FTA (изд-во). Flexography: Principles And Practices. Vol.1-6

Подождите немного. Документ загружается.

tatistical process control (SPC)

represents a “tool box” from

which the printer can draw in

order to define the printing

process, measure and control its

key parameters, and improve

upon its ability to deliver a satisfying product

to the customer. Measuring, collecting and

using critical data is a cornerstone of a quali-

ty program.

100% INSPECTION AND SAMPLING

The use of 100% percent inspection is not

a statistical tool; however, in the flexo indus-

try it is sometimes necessary to inspect

every item in the process. This is true in

areas such as artwork and designs, comput-

er disk files, printing plates, cost estimates

and billing. Items in these areas are one of a

kind and irreparable damage can be caused

if they are not caught and corrected. In this

type of inspection, it is very important that

the inspector be knowledgeable and alert

and have the appropriate time to accomplish

the inspection process. Sometimes it is nec-

essary to have more than one person inspect

the process because of the criticality of the

object being inspected. It is helpful to have

an inspection form (checklist) so that the

inspector does not forget items. A record

can be kept of the item checked and noted

as approved, or flagged for correction. Table

18, is an example of a checklist.

STATISTICAL INSPECTION

AND SAMPLING

In contrast to 100% inspection, statistical

sampling means inspecting a limited number

of samples. Statistical sampling offers econ-

omy of scale while remaining a very effec-

tive quality tool. In a 30,000 run of flexible

packages it would be prohibitive in time and

cost to inspect each bag. Therefore a more

effective procedure is to develop a statisti-

cally valid sampling plan to validate the qual-

ity of the product being produced. A mini-

mum of 30 samples is needed to adequately

develop an SPC charting system. Table 21

shows the numbers for a statistically sound

sampling plan.

ATTRIBUTES AND VARIABLES

An attribute is defined as a characteristic

that is either present or absent. Some exam-

ples of attributes include whether the die is

cutting or not, whether the seal holds or not,

whether the typography is present or not. An

attribute can be classified as yes or no, 0 or

1, present or absent.

QUALITY CONTROL 97

Statistical

Process Control

A Illustrates the air

flow mpattern through

recuperative thermal

oxidizer

Table 21

RUN LENGTH SAMPLE SIZE

1–5,000 2 samples per 100

5,001 – 100,000 1 sample per 100

100,000 up 1 sample per 200

STATISTICAL SAMPLING PLAN

98 FLEXOGRAPHY: PRINCIPLES & PRACTICES

A variable is the result of a measurement

and has a tolerance or ± associated with it.

During a flexo production run variables will

never be constant but always have some

variation. Some common variables are ink

viscosity and pH, solid ink density, dot gain,

color value, plate, stickyback and substrate

thickness, and registration.

MILITARY STANDARD

(MIL-STD-105E)

Military Standard (MIL-STD-105E)

is a

method of attribute-acceptance-sampling

that has been developed by the United States

Department of Defense and is widely accept-

ed by industry as an effective procedure for

attribute sampling. This standard includes a

sampling plan, which is the acceptable qual-

ity level (AQL), run-length size and corre-

sponding sample size, and acceptance and

rejection numbers. A sample is shown

(Figure

To review how to use MIL-STD 105E use

Figure

. The figure has two charts,

“Sample Size Code Letters Chart” and

“Acceptable Quality Level Chart.” The first

lists code letters for inspection levels for a

given lot or batch size. The inspection levels

allow for more or less sampling depending

on the history or established quality level of

a given supplier. For example, if the flexo

run length is 100,000 and there is no history,

the normal or default level II (letter N)

would be used. With a quality supplier with a

history of quality success, column I, (letter

L) could be used. On the other hand, with a

poorer quality supplier, column III (letter P)

might be appropriate.

Next, the Acceptable Quality Level Chart

is used. Using the above example of a run

length of 100,000 and the letter N, the second

column of the chart shows the sample size

needed. In this case the number is 500 sam-

ples, which need to be taken in a random

manner. Finally, the number of samples

allowed to be out of specification to achieve

an Acceptable Quality Level (AQL) is given

in the right hand side of the chart. Most com-

panies in the U.S.A. choose an AQL of 1.5 or

2.5. Basically, the 1.5 and 2.5 mean there is a

98.5% and 97.5% confidence, respectively, in

the sample plan. This is the customer’s

choice and is dependent upon the chances

one is willing to take that the sample plan

may fail. Using an AQL of 1.5 for this exam-

ple, two numbers, Ac and Rc, are listed in

the column under 1.5. Their values are 14

and 15. This means that the product is with-

in the acceptable tolerance level chosen if 14

or less out of the total sample of 500 are out

of specification. If 15 or more are out of

specification, the product is out of the

acceptable tolerance level and may be

rejectable. One could go to a higher level of

sampling (such as from N to P) or, if feasible,

one could go to 100% inspection to get rid of

all out-of-specification product. Strictly

speaking, even when 100% inspection is

done, this does not guarantee 100% accept-

able product. Letter P with an AQL of 2.5

doesn’t have an entry in the chart. Instead,

the arrow means to use the numbers to

which it points, in this case, 21 and 22.

Military Standard Sampling Procedures and Tables for Insertion by Attributes

(MIL-STD-105E) and Military Standard Sampling Procedures and Tables for

Insertion by Variables (MIL-STD-114) can be obtained from Naval

Publications and Forms Center, 5801 Tabor Avenue, Philadelphia, PA 19120.

QUALITY CONTROL 99

MIL-STD-105E is a

method of attribute-

acceptance-sampling.

In the top chart, a letter is

assigned based on run

length. This letter is used

in the bottom chart to

determine an appropriate

sample size and (reading

across) the “accept” and

“reject” levels, based on

the number of errors

found.

Lot or Batch Size I II III

General Inspection Levels

Sample Size

Code Letters Chart

Acceptable

Quality Level Chart

2–8

9–15

16–25

26–50

51–90

91–150

151–280

281–500

501–1,200

1,201–3,200

3,201–10,000

10,001–35,000

35,001–150,000

150,001–500,000

500,001 and over

Sample Size

Code Letter

Sample Size

125

200

315

500

800

1,250

2,000

0.010 0.015 0.025 0.040 0.065 0.10

Acceptable Quality Levels (Normal Inspection)

0.15 0.25 0.40 0.65 1.0 1.5 2.5 4.0 6.5 10 15

■ Oval Begin or End

■ Rectangle Activity

■ Document Linked to Activity

■ Diamond Decision

■ Arrows Flow of Process

FLOW CHART SYMBOLS

100 FLEXOGRAPHY: PRINCIPLES & PRACTICES

tatistical process control (SPC)

involves not only measure-

ments and tracking those mea-

surements, but also charting

and other tools to quantify and

describe the process. The seven

tools of statistical process control are listed

in Table 22.

FLOW CHARTS,

OR PROCESS MAPPING

Flow charts, which are also known as

process maps, are used to define the key steps

in the flexographic reproduction process.

They help in determining the correct and nec-

essary ways to perform a given operation and

give direction to the development of standard

operating procedures (SOP’s). By following

the standard operating procedures estab-

lished by the flow chart, a printer can be

assured of “doing things right the first time.”

This results in less waste, a more consistent

product, higher productivity and reduction in

the cost of producing the product. Table 23

shows the symbols used in flow charts and

Figure

shows an example of a flow chart

for creating a color target to be used for cus-

tomer approval of a spot color.

CAUSE AND EFFECT ANALYSIS

Cause and Effect Analysis is used to identi-

fy the many causes of quality-related prob-

lems. For example, if the printer wanted to

know the causes of dirty print, cause & effect

analysis using what is called a Fishbone dia-

gram (Figure

) would be a useful tool

which could quickly and efficiently define a

list of probable causes. Most often, this tool

is used by a small group of people utilizing

the brainstorming methodology. This allows

Tools of Statistical

Process Control

Table 23

1.Flow Charts or Process Mapping

2.Cause & Effect Analysis

3.Checksheets and Checklists

4.Pareto Analysis

5.Run and Control Charts

6.Histograms

7.Scatter Diagrams

THE SEVEN TOOLS OF

STATISTICAL PROCESS CONTROL

Table 24

QUALITY CONTROL 101

Flow charts should be

used to define key steps

in a given process. This

chart shown might be

used in creating a color

target.

Find Out Why

and Correct

COLOR APPROVED

Make 6 Inkroom Proofs

Send 2 Proofs to Customer

Find Out Why

and Correct

Make Ink for Press

OK Color on Press

Visually and Numerically

Send For Approval

YES

APPROVED

Create Color Target

START

YES

File All Paperwork

Get Data on Proof

PROOF

FINISH

NEED

CUSTOMER

APPROVAL

Form A2

Form B7

Form B9

Adapted from Progressive Inks.

102 FLEXOGRAPHY: PRINCIPLES & PRACTICES

A fishbone diagram is

helpful in defining the

cause and effect of a

problem. The chart

to the right shows

possible causes of a

dirty print.

FISBHONE DIAGRAM

Causes of Dirty Print

ENVIRONMENT

MANPOWER

MATERIALS

EFFECT (Problem)

DIRTY PRINT

Shift Change

Operator

Communication

Didn’t Check

Standard

Operator/Helper

Training

Arrive at Workstation

on Time

Not Watching

METHODS

Ink Film Too Heavy

BC Air Temp

Too Much

Impression

Chill Roll

Temperature

Poor Setup

Poor Quality

Standards

Printing Wrong

Side of Web

Inconsistent Standards

Wrong Viscosity

Wrong pH

Dirty Ink Pan

Poor Ink

Film Treatment

Foamy Ink

Wrong Stickyback

Too Much Alcohol

Defective Plates

Old Plates

Wrong Extender

Slip in Film

Dirty Plates

MACHINES

Slow PumpRubber Roll

Durometer

No Cover Pans

Roller Speed

Station Design

Air on Plates

Bad Bearings

Dirty Drum

Bad Plate Cylinder

Dirty

Journals

Weather

Shift Time

Output of Process

Desired quality

is clean print

Adapted from Progressive Inks.

the printer to draw upon the collective exper-

tise of process “experts” (electronic pre-

press, press operators, supervisory person-

nel, helpers, strippers and others directly

working in the system).

CHECKSHEETS AND CHECKLISTS

Checksheets are tools which allow for the

easy collection and analysis of data (Table 25).

They are simple, systematic ways to collect

and organize data. Checksheets can be used

to determine where, when and why problems

such as hickies, wrinkles, marking, and other

printing defects occur. The data collected can

easily be turned into a histogram (Figure

Checklists are familiar to most people from

ordinary experience. An example is using a

shopping list when going to the store. The

benefits of doing so far outweigh the costs of

trying to remember. In the printing environ-

ment checklists can be used in the same way

to consistently check items against a list and

not leave anything to memory. Table 18

already showed such a list. Table 25 shows a

checklist for proof approval.

PARETO ANALYSIS

Pareto Analysis is a tool for identifying

cost-effective solutions for quality improve-

ment. The principle of Pareto Analysis is the

familiar 80/20 rule: the bulk of printing prob-

lems (80%) are due to only a small minority of

the related causes (20%). Most customer

complaints can be tied to a few systemic

problems such as late delivery or printing

defects (Figure

). Most printing defects

are caused by a few items such as a specific

stock, press or ingredient. The key is to col-

lect data on the relative frequency of each of

the causes and then find solutions to the

largest of these. Most importantly, do not

make assumptions about how important an

item may be – collect the data first.

QUALITY CONTROL 103

This histogram, based on

data collected in Table 24,

plots pH readings in an

easy-to-reference format.

Table 24

READING FREQUENCY TOTAL

8.0 ✓✓ 2

8.3 ✓✓✓✓ ✓ 6

8.6 ✓✓✓✓ ✓✓✓✓ ✓✓ 12

8.9 ✓✓✓✓ ✓✓✓✓ ✓✓✓✓ ✓✓✓✓ ✓ 21

9.2 ✓✓✓✓ ✓✓✓✓ ✓✓✓✓ ✓✓✓✓ ✓✓ 22

9.5 ✓✓✓✓ ✓✓✓✓ ✓ 11

9.8 ✓✓✓✓ ✓✓ 7

10.1 ✓✓ 2

CHECKSHEET FOR pH

Table 25

Check against customer original art, board or a

proof of electronic file.

■ Color Breaks.

■ Copy: Location and verbiage.

■ Bleed off panels (0.375" min.)

■ Special instructions on Mylar.

■ Printability of small copy © ®

■ Ensure verbiage is not less than 0.25" to

score

■ Check process work against customer

target.

■ Affix sign off label, sign and date.

Adapted from Checklist developed by Schiffenhaus

Packaging Corp.

PROOF OF APPROVAL

CHECKLIST

8.0 8.3 8.6 8.9 9.2 9.5 9.8 10.1

Frequency

104 FLEXOGRAPHY: PRINCIPLES & PRACTICES

RUN AND CONTROL CHARTS

Run and control charts are tools used by

operators for monitoring the printing process

on an ongoing basis and making adjustments

as necessary. Instead of waiting for things to

go wrong, the control chart serves as an early

warning device for the operator who can

then take the appropriate action long before

the occurrence of substandard production.

The cost of using control charts is the time

and training required so that operators have

the knowledge and resources necessary to

use them properly. The benefits are reduced

spoilage and much more consistent results.

Figure

showed a run chart. With the addi-

tion of upper and lower control limits, this

chart becomes a control chart as shown

(Figure

). If the measured values fall out-

side these control limits, some corrective

action needs to be taken, since out of specifi-

cation product can be produced. A control

chart is used in conjunction with a range

chart so that a whole picture can be seen of

the process. It is important to note that the

control limits must be well within the speci-

ifcation limits. Control charts should be used

to determine whether the process is in con-

trol before using a histogram to determine

process capability. It is advisable to use

spreadsheets or specific statistical computer

programs when working with control charts.

Refer to Appendix C for additional details.

HISTOGRAMS

Histograms are used for comparing the

flexo product to its specifications and to

assist in the determination of press capabili-

ty. Histograms do not show variation over

time, but the overall variation of the process

being statistically monitored. For a particular

variable, measurements are taken and the

frequency of the results are graphed as was

shown in Figure

. The histogram will show

if the natural variation in the variable is larg-

er or smaller than the desired variation.

Figure

is shown again in Figure

with

the natural variation shown as a bell-shaped

curve (normal distribution). An important sta-

A Pareto chart can be

used to plot the different

types of customer

complaints.

A control chart shows the

upper and lower limits for

your process. If the plotted

values fall outside of these

limits, corrective action

must be taken.

A histogram shows

whether the variation or

tolerance for a variable is

within desired limits.

100

Gloss Wrong

Color

Strength

Adhesion Late

Delivery

Other

Type of Problem

Number of Complaints

Upper Specification Limit

Upper Control Limit

Specification

Process Average

Lower Specification Limit

Lower Control Limit

123456789101112

Sample Number in Increments of Time

Density

1.58

1.56

1.54

1.52

1.50

1.48

1.46

1.44

1.42

1.60

99.7% (3 Sigma)

95% (2 Sigma)

68%

(1 Sigma)

-.08 -.06 -.04 -.02 -.00 .02 .04 .06 .08

Print to Die Deviation

Frequency

tistical measure associated with this curve is

the standard deviation called sigma (σ or Σ).

In Figure

the sigma value is 0.02. This

means that 68% of the values will have a devi-

ation of ±0.02, 95% a deviation of ±0.04 (2 σ),

and 99.7% a deviation of ±0.06 (3 σ). Control

limits should be 3 σ or larger in order for the

process to consistently and reliably produce

the target value.

In the example shown in Figure

, the

histogram shows that for a specification or

tolerance of 0.06 or greater, the process

would be capable of producing acceptable

product. For a tighter tolerance, the process

is not capable of reliably producing accept-

able results.

For more information on histograms, refer

to Appendix B.

SCATTER DIAGRAMS

Scatter diagrams are tools used for deter-

mining how important the cause and effect

relationship is between two variables. This

method could be used for testing out such

hypotheses as “Running the press faster

causes more spoilage,” “Customer turnover

is lower when jobs are delivered on time,” or

“Thicker ink films cause higher dot gain”

(Figure

). While these statements may be

intuitively appealing, it is hard to know how

important press speed, on-time delivery or

ink film thickness are without collecting

data. A scatter diagram can be used for show-

ing the correlation (Figures

and

QUALITY CONTROL 105

Scatter diagram of dot

gain vs. film thickness.

Scatter diagram showing

positive correlation.

Scatter diagram showing

negative correlation.

Scatter diagram showing

no correlation.

Film Thickness

Dot Gain

Effect

Cause

Positive Correlation

Effect

Cause

Negative Correlation

Effect

Cause

No Correlation

106 FLEXOGRAPHY: PRINCIPLES & PRACTICES

he pursuit of quality is an ongo-

ing process. Using SPC tools the

process should be monitored

and corrected as required. A

good way to do this is to include

a control target, as shown in

Figure

, on every job. If the job precludes

this target, at the very minimum a run target

should be included in the live area of the job

(Figure

The control target allows continual inspec-

tion and measurement of key quality para-

meters of the process. These parameters can

then be charted to make sure the process

stays in control. The flexo printer should

have in house standards which are used to

control the process. These standards may be

industry guidelines, such as FIRST, or stan-

dards specific to the printer. The standards

need to be understood and communicated to

the entire organization, including suppliers

and customers.

VISUAL INSPECTION

A number of quality characteristics can be

checked visually. These include slur, regis-

tration, trap, gray balance and color. Some

of these characteristics can also be mea-

sured and quantified, but a visual check is a

quick verification that the process is still

under control.

• Registration can easily be measured

through visual inspection. Accuracy for

this measurement can be greatly in-

creased with the use of a magnifying

glass. A 12x or greater power device is

very effective. Register marks can be

designed that visually assist in determin-

ing how far out of register the colors are

from each other.

• Slur targets will assist the press operator

in visually determining the accuracy of

the impression and anilox pressure set-

tings. These targets can also assist in

determining worn gears, out of round

Elements of Process

Control in Flexography

4# A control target is an

excellent means of

monitoring a process.

A simple control target

is shown here. This is

the FTA control target,

adapted from the 1997

FIRST Standards.

B D E

F G

H J K

A Ink Trap Patch

B Solid Process Patches

C Exposure Guide

D Solid Density Patches

E Slur Patch

F Reference Code

G First Logo

H Tonal Scale

I Dot Gain Values

J Highlight Grey Balance

K Shadow Grey Balance

Three

Color

Black

Only

Three

Color

Black

Only

Adapted from FIRST, 1998