7. 品質意義-品質特性八構面
美國哈佛商學院 David Garvin（1987）對品質特性的構面分類
No. Item English 說明
5 美感性 Ae s t h e t i c s
外觀的視覺效果所產生之主觀
的感知程度。
6 性能 Fe a t u r e s 能展現的特性。
7 認知品質 Pe r c e i v e d Qu a l i t y
因口碑信譽而所設定的品質程
度。
8 品質的符合性
Co n f o r m
a n c e t o
St a n d a r d s
與設計期許要求的符合程度。
品質特性的八構面
No. Item English 說明
1 功能 Pe r f or m
a nc e 欲執行使用的目的。
2 可靠度 Re l i a bi l i t y 失效的頻次。
3 耐久性 Dur a bi l i t y 可使用的年限。
4 維護性 Se r vi c e a bi l i t y 維修保養的難易度。
品質特性的八構面
28. www.anser.com.tw 版權所有 請勿翻印 28
- States the effect (what is wrong，not why it is wrong)
- Provides understanding of the issue to be addressed. The problem or issue
needs to be carefully defined and bounded. Careful definition is an essential
first step to problem containment and resolution.
- May include internal and external description - May be refined as process progresses
- Analyze existing data - Establish operational definition
- Who, what, when, where, why, how, how many - Be specific
- What is wrong with what - Establish extent of problem
Issue Description, 4 examples:
1. You have suffered a severe sun burn
2. Your car engine begins to run roughly
3. Your neighbor’s dog is on your property chasing your pets
4. Your baby has just got shocked from sticking its tongue in an electrical outlet
Issue Description
30. www.anser.com.tw 版權所有 請勿翻印 30
2.1 State all known data given by customer. 列出所有在客戶端知道的讯息。
Customer name (IBM, UK) 顧客名字 ( 如 : IBM, UK )
Model (MS-6188 1.0A) 模型 : ( MS - XXXX 1.0 一 )
Phenomenon (Failed on IAT test program) 现象 :
(如XXX测验程序方面失败或XXX组装不良)
Time – 04/10/2001發生時間 : XX年XX点XX分
Station – First burn-in, Post test, and so on. 發生站别- 修補站，测试站等。
Test configuration – Equipment brand, model, speed ……
2.2 State the product affected. 評估影響範圍。
- Customer bar code ( xxxxxxxxxxxxx ) 客户端的條形码 ( xxxxxxxxxxxxx )
- Control code (04) 客户端控管代码 ( XX )
2.3 State if there has been multiple occurrences of problem. 是否還有其它問题
發生(多重不良状態)。
2.4 State the specification requirements. 實際状態规格需要為多少。
2.5 What is the deviation. (measurement 6mm which is out of spec.) 問题發生
偏差為多少。
37. www.anser.com.tw 版權所有 請勿翻印 37
4.3. 產品不良暫行對策之機制建構 ( sorting, repair, replace)
3.1 Define the suspect product. 定義嫌疑批產品。
Do some experiment if necessary 如有必要需做一些實驗找出嫌疑批
3.2 All stock locations purged/sequestered of suspect stock.
嫌疑批所有庫存的位置需做隔離。如:
- Supplier site. 供應商廠内储存地点
- Working in process. (include hub) 在制程當中(包含廠外倉庫)
- Ship-out. 出货旅途運送當中。
3.3 Suspect product is identified by customer bar code.
從顧客條形码識别出嫌疑批產品。( 號码 : 起 ~ 迄 )
3.4 The containment action can detect the suspect product.
從這隔離動作能够探查這個嫌疑批產品。
3.5 Inspection/Analysis results list below items.
检查或分析报表需要有以下產生下列结果。
- #Sorted. 分類(良品與不良品) - #Found. - % Defective. 不良率 %是多少 .
3.6 Containment actions are validated as effective.
证明出隔離做法實施是有效。
3.7 The containment will not introduce other issues/defects.
這项隔離做法将不會引起其它缺陷問题发生。
38. www.anser.com.tw 版權所有 請勿翻印 38
What do we mean by “Containment”? What is being done to contain the issue or problem?
-Containment does not only apply to product issues. We need to “contain” other process or
similar issues as required.
- The problem didn't suddenly appear — it's been around for a while. It has, however, only
just become apparent.
- Immediate action needs to be taken to minimize the potential for additional loss.
- The primary obligation is to prevent harm (economic or physical) to additional
customers or departments. The critical factor is immediacy.
- Containment action planning must begin as soon as a problem becomes known.
- Containment has the potential to be very costly and painful depending on the issue.
Containment for the 4 example issues includes:
1. Getting out of the hot sun and going to the doctor for treatment of a severe sun burn
2. Turning off the car engine when it begins to run roughly
3. Chasing your neighbor’s dog off your property when it has been chasing your pets
4. Taking your baby out of the room where it has just got shocked from an electrical outlet
暫行對策原則
41. www.anser.com.tw 版權所有 請勿翻印 41
4.1 Fishbone analysis completed (if applicable 用鱼骨圖分析 。
4.2 All pertinent data been provided. 所有相關资料需被各相關人员深思熟慮過了
4.3 If issue is a “repeat”, has deficiency of previous corrective action been
identified? 如果問题被“ 重複”發生 , 那表示以前的矯正行動的無效吗?
4.4 Process of continuous to ask why until true root cause is identified.
需要不断的問 5次為什麼，直到真實根源原因被確定。
4.5 List why other potential root causes has been dismissed.
列出一张表，為什麼其它潜在根源已經解除。
4.6 Reason for “escape” from system is identified.
列出原因為什麼這個系统會(侦测不到)的原因。
4.7 List both process and system root cause.
列出製程和系统的發生根源原因。
4.8 Stated root causes are causes, not symptom or effects.
陳述的根源是原因，不可说明這問题表面现象症状。
4.9 Verification of root cause has been done. Simulation done.
42. www.anser.com.tw 版權所有 請勿翻印 42
10 6 4 3 2 1
38.5 23.1 15.4 11.5 7.7 3.8
38.5 61.5 76.9 88.5 96.2 100.0
0
10
20
0
20
40
60
80
100
Defect
C ount
Per c ent
C u m %
Percent
Count
Pareto Exa mple
Pareto Analysis
1. Tool sorting
2. Mapping
3. SEM等FA工具
找出與Y最相關
之製程
找出各相關站相關
要因(Note解析到原
因，不要現象)
魚骨圖
PostingCorre
ctions
errorfillingo
ut fo
errorre
adingform
errorkeyp
unching
co
mp
uterclitch
ba
dfo
rms
o
utdatedg
uides
o
u
tdatedpro
ce
dure
s
differentdep
tdefini
Pe
rsonnel
Machines
M
aterials
Meth
ods
M
easurem
ents
Environ
m
ent
C
ause-and-EffectDiagram
流程圖
Importance of Each
Y to the Customer
(1 = not important;
10 = very
important)
3 9 6 6
Description of Y's
Speed
of
Service
Taste
Peaceful
ambience
Cost
X ID X's
X1 Bean Type 0 9 0 9 135
X2 Brew Type 9 9 0 3 126
X3 Amount of Grounds 0 9 0 3 99
X4 Cleanliness of shop 1 0 9 0 57
X5 Water temperature 0 9 0 3 99
X6
Styrofoam vs paper
designer cups
0 0 3 9 72
Correlation with Y's (0 = none; 1 =weak; 3 = moderate; 9 = strong)
Totals
(Priority)
因果分析矩陣
依團案成員客觀評
估找出與Y最相關之
要因X
進入 A Phase
1. 了解Y現況與目標之差距
2. 以適當指標如Z值
/RTY/DPU等衡量現況
3. 確認量測系統之誤差
48. 以統計手法解析之
Data Analysis to
Identify Problem,
Root Cause,
and Solution
Potential
Process
Change
Identified
Validation
Experiment
Decision
Regarding
Process
Change
Problem or
Improvement
Opportunity
Identified
Statistical Analysis Tools are used here:
These steps are part of the Change Control Process.
• Single-Sample Analysis
• Two-Sample Analysis
• Regression Analysis
Paired Analysis
55. 技术统计学
变量 Process N 平均 标准偏差
数值 A 10 84.24 2.90
B 10 85.54 3.65
假设检(例)
實務上我們會問:
能说改善工程 B的数值比原有工程A的数值好吗?
统计上我們會問:
工程B 的平均(85.54)和工程 A 的平均(84.24)差異， 在统计上是否有意
的差異? 或者，平均差異只是随时间变动而出现的差異?
继续
Two-sample T for 工程 A vs 工程 B
N Mean StDev SE Mean
工程 A 10 84.24 2.90 0.92
工程 B 10 85.54 3.65 1.2
Difference = mu (工程 A) - mu (工程 B) Estimate for difference: -1.30000
95% CI for difference: (-4.41120, 1.81120)
T-Test of difference = 0 (vs not =): T-Value = -0.88 P-Value = 0.390 DF
P-Value > 0.05 没差異
Minitab
檢定
59. www.anser.com.tw 版權所有 請勿翻印 62
CAUSES:
Root Cause - guidance
What do we mean by “Root Cause”?
What is the root cause of the issue or problem?
- While it's desirable that corrective action be taken as quickly as possible, it's just as
important that the true root cause of the problem is identified and fixed.
-If the previous step - that of containment - was accomplished effectively, then downstream
processes and the customer have been protected from further loss. Sufficient time should be
spent in causal analysis so that the problem is fixed, fixed once, and stays fixed.
-
Root Cause for the 4 example issues could be:
1. Extended exposure to direct sun light caused severe sun burn
2. Debris in the engine oil and low oil level caused excessive engine wear
3. Lack of proper restraints allowed neighbor’s dogs to run free and chase your pets
4. Babies don’t understand that sticking their tongue in an electrical outlet will shock them
60. www.anser.com.tw 版權所有 請勿翻印 63
Discipline 4 DECRIBE ROOT CAUSE 根本原因描述
ROOT CAUSE 發生原因
WHY WAS NOT
DETECTED為何未被偵
測到
責屬
單位
責屬人
員
描述問题的根本原因
如果失效是由於元器件的失效，该元件失效的根本原因也需要描述
使用問题解决工具，如鱼骨圖，制程圖等来找出根本原因
第四步完成日期
62. www.anser.com.tw 版權所有 請勿翻印 65
5.1 Alternatives are examined/screened. 提出方法可選擇性来執行检查 / 监控。
5.2 Poka Yoke solutions are considered. 需使用防呆方式考量。
5.3 No other potential issues/defects are introduced from corrective action.
任何其它有潜在問题 / 缺陷是從這矯正行动不被導入的。
5.4 Stated actions cover all process root cause.
這项矯正措施必需包含所有生產流程的真正问题發生原因。
5.5 Stated actions cover all system root cause. System corrective action can
prevent“escapes”. 這项矯正措施必需包含所有品質系统的真正問题發生原因。
5.6 Corrective actions are demonstrated to be effective? List evidence to support it.
矯正行动顯示有效吗? 把证据列成表以支持它。
5.7 State the targeted implementation dates of all corrective action.
所有矯正行動目标的需填寫執行日期。
76. www.anser.com.tw 版權所有 請勿翻印 79
What do we mean by “Corrective Action”? What is being done to provide corrective
action to address the issue or problem?
- After the root cause of the problem has been determined, it's time to take
corrective action. This may involve both interim and permanent actions to deal with
all of the problems in evidence.
- Corrective action differs from the Containment action performed when the problem
became evident. Containment action is protective in nature and is done to prevent
further harm (to customer and/or process) from occurring.
- Corrective action is the implementation of a solution believed to eliminate the root
cause of a problem, defect or failure.
- Many organizations consider corrective action primarily as taken against a
product, while they consider preventive action applied against a process.
Corrective Action for the 4 example issues could include:
1. Limit the amount of time in direct sun light in order to avoid severe sun burn
2. Taking your car to the mechanic to repair damaged engine
3. Warning your neighbor to keep their dog off your property or you will shoot it
4. Installing electrical outlet covers to prevent the baby from getting shocked again
77. www.anser.com.tw 版權所有 請勿翻印 80
Discipline 5 LONG-TERM ACTION長期措施
CORRECTIVE ACTION 改善措施
RECHECK FOR RELATIONAL
PROCESS AND PRODUCT類似
製程和產品的考量
負責人
預定完成
日
是否修正□YES □NO
FMEA/FLOW
CHART/CONTROL PLAN
Department Manager/單位主管﹕ Respondor/承辦人﹕
提供长期的预防對策,包括:
针对根本原因的预防对策
進行總的检查以消除任何潜在的相似问题
更改有關部門的指導書/程序
對相關人员提供必要的培训
對相關的ECR需要客戶的批准
第六步完成日期
78. 永久改善行動的對策實施與效果確認
請勿以此步確認 D3 的有效性
具體實施方案完成後，即須按決定的方案去實施每項改善案，並儘量以
分工合作的方式在限制的期限內完成.
7. D6:Implement Permanent Corrective Action
執行永久對策及效果確認
7.1 All process documents are revised and deployed.
所有生產流程文件需被修正與展開。
7.2 Are process corrective actions applicable elsewhere?
所有生產流程在其它地方是矯正行動可適用吗?
7.3 Are system corrective actions applicable elsewhere?
品質系统在其它地方是矯正行動可適用吗?
7.4 Identify the target date for accomplishing the prevent actions from above.
從上述7.1~7.3 ，完成预防行動必须確定這個目標日期。
80. 以統計手法驗証之
Data Analysis to
Identify Problem,
Root Cause,
and Solution
Potential
Process
Change
Identified
Validation
Experiment
Decision
Regarding
Process
Change
Problem or
Improvement
Opportunity
Identified
Statistical Analysis Tools are used here:
These steps are part of the Change Control Process.
• Single-Sample Analysis
• Two-Sample Analysis
• Regression Analysis
Paired Analysis
84. www.anser.com.tw 版權所有 請勿翻印 87
What do we mean by “Verification”?
What is being done to verify that the previous steps in the Dell CLCA Process have
been effective to Contain the problem, establish Root Cause, enact Corrective Action
and Preventive Action?
- Problem closure cannot occur until all changes have been validated and
documented.
- In addition, Closure activity includes knowledge sharing so that other Dell
locations can act proactively to preclude a similar problem from occurring
Verification for the 4 example issues could include monitoring or measuring:
1. . . . the degree of sun tan vs. sun burn when using a given amount of sun screen
2. . . . how dirty engine oil gets when changed on recommended cycle
3. . . . how often the neighbor’s dog gets out of its yard
4. . . . the baby’s activity when free to play around covered electrical outlets
85. www.anser.com.tw 版權所有 請勿翻印 88
Discipline 6 VERIFICATION OF EFFECTIVENESS 效果確認
審核
者
效果追
蹤日
提供证據表明對策的有效性，包括:
短期對策的有效性
effectiveness of long term preventive actions
若對策由分供應商進行，需要有分供應商/供應商製程/OQA数據進行验证
需要對整個製程進行追踪验证，包括:
文件/培训的完成
對策導入日期/良品的批号及performance
第七步完成日期
86. 指文件更新，要有以下兩種文件
1. 實際作業規範. 程序及製程來避免此問題及其他相關問再發生,
2. 並將之納入品質作業作業系統及推 廣, 如: FMEA 或CONTROL PLAN 及
General Rule 文件
8. D7:Prevent Recurrence預防再發生及標準化
Discipline 7 PREVENT RECURRENCE 再發防止
審核
者
效果追
蹤日
Case close (結案) □YES 是 □NO 否
87. www.anser.com.tw 版權所有 請勿翻印 90
What do we mean by “Preventive Action”?
What is being done to provide preventive action to address the issue or problem?
- Action taken to prevent known issues before they occur or reoccur.
- Determine how action(s) can be applied to other areas.
- Implement method/system to prevent this and like issues.
- Preventive action is done to assure that the problem won't recur in like products or
processes.
- “Lock in” the new system.
- Update FMEAs, Control Plans, Procedures, Work Instructions, Processes.
PRP evaluation; can anything be done in development to prevent this and like issues from
occurring in similar products?
- Many organizations consider preventive action primarily as taken against a process, while
they consider corrective action applied against a product.
Preventive Action for the 4 example issues could include:
1. Putting on sun screen to prevent sun burn or skin cancer
2. Regularly changing your car’s engine oil (and adding oil) to prevent excessive engine wear
3. Putting up a fence to prevent the neighbor’s dog from chasing your pets
4. Installing electrical outlet covers to prevent the baby from getting shocked in the first place
88. 修正管理系統
針對現有的作業或系統做修正
另尋解決對策所替換的系統
8.1. 問題再發防止之Lesson Learn機制建構
機種 失效模式
發生階段
EVT
有效對策
DVT 送樣 OEM 市場
A
B
C
建立 EVT/DVT/PP/MP/Field產品失效資訊庫
1.1. 分析管道及資訊來源----------如何建立
1.2. 分析結果回饋-------------------如何轉換成 Design Review / FMEA作業
1.3. 定管ORT成效評估-------------如何讓 ORT 發揮把關功能
89. www.anser.com.tw 版權所有 請勿翻印 92
FA/
CA
客服
單位
CAR
審查
生產 產技 SQE 設計
Material/
Product
impact
Purge/
Sort/Rework
ECN/Purge
form/rework
Instruction
Supplier
WIP
Stock
工廠 Site
市場Field
Root Cause
Analysis
C/A’s plan &
Final Report
Review By
RCA Team
Issue still open,
continue CA review
Issue Report
to
customer for
approval
Approved
by Customer
Closed
CAR
NO
Yes
Yes
No
Yes
No
Tracking in
Production
& Field(No
same CAR
repeat)
Escalation if over
RCA Timeliness
Failure
Analysis
Start
供應商 FA
QA
Escalation if over
RCA Timeliness
CP/CIP
Web
客戶產線/市場不良
2nd day
5th day
10th day
DFMEA/PFMEA
update
8.1. 問題再發防止之Lesson Learn機制建構
90. 8D GLOBAL 8D DISCIPLINE /
FORMAT
NEW FEATURES REMARKS
both have Title
none have Date opened
none have Last date updated
none have
Product process
Infromation
none have Organization Information
none have
D0
Symptoms
Emergency Response
Action
- evaluate the need for the 8D process
- if applicable, provide Emergency Response
Action(ERA) to protect the customer
- if ERA is applied, need to have
* Verification/validation of effectiveness of
ERA
* percent(%) effectiveness
-ERA should be
assessed 100%
effective
same -description
none
D1
Establish a Team
- indicate team members designation and
telephone no.
have have
D2
Describe the
Problem
-need to define “problem statement”
- same description as normal 8D
same - description
none have
D3
Interim
Containment
Action
- Verification/validation of effectiveness of
interim containment actions should be
performed
- the identified actions
should represent the full
containment of the
problem by 100%
same - description
none have
D4
Define and Verify
Rootcauses
- verification of rootcause is required (percent
contribution should be assessed in %)
-the identified rootcause
should represent the
origin of the problem by
100%
same - description
none have
D5
Chosen
Permanent
Corrective Action
- verification of permanent corrective action is
is required
-the identified
permanent corrective
action should represent
the solution to the
problem by 100%
same
- description as normal 8D(remove interim
containment action (ICA)
none have
D6
Implement
Permanent
Corrective Action
- validation of permanent corrective action and
monitoring of long –term result are required
same
D7
Prevent Action
- description
none have
D7
Systemic Prevent
Recommendations
- introduce new system, procedure, technology
as necessary
have
D8
Team and
Individual
Recognition
-description
91. www.anser.com.tw 版權所有 請勿翻印 94
What do we mean by “Lessons Learned”? What is being done to review the
long term effectiveness of the Dell CLCA Process enacted on selected issues?
What are the Lessons Learned?
- Lessons Learned are the key fundamental lessons that can have a
significant impact when applied to prevent additional occurrences of the issue
or problem.
- Lessons Learned may be applied at the factory, at the supplier, within
engineering, or during the Phase Review Process. Lessons Learned may be
applied through implementation in any process or document that impacts
product quality or customer experience.
- Lessons Learned may incorporate many of the same steps taken for
effective Corrective and Preventive Action, but may extend beyond those
steps.
- Lessons Learned may extend the CLCA issue into a BPI project.
Lessons Learned - guidance
92. 8 D 如何納入知識管理體系
8D
FMEA
Positrol Plan
OCAP
TCM Form
Process
Control
TECN:
Containment
Action
ECN:
Permanent
Action
D3
D7
D6
8.2. 有效對策標準化系統管制作業
93. www.anser.com.tw 版權所有 請勿翻印 96
(Customer)
1Product designing requirement(or
design change)
2Product (functional) failure
3.others
Yes Yes
No No
No
Designs issue
Process issue
No
No
Data gathering
Data analysis
&
Root cause analysis
Improvement
Corrective
preventive action
form
Yes
FAE RMA QC system
RMA PE Design
Design
FAB
PFMEA DFMEA
Yes
New mass product
ORT TEST
Lesion’s learn database
New product design
rule
Market information
New product try run
New product
evaluation
Feedback Flow
8.2. 有效對策標準化系統管制作業
94. Part Function
Process Function
Potential
Failure Mode
Severity
Potential
Effect(s) of Failure
Classif. Occurrence
Potential
Cause(s)/Mechanism(s)
of Failure
Current
process Controls
Detection
1
2 3 4 5 6 7 8 9
The section of the FMEA form duplicated above has several logic
traps that cause teams to get confused and go off track.Follow this
guide if your team is confused.
1.Design FMEA ---- Be sure to list all part functions and engineering
specifications.
Process FMEA ---- Be sure to list all process functions and
requirements.
2.Failure modes occur in two major ways:
•Complete failure (fails to function).
•Partial failure (Intermittent;Overperformance;Degrades over time
faster than specified.)
3. Effects are all the downstream consequences of the failure mode.
4. Severity measures the seriousness of the effects.
5..Note critical or significant characteristics in the Classification
column.
6. In this process,Causes create failure modes,not effects.
7.Occurrence is the measure of the cause.
8. Current Design/Process Controls are directed at causes and
failure modes.
9. Detection is directed at causes and failure modes.
FMEA Logical Traps
