SPC 英语程序

1.0 PURPOSE

1.1 To promote basic SPC awareness to all departments.
1.2 To establish guidelines and procedure in using SPC towards quality improvement.
1.3 To interpret and simplify the technicalities of SPC and standardize its use for uniform analysis and interpretation.

2.0 SCOPE

Covers all statistical process control program used in Company.


3.0 DEFINITION AND TERMINALOGIES

3.1 Statistic – Branch of mathematics which involves the collection, analysis, interpretation and presentation of masses of numerical data.
3.2 SPC – Statistical Process Control
3.3 USL – Upper Specification (Specs) Limit
3.4 LSL – Lower Specification (Specs) Limit
3.5 Specification Limits – Tolerance specified by customer documents or Process/Product Engineer
3.6 UCL – Upper Control Limit
3.7 LCL – Lower Control Limit
3.8 Control Limits – Tolerance obtained according to actual process
3.9 X – Individual readings or measurements.
___
3.10 X – Mean or average of individual readings
----
3.11 X – Average of all averages
3.12 s - Standard Deviation (
3.13 Summation
3.14 R – Range
3.15 Rm – Moving Range

4.0 APPENDICES
_
4.1 X-R Chart
4.2 X-Rm Chart
4.3 P-Chart
4.4 C-Chart
4.5 Out-of-control Record Sheet

5.0 PROCEDURE

5.1 Establish an SPC team to develop and oversee SPC implementation. An initial training may be conducted by SPC facilitator to confirm uniformity of SPC knowledge among members.

5.2 Discuss the objective and importance of SPC:
Objective – Install and maintain an statistical process control system across the reduction of variability to provide continuous quality, reliability and service improvement through the use of statistic techniques.

5.3 Provide over-all awareness to middle management, line supervisors and operators through the use of seminar, trainings or posters.

5.4 Identify the process/station or machine/equipment to be control as well as the parameter to be controlled based but not limited to what are indicated in the PMP.

5.5 Assign the appropriate control chart to be used in each parameter of a process or station.

Variable Control Charts – Use for quality characteristics that can be measured and expressed in numbers such as height, length, temperature, time, pressure, speed, etc.
Samples of Variable Control Charts
_
X-R – Use for small samples (usually 3 to 8).
_
X-s – Use for large samples

X-Rm – Use for single sampling (1 sample only).

Attribute Control Charts – Use for quality characteristics that can not be measured and can be classified as either conforming or non-conforming to the requirements or specifications such as yield, absenteeism, dimensions checked by Go-No Go gauges, etc.
Samples of Attribute Control Charts

Defective Data – An entire unit fails to meet acceptance criteria, regardless of the number of defects on the unit.

P-Chart – Use for variable and large samples

nP-Chart – Use for constant and large samples


Defect Data – Failure to meet one of the acceptance criteria. A defective unit might have multiple defects.

C-Chart – Use in operation with constant and small samples.

u-Chart – Use for variable sample size.

5.6 Fill-up all the applicable information in the header portion of the control chart for items specified.

5.7 Sample size and frequency may depends on customer’s requirement, fixture/tooling and machine output or capability, current rejection rate, manpower availability and criticality of station or process. Normally, for process with high rejection rate and critical process, samples and frequency are more compared to process with high yield, not critical or already stable.

5.8 Initiate data collection for at least 25 sub-groups (1 column is equal to 1 subgroup). Record data in the appropriate control chart and plot the results. Ensure that during data collections, process should run on normal condition with equipment, machines, tooling and fixtures are in good condition and calibrated.
_ _ _
For X-R and X-s, X is computed as the average reading for each subgroup or column.

5.9 After completing at least 25 sub-groups, start to compute the control limits of the certain process. Refer to below table in computing central line and control limits per control chart. Different control chart has different formula in computing central line and control limits.

ATTRIBUTE CONTROL CHARTS

Central Line

Upper Control Limit
( UCL )
Lower Control Limit
( LCL )

u-Chart


_ # of Defects in a subgroup
u = -----------------------------
# of insp. in a subgroup
_
u
UCL = u + 3 ---------
n
_
u
LCL = u - 3 ---------
n


C-Chart


_ Total # of Defects
C = -----------------------
Total # of Samples

_ _
UCL = C + 3 C


_ _
LCL = C – 3 C



P-Chart

_ Total # of defectives
P = --------------------------
Total # of samples
_ _
_ P(1-P)
UCL = P + 3 ------------
N
_ _
_ P(1-P)
LCL = P - 3 ------------
n


np-Chart

Total # of defectives
np = ----------------------------
# of subgroups
_ _ _
UCL = np - 3 np(1 – P)


_ _ _ UCL = np - 3 np(1 – P)












VARIABLE CONTROL CHARTS
Central Line Upper Control Limit
( UCL) Lower Control Limt
( LCL)
_
X-R
_ _
X = X / # of Subgroups

R = R / # of Subgroups __ _
UCLX = X + A2R
_
UCLR = D4 R __ _
LCLX = X - A2R
_
LCLR = D3 R
_
X-s
_ _
X = X / # of Subgroups

s = s / # of Subgroups __ _
UCLX = X + A3s
_
UCLs = B4 s __ _
LCLX = X – A3 s
_
LCLs = B3 s

X-Rm
_ _
X = X / # of Subgroups

Rm = s / # of Subgroups __ _
UCLX = X + E2Rm
_
UCLRm = D4 Rm
__ _
UCLX = X - E2Rm
_
UCLRm = D3 Rm


Samples
n
_
X-R _
X-s
X-Rm

A2
D3
D4
A3
B3
B4
c4*
E2
D3
D4
d2*

1
---
---
---
---
---
---
---
2.659
0
3.267
1.128

2

1.880

0
3.267
2.659
0
3.267
0.7979

2.659
0
3.267
1.128

3

1.023
0
2.574
1.954
0
2.568
0.8862

1.772
0
2.574
1.693

4

0.729
0
2.282
1.628
0
2.266
0.9213
1.457
0
2.282
2.059

5

0.577
0
2.114
1.427
0
2.089
0.9400
1.290
0
2.114
2.326

6

0.483
0
2.004
1.287
0.030
1.970
0.9515
1.184
0
2.004
2.534

7

0.419
0.076
1.924
1.182
0.118
1.882
0.9594
1.109
0.076
1.924
2.704

8

0.373
0.136
1.864

1.099
0.185
1.815
0.9650
1.054
0.136
1.864
2.847

9

0.337
0.184
1.816
1.032
0.239
1.761
0.9693
1.010
0.184
1.816
2.970

10

0.308
0.223
1.777
0.975
0.284
1.716
0.9727
0.975
0.223
1.777
3.078

Note: If the computed value is negative, use zero (0) instead of the computed negative value. Exception on the X value where there is a negative specification required.
If the computed value of control limits exceed either the USL or LSL, re-study and re-calculate the data and computed value. New data gathering could be necessary as well.