weighing and use Parting line flash. Stress cracking.

Initially the oven quickly reaches the set- Cooling Warping is a major
point but the air takes longer to heat up Cooling is also important in rotational issue and one of
until it reaches the point at which the the main causes of
moulding. During cooling, the product will warping is non-
material starts to stick to the mould. The shrink due to thermal shrinkage and, if it uniform cooling.
rate increase of the internal air is a semi-crystalline material, due to
temperature then slows as the material crystallisation shrinkage. The cooling rate
absorbs the input energy. affects the degree of crystallisation and
After the mould is fully coated with the shrinkage. If the material is cooled too
polymer, i.e. all the material is effectively rapidly during cooling then warping will
molten, then the internal air temperature increase.
will increase again as the sintering and Under-fused parts
densifying process begins. The oven Quality and troubleshooting will show many
heating is then turned off but the internal small bubbles in the
Quality is linked to general
air temperature continues to rise and wall and a rough
troubleshooting and both Crawford and slightly powdery
drive coalescence of the polymer powder. Throne1 and LyondellBasell4 have internal surface.
This sintering and densifying continues excellent charts for process
and is largely complete when a specified troubleshooting and quality improvement.
peak internal air temperature is reached.
The peak internal air temperature defines
•1. Crawford, R.J. and Throne, J.L. 2002.
the point at which the material is fully
‘Rotational Molding Technology’. William
coalesced and cooling can start. Andrew.
As the mould cools, the material cools and
•2. Beall, G. 1998. ‘Rotational Molding’.
eventually starts to solidify. During Hanser.
solidification the rate of change of the
internal air temperature decreases until •3. Spence, A. and Scott, R. 2003. ‘Rotational
Molding Process Control’, ANTEC. Available The degree of cure
all the material is solid and it will then can be related to
from http://www.centroinc.com/Content/Our-
increase again until cooling is complete. Processes/Case-Studies.aspx. impact strength and
bubble content and
Peak internal air temperature and •4. LyondellBasell. Undated ‘A Guide to is a sensitive
‘degree of cure’ Rotational Molding’. Free from process control
www.lyondellbasell.com. measure.
Initial work1 showed that it was essential
for the internal air temperature to reach a
specified level to give sufficient time for Oven temperature
sintering and densifying. If this Internal air temperature
temperature was reached then moulders
could reliably produce good mouldings.
Later work3 has shown that it is not Peak
simply the peak internal air temperature internal air
that counts but also the time at this temperature
temperature and this was called the
Temperature oC
‘degree of cure’. This is unsurprising as All material

All material solid
coalescence will be affected not only by the melted
temperature but by the time at the Degree
temperature, i.e. the amount of input of cure
Tm
energy. Degree of cure is defined in units
Tg
of (°C x min) where the °C is measured as
being above the melting point and the Material starts
to melt locally Critical area for
time is also that above the melting point. crystallisation
The concept of the ‘degree of cure’ is and shrinkage
shown on the right.
• Tip - Controlling the degree of cure gives Time
good control of the process as seen by the
polymer. Heating Cooling
The ‘degree of cure’ can be considered
analogous to the cavity pressure curve (see Temperature changes in rotational moulding
Section 10.10) for process control. A
The oven temperature is not the same as the temperature that
controlled and consistent ‘degree of cure’
the polymer sees. What the polymer sees is the important thing.
will give a controlled and consistent Understanding and controlling the internal air temperature rather
process. than oven temperature is vital to good quality parts.

10.20 Processing systems - where are you now?
Get the systems right • Planning for process design and Good plastics
development. processing needs
High quality design and development good systems to
work counts for nothing if the transfer to • Creation and completion of FMEAs for
transfer processes
production is not managed consistently the complete process and machinery. from design and
and well to deliver the required results. • Control planning for the process, development into
including the necessary reaction series production.
The process models, e.g. APQP, for this
transfer are widely available but poorly or planning. Get the systems
little used in many plastics processors. • Good machine setting at the transfer right to get the
These models provide both the tools and production right.
stage.
good practices to set up processing • Full product and process validation
correctly from the start and for clearly before series production.
specifying what should happen in the
event of a process not performing Completing the chart
correctly.
This chart is completed and assessed as
The keys to good processing systems are: for the previous charts.
Processing systems
Process design &

development
4
2
Product & Process &
process machinery
validation 1 FMEA
A systematic
Initial Control approach to
machine setting planning process and
machinery planning
and control will
always give
reduced quality
issues in series
production.
Use the scoring chart to assess where you are in processing systems
Failing to plan is the
The numbers from the scoring chart can be transferred to the radar chart for a quick visual same as planning to
assessment of where you are in terms of processing systems. fail.

Processing systems
Process & Product &
Process design & Control Initial
Level machinery process
development planning machine setting
FMEA validation
Full process design & Full process & Full & detailed control Initial machine setting Full product & process
development carried machinery FMEA plan available with all carried out in logical validation carried out
out using logical steps. carried out & available. process & customer manner taking account using logical steps.
Records show all requirements of process variability. Records show all
required inputs specified. Process setting always required inputs
4 considered & all Reaction plans in the based on multiple considered & all
outputs produced. event of failure samples. outputs produced.
specified.
Majority of process Partial process & Partial control plan Initial machine setting Majority of product &
design & development machinery FMEA available. Identifiable carried out in a logical process validation
carried out using carried out. gaps in process & manner but little carried out using
logical steps. Identifiable gaps in customer consideration of logical steps.
Records show most FMEA coverage. requirements. process variability. Records show most
3 required inputs Written records Partial reaction plans Process setting mostly required inputs
considered & most available. in the event of failure based on multiple considered & most
outputs produced. specified. samples. outputs produced.
Some process design Poor process & Poor control plan Initial machine setting Some product &
& development carried machinery FMEA produced. Large gaps based primarily on process validation
out using logical steps. carried out. in process & customer previous ‘best’ settings carried out using
Records show some Large gaps in FMEA requirements. with no consideration logical steps.
required inputs coverage. Poor specification of of process variability. Records show some
2 considered & some Written records reaction plans in the required inputs
outputs produced. available. event of failure. considered & some
outputs produced.
Informal & poorly Informal process & Informal & poorly Initial machine setting Informal & poorly
documented process machinery FMEA documented control carried out with little documented product &
design & development carried out. plan. regard to process process validation
procedure. Poor written records Little reference to variability. procedure.
Poor written records available. process & customer Poor written records
1 available. requirements. available.
No reaction plans
available.
No discernable No process & No control plan Initial machine setting No discernable product
process design & machinery FMEA prepared. is poorly controlled & & process validation
development carried out. managed. procedure.
procedure. No written records No written records
No written records available. available.
0 available.
Score

10.21 Processing - where are you now?
Get the systems right The key to controlling any process is to Using attributes
identify and understand the effects of the data for process
The variety of plastics processing methods and product control
main process variables affecting the
and products means that the controls is sometimes seen
quality of the output product. These
applicable and those used vary widely as being easier
variables are different for every plastics
with the process and also throughout the than using variables
processing method but controlling the data.
industry. Despite these process and
process will always be more effective than
product differences, the standard methods Variables data
simply trying to control the product.
of control charting for process control of gives much more
variables and attributes can be applied for information that can
Completing the chart be used to control
most of the processes (see Chapter 5). This
This chart is completed and assessed as the process and
is more difficult when processes are
for the previous charts. reduce the
effectively multi-variate (see Section possibility of
10.11) but any attempt at process control producing non-
by controlling the key process variables conforming product.
will always give better results than
Where possible, my
traditional simple end-of-line inspection. preference is to
convert attributes
data to variables
Processing data for better
control, e.g. a
colour attribute can
Injection be OK or not OK
but an actual colour
moulding measurement gives
4 much better
opportunity for
control.
3
2 “The disciplines
we've created don't
Rotational correspond to real-
Extrusion world problems.
moulding 1
One of my friends
says, ‘The world
has problems, and
0 we have
departments.’
The fact is, there is
no electrical
engineering. There
is no mechanical
engineering. There
are problems, and
we solve them with
whatever tools we
Blow have available.”
Thermoforming Jim Melsa
moulding
Think of this when

you next approach
a quality concern
and equip yourself
Use the scoring chart to assess where you are in processing with the best tools
you can possibly
The numbers from the scoring chart can be transferred to the radar chart for a quick visual find - from wherever
assessment of where you are in terms of processing. you can find them.

Processing
Injection Blow Rotational

Level Extrusion Thermoforming
moulding moulding moulding
Scientific moulding Key direct process & Key direct process & Key direct process & Key direct process &
used for all setting. indirect process indirect process indirect process indirect process
Cavity pressure curves variables monitored for variables monitored for variables monitored for variables monitored for
used to control process control. process control. process control. process control.
process variables Product dimensions Product dimensions Product dimensions Product dimensions
4 Attribute charts used automatically automatically automatically automatically
to control surface measured & monitored measured & monitored measured & monitored measured & monitored
features. for control. for control. for control. for control.
Attributes monitored Attributes monitored Attributes monitored Attributes monitored
for process control. for process control. for process control. for process control.
Scientific moulding Some direct process & Some direct process & Some direct process & Some direct process &
used for majority of indirect process indirect process indirect process indirect process
setting. variables monitored for variables monitored for variables monitored for variables monitored for
Cavity pressure process control. process control. process control. process control.
sometimes used to Product dimensions Product dimensions Product dimensions Product dimensions
3 control process. manually measured & manually measured & manually measured & manually measured &
Attribute charts monitored for control. monitored for control. monitored for control. monitored for control.
sometimes used to Attributes monitored Attributes monitored Attributes monitored Attributes monitored
control surface for process control. for process control. for process control. for process control.
features.
Scientific moulding Some direct process & Some direct process & Some direct process & Some direct process &
used for some setting. indirect process indirect process indirect process indirect process
Cavity pressure rarely variables measured variables measured variables measured variables measured
used to control but not for control. but not for control. but not for control. but not for control.
process. Product dimensions Product dimensions Product dimensions Product dimensions
2 Attribute charts rarely manually measured as manually measured as manually measured as manually measured as
used to control surface part of inspection. part of inspection. part of inspection. part of inspection.
features. Attributes assessed as Attributes assessed as Attributes assessed as Attributes assessed as
part of inspection. part of inspection. part of inspection. part of inspection.
Product weight is main No process variables No process variables No process variables No process variables
control variable monitored. monitored. monitored. monitored.
monitored. Product dimensions Product dimensions Product dimensions Product dimensions
No use of attribute manually measured for manually measured for manually measured for manually measured for
controls apart from inspection. inspection. inspection. inspection.
1 inspection. Attributes assessed for Attributes assessed for Attributes assessed for Attributes assessed for
inspection only. inspection only. inspection only. inspection only.
No use of variables or Control is by regular, Control is by regular, Control is by regular, Control is by regular,
attributes charts. or irregular, inspection or irregular, inspection or irregular, inspection or irregular, inspection
Control is by regular, of final product only. of final product only. of final product only. of final product only.
or irregular, inspection
of final product.
0
Score

Key tips
• Taking product design to the required moulding provides a method of

process design needs a process and understanding ‘what the plastics sees’.
APQP provides an excellent model for • Managing the quality of attributes in
this. injection moulding needs a defined
• APQP provides a template and ‘aide process to get the most information from
memoire’ for all the things that you need the process.
to do to make the transition from the • Extrusion can use direct or indirect
design to the product. process measurement of variables data
• Carrying out a process and machinery for process control.
FMEA can not only help make the • Managing the quality of attributes in
transition easier but also help with extrusion needs a defined process to get
proving that you carry out risk-based the most information from the process.
thinking for ISO 9001.
• Blow moulding is a range of processes
• Control planning helps everyone to but both variables and attributes can be
understand what is needed to control controlled using standard process control
quality, when and where the controls techniques.
should take place and, most importantly,
• The critical process control in
what to do when things don’t go to plan
thermoforming is sheet heating and
(which is always the case).
getting a the complete sheet into the
• Process variability is always present and process window is key to production
process setting must take this into quality.
account.
• The critical process control in rotational
• Process setting should always use moulding is the ‘degree of cure’ which
multiple samples to account for defines the temperature and time needed
variability. for coalescence of the polymer powder.
• After a process is fully designed and
implemented then it must be validated
to make sure that it delivers what it was
designed to deliver.
• APQP provides a template and ‘aide
memoire’ for all the things that you need
to do to validate the product and process.
• Pre-control (zone charts) offers a rapid
method of assessing process control at
start-up and during running.
• Pre-control is an alternative to either
inspection or full control charts.
• Engineering process control (EPC) uses
established process control methods to
dynamically change the input variables
to control the output variables.
• EPC and SPC both try to control quality
but approach the issue from different
directions.
• Scientific moulding provides a rational
approach to setting injection moulding
machines by understanding what the
process is actually doing.
• Injection moulding is a multi-variate
process and this must be considered
when making any process change.
• Cavity pressure control in injection

Kent 2016

Uploaded by

Copyright:

Available Formats

Kent 2016

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Kent 2016

Uploaded by

Copyright:

Available Formats

Chapter 10

Chapter 10 - Processing quality management 293

Taking product design to • Consider ‘push/pull’ production If you are inspecting

294 Chapter 10 - Processing quality management

Floor plan layout

Chapter 10 - Processing quality management 295

Severity of event Probability of event Current controls

296 Chapter 10 - Processing quality management

Current process controls Recommended actions Action results

Extruder and upstream see previous sheet

Regular Start-up check by Production control

Chapter 10 - Processing quality management

A documented description of Stages of control planning Every part must

298 Chapter 10 - Processing quality management

Calibrators 1 to Cease production and

If chart indicates that

Chapter 10 - Processing quality management 299

Getting the best from the start

The effect of process variability Offset 1

300 Chapter 10 - Processing quality management

Chapter 10 - Processing quality management 301

302 Chapter 10 - Processing quality management

critical dimensions that do not require a

Chapter 10 - Processing quality management 303

Start-up control The pre-control rules Pre-control was

304 Chapter 10 - Processing quality management

Chapter 10 - Processing quality management 305

What do we want to control? Pre-control (product and

Inspection (product control)

306 Chapter 10 - Processing quality management

Chapter 10 - Processing quality management 307

The basics of EPC

308 Chapter 10 - Processing quality management

• Proportional control (P) - where the Disturbance to system

• Derivative control (D) - where the

Chapter 10 - Processing quality management 309

Can we use both? Conventional SPC EPC is rapidly

310 Chapter 10 - Processing quality management

Chapter 10 - Processing quality management 311

curves are not aligned because of time

The stages of moulding

312 Chapter 10 - Processing quality management

consistent in production - this improves

Chapter 10 - Processing quality management 313

A multivariate process -3σ +3σ

The best work in multivariate analysis

UMass Lowell and they have published

set at X ± 3σ and these limits will contain

99.73% of the population, i.e. there is a

single set of control limits for the process.

If the variable is within these limits then

the process is regarded as being in control.

314 Chapter 10 - Processing quality management

Chapter 10 - Processing quality management 315

316 Chapter 10 - Processing quality management

Chapter 10 - Processing quality management 317

Visual properties Follow a process Attributes are

Have we got a concern? Surface and visual (attribute) concerns

318 Chapter 10 - Processing quality management

Chapter 10 - Processing quality management 319