HALT to HASS to HASA

1

The following presentation materials are copyright protected property of

Ops A La Carte LLC.

2

How to Move from HALT to HASS to HASA

for

byMike Silverman, CRE

Managing Partner, Ops A La Carte [email protected] // www.opsalacarte.com // (408) 472-3889

3

Presenter Biography• Mike is founder and managing partner at Ops A La Carte, a Professional Consulting Company

that has in intense focus on helping customers with end-to-end reliability. Through Ops A La Carte, Mike has had extensive experience as a consultant to high-tech companies, and has consulted for over 125 companies including Cisco, Ciena, Siemens, Abbott Labs, and Applied Materials. He has consulted in a variety of different industries including power electronics, telecommunications, networking, medical, semiconductor, semiconductor equipment, consumer electronics, and defense.

• Mike has 20 years of reliability and quality experience. He is also an expert in accelerated reliability techniques, including HALT&HASS, testing over 500 products for 100 companies in 40 different industries. Mike has authored and published 7 papers on reliability techniques and has presented these around the world including China, Germany, and Canada. He has also developed and currently teaches 10 courses on reliability techniques.

• Mike has a BS degree in Electrical and Computer Engineering from the University of Colorado at Boulder, and is both a Certified Reliability Engineer and a course instructor through the American Society for Quality (ASQ), IEEE, and Effective Training Associates. Mike is a member of ASQ, IEEE, SME, ASME, PATCA, and IEEE Consulting Society and is an officer in the IEEE Reliability Society for Silicon Valley.

4

Ops A La Carte assists clients in developing and executing any and all elements of Reliability through the Product Life Cycle.

Ops A La Carte has the unique ability to assess a product and understand the key reliability elements necessary to measure/improve product performance and customer satisfaction.

Ops A La Carte pioneered “Reliability Integration＂ – using multiple tools in conjunction throughout each client＇s organization to greatly increase the power and value of any Reliability Program.

5

Ops A La Carte Services

Reliability Integration in the Concept Phase1. Benchmarking2. Gap Analysis3. Reliability Program and Integration Plan Development

Reliability Integration in the Design Phase1. Reliability Modeling and Predictions2. Derating Analysis/Component Selection3. Tolerance/Worst Case Analysis/Design of Experiments4. Risk Management / Failure Modes, Effects, & Criticality Analysis (FMECA)5. Fault Tree Analysis (FTA)6. Human Factors/Maintainability/Preventive Maintenance Analysis7. Software Reliability

6

Ops A La Carte Services, continued

Reliability Integration in the Prototype Phase1. Reliability Test Plan Development2. Highly Accelerated Life Testing (HALT)3. Design Verification Testing (DVT)4. Reliability Demonstration Testing5. Failure Analysis Process Setup

Reliability Integration in the Manufacturing Phase1. Highly Accelerated Stress Screening (HASS)2. On-Going Reliability Testing3. Repair Depot Setup4. Field Failure Tracking System Setup5. Reliability Performance Reporting6. End-of-Life Assessment

7

Ops A La Carte Services, continued

Reliability Training/Seminars

1. Reliability Tools and Integration for Overall Reliability Programs2.Reliability Tools and Integration in the Concept Phase3.Reliability Tools and Integration in the Design Phase4. Reliability Tools and Integration in the Prototype Phase5.Reliability Tools and Integration in the Manufacturing Phase6. Reliability Techniques for Beginners7.Reliability Statistics8. FMECA9.Certified Reliability Engineer (CRE) Preparation Course for ASQ10.Certified Quality Engineer (CQE) Preparation Course for ASQ

8

RELIABILITY INTEGRATION

“the process of seamlessly cohesively integrating reliability

tools together to maximize reliability and at the lowest

possible cost”

9

♦ Intuitively, one recognizes that there is some minimum total cost that will be achieved when an emphasis in reliability increases development and manufacturing costs while reducing warranty and in-service costs. Use of the proper tools during the proper life cycle phase will help to minimize total Life Cycle Cost (LCC).

Reliability vs. Cost

CRE Primer by QCI, 1998

10

Reliability vs. Cost, continuedC

OST

RELIABILITY

TOTAL COST CURVE

RELIABILITY PROGRAM COSTS

WARRANTY COSTS

OPTIMUM COST POINT

11

Reliability vs. Cost, continued

In order to minimize total Life Cycle Costs (LCC), a Reliability Engineer must do two things:

♦ choose the best tools from all of the tools available and must apply these tools at the proper phases of a product life cycle.

♦ properly integrate these tools together to assure that the proper information is fed forward and backwards at the proper times.

12

Reliability vs. Cost, continued

As part of the integration process, we must choose a set of tools at the heart of our program in which all other tools feed to and are fed from. The tools we have chosen for this are:

HALT and HASS

13

HALT and HASS Summary

♦ Highly Accelerated Life Testing (HALT) and Highly Accelerated Stress Screening (HASS) are two of the best reliability tools developed to date, and every year engineers are turning to HALT and HASS to help them achieve high reliability.

14

HALT and HASS Summary, continued

♦ In HALT, a product is introduced to progressively higher stress levels in order to quickly uncover design weaknesses, thereby increasing the operating margins of the product, translating to higher reliability.

♦ In HASS, a product is “screened” at stress levels above specification levels in order to quickly uncover process weaknesses, thereby reducing the infant mortalities, translating to higher quality.

15

HALT and HASS Summary, continued

This presentation shall review the best reliability tools to use in conjunction with HALT and HASS and how to integrate them together.

16

RELIABILITYINTEGRATION

TOOLS

17

Reliability Integration Tools - Summary♦ PHASE I: Concept Phase

• Reliability Integration in the CONCEPT Phase -Tools that are used in the concept phase of a project in order to define the reliability requirements of a program. Benchmarking is usually required.

• The output of this phase is the Reliability Program and Integration Plan. This plan will specify which tools to use and the goals and specifications of each. This is the plan that drives the rest of the program.

18

Reliability Integration Tools - Summary♦ PHASE II: Design Phase

• Reliability Integration in the DESIGN Phase - Tools that are used in the design phase of a project after the reliability has been defined.

• Predictions and other forms of reliability analysis are performed here.

• These tools will only have an impact on the design if they are done very early in the design process.

19

Reliability Integration Tools - Summary♦ Phase III: Prototype Phase

• Reliability Integration in the PROTOTYPE Phase -Tools that are used after a working prototype has been developed.

• This represents the first time a product will be tested.• The testing will mostly be focused at finding design

issues

20

Reliability Integration Tools - Summary♦ Phase IV: Manufacturing Phase

• Reliability Integration in the MANUFACTURING Phase - Tools here are a combination of analytical and test tools that are used in the manufacturing environment to continually assess the reliability of the product.

• The focus here will be mostly at finding process issues

21

Reliability Integration Tools - Summary♦ In this seminar, we shall concentrate on the

tools used in the Prototype and Manufacturing Phases.

♦ We offer other seminars on tools used in the Concept and Design Phases.

22

RELIABILITYINTEGRATION IN THE PROTOTYPE

PHASE

23

Reliability Integration in thePROTOTYPE Phase

• Highly Accelerated Life Testing (HALT)• Failure Reporting, Analysis and Corrective Action

System (FRACAS)• Reliability Demonstration Test

24

HIGHLY ACCELERATEDLIFE TESTING (HALT)

• Quickly discover design issues.• Evaluate & improve design margins.• Release mature product at market introduction.• Reduce development time & cost.• Eliminate design problems before release.• Evaluate cost reductions made to product.

Developmental HALT is not really a test you pass or fail, it is a process tool for the design engineers.

There are no pre-established limits.

HALT - Highly Accelerated Life Test

26

HALT, How It Works

Stre

ss

Start low and step up the stress, testing the product during the stressing

27

HALT, How It Works

Failure

Gradually increase stress level until a failure occurs

Stre

ss

28

HALT, How It Works

Stre

ss

Failure

Analys

isAnalyze the failure

29

HALT, How It Works

Stre

ss

Failure

Analys

isImproveMake temporary improvements

30

HALT, How It Works

Stre

ss

Failure

Analys

isImprove

(incr

ease

)

Increase stress and start process over

31

HALT, How It Works

Stre

ss

Failure

Analys

isImprove

(incr

ease

)Fundamental Technological

Limit

32

HALT, Why It Works

Classic S-N Diagram(stress vs. number of cycles)

Stress

N0

S0= Normal Stress conditions

N0= Projected Normal Life

S1

S2

N1N2

Point at which failures become non-relevant

S0

33

Product Operational

Specs

Stress

Upper Oper. Limit

Upper Destruct

Limit

Lower Destruct

Limit

LowerOper. Limit

Margin Improvement Process

34

Product Operational

Specs

Stress

Upper Oper. Limit

Upper Destruct

Limit

Lower Destruct

Limit

LowerOper. Limit

Margin Improvement Process

Operating Margin

Destruct Margin

35

SUMMARY OF HALT

RESULTS AT AN

ACCELERATED

RELIABILITY TEST

CENTER

36

Summary of CustomersIndustry Types Number of

CompaniesProduct Type

1 Networking Equipment 6 Electrical2 Defense Electronics 4 Electrical3 Microwave Equipment 4 Electrical4 Fiberoptics 2 Electrical5 Remote Measuring Equipment 2 Electrical6 Supercomputers 2 Electrical7 Teleconferencing Equipment 1 Electro-mechanical8 Video Processing Equipment 1 Electrical9 Commercial Aviation Electronics 1 Electrical10 Hand-held Computers 1 Electrical11 Hand-held Measuring Equipment 1 Electrical12 Monitors 1 Electrical13 Medical Devices 1 Electro-mechanical14 Personal Computers 1 Electrical15 Printers and Plotters 1 Electro-mechanical16 Portable Telephones 1 Electrical17 Speakers 1 Electro-mechanical18 Telephone Switching Equipment 1 Electrical19 Semiconductor Manufacturing 1 Electro-mechanical

TOTAL 33

37

Summary of Products by Customer Field Environment

Environment Type

Number of Products

Thermal Environment

Vibration Environment

Office 18 0 to 40°C Little or no vibration

Office with User 9 0 to 40°C Vibration only from

user of equipment

Vehicle 8 -40 to +75°C 1-2 Grms vibration, 0-200 Hz frequency

Field 7 -40 to +60°C Little or no vibration

Field with User 4 -40 to +60°C Vibration only from

user of equipment

Airplane 1 -40 to +75°C 1-2 Grms vibration, 0-500 Hz frequency

TOTAL 47

38

Summary of Results- by attribute -

Thermal Data,oC Vibration Data, Grms

Attribute LOL LDL UOL UDL VOL VDL

Average -55 -73 93 107 61 65

Most Robust -100 -100 200 200 215 215

Least Robust 15 -20 40 40 5 20

Median -55 -80 90 110 50 52

39

Summary of Results- by field environment -

Environment

Office

Office with User

Vehicle

Field

Field with User

Airplane 110 18 29-60 -90 110

124 66 69

-49 -68 81 106 62 62

-66 -81 106

76 32 36

-69 -78 116 123 121 124

-21 -50 67

-62 -80 92 118 46 52

Vibration Data, GrmsLOL LDL UOL UDL VOL VDL

Thermal Data,oC

40

Summary of Results- by product application -

Data, Grms

115 64 66

Commercial -48 -73 90 95 32 39

Field -57 -74 94

Military -69 -78 116 123 121 124

Application LOL LDL UOL UDL VOL VDL

VibrationProduct Thermal Data,oC

41

Summary of Results- by stress -

Cold Step Stress: 14%

Hot Step Stress: 17%

Rapid Thermal Transitions: 4%

Vibration Step Stress: 45%

Combined Environment: 20%

Significance:Without Combined Environment, 20% of all failures would have been missed

42

Failure Details by Stress- Cold Step Stress -

Failure Mode Qty

Failed component 9

Circuit design issue 3

Two samples had much different limits 3

Intermittent component 1

43

Failure Details by Stress- Hot Step Stress -

Failure Mode Qty

Failed component 11

Circuit design issue 4

Degraded component 2

Warped cover 1

44

Failure Details by Stress- Rapid Temperature Transitions -

Failure Mode Qty

Cracked component 1

Intermittent component 1

Failed component 1

Connector separated from board 1

45

Failure Details by Stress- Vibration Step Stress -

Failure Mode Qty Failure Mode Qty

Broken lead 43 RTV applied incorrectly 1Screws backed out 9 Potentiometer turned 1Socket interplay 5 Plastic cracked at stress point 1Connector backed out 5 Lifted pin 1Component fell out of socket 5 Intermittent component 1Tolerance issue 4 Failed component 1Card backed out 4 Connectors wearing 1Shorted component 2 Connector intermit. contact 1Broken component 2 Connector broke from board 1Sheared screws 1 Broken trace 1

46

Failure Details by Stress- Combined Environment –(combination of vibration with rapid temp transitions)

Failure Mode Qty

Broken lead 10Component fell off (non-soldered) 4Failed component 3Broken component 1Component shorted out 1Cracked potting material 1Detached wire 1Circuit design issue 1Socket interplay 1

47

HALT Flow Chart

Reliability - Highly Accelerated Life Testing (HALT) Flow

Rel

iabi

lity

Yes

No

ResearchEnvironmental

Limitations on All"Exotic"

Technologies BeingUsed

Perform HALT, TakingProduct Outside

Environmental andPerformance Specs to

Find Weakpoints

Evaluate Failures/Weaknesses and

Fix Those That AreRelevant and Cost-

Effective

Retest Product toDetermine New

Limits

Are MarginsAcceptable for

ReliabilityReqts?

Publish Results

Send failureinformation to

FRACAS

Use Results toDevelop a HASS

Profile

Use Results toDevelop aReliability

DemonstrationTest

Perform a FailureModes and EffectsAnalysis (FMEA) to

DetermineWeakpoints in a

Design

Use ReliabilityModeling/

Derating Data asInput

48

FAILURE REPORTING, ANALYSIS, AND

CORRECTIVE ACTION SYSTEM (FRACAS)

49

FRACAS


♦ This is also sometimes referred to as Closed Loop Corrective Action (CLCA) or Corrective and Preventive Action (CAPA).

♦ The purpose of the FRACAS is to provide a closed loop failure reporting system, procedures for analysis of failures to determine root cause, and documentation for recording corrective action.

50

FRACAS, continued


♦ This closed loop system should include:• Assurance that the root cause for each failure

is found and clearly defined.• Provisions to assure that effective corrective

actions are taken on a timely basis• Follow-up audits for all open failure reports,

failure analyses, and corrective action suspense dates

• Reporting all delinquencies to management♦ An integral part of the FRACAS is the failure

review board (FRB). The FRB is responsible for initiating and reviewing corrective action to ensure reliability improvement.

51

♦ When performing HALT, failures are identified and each must be taken to root cause. FRACAS is the perfect tool for this. A FRACAS can:• Help classify failures as to their relevancy• Help choose the appropriate analysis tool• Keep track of the progress on each open issue• Help communicate results with other departments and

outside the company

FRACAS: How to use in conjunction with a HALT

52

♦ A FRACAS can help classify failures as to their relevancy• During HALT, many failures are likely to be

uncovered. However, not all failures will be relevant. The FMECA process will find many of these non-relevant failures, but for those that are first found in HALT, a FRACAS will help make the determination of the relevancy by use of a variety of tools.

FRACAS: How to use in conjunction with a HALT, continued

53

♦ When performing a failure analysis, there are many tools that can be helpful. Some of these are:• Fault Tree Analyses (FTA’s)• Fishbone diagrams• Pareto charts• Designs of Experiments• Tolerance Analyses


54

♦ A FRACAS can keep track of the progress on each open issue• Each failure is assigned a unique FRACAS Report ID• Each report requires detailed information about the

corrective action and must be signed off• During critical stages in a project, regular FRACAS

review meetings are typically held


55


♦ A FRACAS can help communicate results with other departments and outside the company• FRACAS databases are typically kept on a network

drive for general viewing• FRACAS can be sent to a vendor to track failure

analysis• FRACAS can be used to communicate with customers

on product development or field issues

56

Reliability - Failure Reporting Analysis and Corrective Action System (FRACAS) FlowR

elia

bilit

y

Duplicate Failure, ifpossible

ImplementCorrective Action Test Solution Did Solution Fix

Problem?

Yes

Report Findingsand

Recommendations

No

Report Solution andClose Failure

Analysis

MonitorEffectiveness of

Solution / PerformVerification HALT

Gather FailureInformation

Contact Customeror Supplier (ifappropriate) toInform Them of

Plan

Analyze Failure toRoot Cause

Send Sample ofFailure Back to

ComponentManufacturer (if

appropriate)

Develop FailureAnalysis Plan forSpecific Failure

Including ResourcePlan

TrendDiscovered in

RepairCenter

FailureDiscovered

in HALTProcess

FailureDiscoveredin HASSProcess

Modify HASSProfile, if necessary

FRACAS Flow Chart

57

RELIABILITY DEMONSTRATION

TESTING (RDT)

58

Reliability Demonstration Testing (RDT)♦ A sample of units are tested at accelerated

stresses for several months. ♦ The stresses are a bit lower than the HALT

stresses and they are held constant (or cycled constantly) rather than gradually increasing.

♦ This enables us to calculate the acceleration factor for the test.

♦ The RDT can be used to validate the reliability prediction analyses.

♦ It is also useful in finding failure modes that are not easily detected in a high time compression test such as HALT.

59

RDT, continued


60

RDT, continued


61

RDT, continued


62

RDT, continued


63

RDT, continued

Classic S-N Diagram(stress vs. number of cycles)

Stress

N0

S0= Normal Stress conditions

N0= Projected Normal Life

S1

S2

N1N2

Point at which failures become non-relevant

S0

64

RDT, continued


65

RDT, continued


66

RDT: How to Use the Results of HALT in Planning an RDT♦ Two of the most important pieces of information to

decide upon when planning an RDT is which stresses to apply and how much. From this, we can derive the acceleration factor for the test. HALT can help with both of these.• HALT will identify the effects of each stress on the

product to determine which are most applicable.• HALT will identify the margins of the product with

respect to each stress. This is critical so that the highest amount of stress is applied in the RDT to gain the most acceleration without applying too much, possibly causing non-relevant failures.

67

RDT: How to Use the Results of Reliability Predictions in Planning an RDT

♦ Another key factor in planning an RDT is the goal of the test. This is usually driven by marketing requirements, but the Reliability Prediction will help determine how achievable this is• Although the prediction may not be able to give an

exact MTBF number, it will give a number close enough to help determine how long of an RDT to run and what type of confidence in the numbers to expect.

• Many times, the reliability of the product will far exceed initial marketing requirements. If this is the case, the RDT can be planned to try to prove these higher levels. Once achieved, the published specs from marketing can be increased.

68

RDT Flow Chart

Reliability - Reliability Demonstration Testing FlowR

elia

bilit

y

YesHave Reliability

Goals BeenMet?

Publish Results

Develop Test Plan, including1. Number of Units2. Acceleration Factors3. Total Test Time4. Confidence Levels

Set up and BeginTest Monitor Results

Review ReliabilityGoals Based onMarketing Input

Input FromReliabilityModeling/Derating

Input FromHALT

69

RELIABILITYINTEGRATION IN THE

MANUFACTURING PHASE

70

Reliability Tools and Integration in the MANUFACTURING Phase

• Highly Accelerated Stress Screening (HASS)• Highly Accelerated Stress Auditing (HASA)• On-Going Reliability Testing (ORT)• Repair Depot Setup• Field Failure Tracking System• Reliability Performance Reporting• End-of-Life Assessment

71

HIGHLY ACCELERATEDSTRESS SCREENING

(HASS)

72

HASS - What Is It?

• Detect & correct PROCESS changes.• Reduce production time & cost.• Increase out-of-box quality & field reliability.• Decrease field service & warranty costs.• Reduce infant mortality rate at product introduction.• Finds failures that are not found with burn-in• Accelerates ones ability to discover process and

component problems.

HASS is not a test, it’s a process. Each product has its own process.

But...before HASS can begin, we must first HALT !!

73

Before HASS, We Must Characterize Product with HALT

♦ Before HASS, we must HALT• Even for mature products in which HASS is the

goal, HALT must be done first to characterize the product margins.

Steps Towards HASS

♦ Begin process during HALT stage (involve mfg)♦ HASS development♦ Production HASS

♦ Even before HALT is complete, we should• determine production needs and throughput• start designing and building fixture• determine which stresses to apply• obtain functional and environmental equipment• understand manpower needs• determine what level HASS will be performed

(assembly or system)• determine location of HASS (in-house or at an

outside lab or contract manufacturer)• for high volume products, determine when to switch

to an audit and what goals should be put in place to trigger this

HASS Process Is Begun Early

♦ After HALT is complete, we must• assure Root Cause Analysis (RCA) completed on all

failures uncovered • develop initial screen based on HALT results• map production fixture (thermal/vibration)• run proof-of-screen

HASS Process

HASS Process, continued♦ Proof-of-Screen Criteria

• Assure that screen leaves sufficient life in product• Assure that screen is effective

Assuring the Screen Leaves Sufficient Life

t

LOL

UOL

LDL

UDL

SPEC

Make dwells long enough to execute diagnostic suite. Execute diagnostics during entire profile.

It is highly recommended to combine six-axis vibration, tickle vibration, power cycling, other stresses with thermal. Powered on monitoring is essential.Minimum

20 passes

. . . .

♦ We run for X times more than proposed screen• When we reach end-of-life, then we can say that

one screen will leave 1 – 1/x left in the product.• Example: We recommend testing for a minimum of

20 times the proposed screen length. A failure after 20 HASS screens tells us that one screen will leave the product with 1 – 1/20 or 95% of its life.

Assuring the Screen Leaves Sufficient Life

Stress

UpperDestructLimit

LowerDestructLimit

UpperOperatingLimit

LowerOperatingLimit

ProductSpecs

H A S S

ESS

HASS Process for Wide Operating Limits

The “Ideal” HASS Profile for wideoperating limits

t

Fast Rate Thermal Make dwells long enough to execute diagnostic suite. Execute diagnostics during entire profile.

LOL

UOL

LDL

UDL

SPEC

It is highly recommended to combine six-axis vibration, tickle vibration, power cycling, other stresses with thermal. Powered on monitoring is essential.

Stress

UpperDestructLimit

LowerDestructLimit

UpperOperatingLimit

LowerOperatingLimit

ProductSpecs

Precipitation ScreenDetection

Screen

ESS

HASS Process for Narrow Operating Limits

The “Ideal” HASS Profile for narrowoperating limits

t

Fast Rate Thermal

Slow Rate ThermalLOL

UOL

LDL

UDL

SPEC

It is highly recommended to combine six-axis vibration, tickle vibration, power cycling, other stresses with thermal. Powered on monitoring is essential.

Make dwells long enough to execute diagnostic suite. Execute diagnostics during entire profile.

HASS Process Is Begun Early

♦ Production HASS• Start screening process with 4x the number of

screen cycles intended for long-term HASS• During production screening (after each production

run), adjust screen limits up and cycles down until 90% of the defects are discovered in the first 1-2 cycles.

• Monitor field results to determine effectiveness of screen. Again, adjust screen limits as necessary to decrease “escapes” to the field.

• Add other stresses, as necessary, if it is impractical to adjust screen limits any further.

It is essential that the product being tested be fully exercised and monitored for problem detection.

Typical HASS Failures♦ Poor solder quality♦ Socket failures♦ Component failures♦ Bent IC leads♦ Incorrect components♦ Improper component placement♦ Test fixture/program errors

HASS Advantages over “Burn-In”

♦ Finds flaws typically found by customers♦ Reduces production time and costs♦ Lowers warranty costs

HASS Defects by Environment

Data from Array Technology (1993).

High TempExtreme

12%

Extreme TemperatureTransitions

13%

29% Low TempExtreme

46%

Combined Temperature and6 Degree-of-Freedom Vibration

HASS – Implementation Requirements

♦ HALT for margin discovery♦ Screen development♦ Powered product with monitored tests♦ Fixturing to allow required throughput

QuaMark, 1998

HASS Cost Benefits

♦ Greatly reduced test time♦ Reduction in test equipment♦ Lower warranty costs♦ Minimized chance of product recalls

QualMark, 1998

90

HASS: How to Use the Results of FMECA and a Reliability Predictions in Planning a HASS♦ How to use the results of FMECA and a Reliability

Prediction in planning a HASS• FMECA results can identify possible wearout

mechanisms that need to be taken into account for HASS.

• Reliability Prediction results can help determine how much screening is necessary.

91

HASS: How to Use the Results of FMECA and a Reliability Predictions in Planning a HASS, continued

♦ Using FMECA results to identify possible wearout mechanisms that need to be taken into account for HASS• As we discussed in the FMECA section, certain

wearout failure modes are not easily detectable in HALT or even in HASS Development. Therefore, when wearout failure modes are present, we must rely on the results of a FMECA to help determine appropriate screen parameters.

92


♦ Using Reliability Prediction results to determine how much screening is necessary• One of the parameters of a reliability prediction is the

First Year Multiplier factor. This is a factor applied to a product based on how much manufacturing screening is being performed (or is planned for) to take into account infant mortality failures.

• The factor is on a scale between 1 and 4. No screening yields a factor of 4, and 10,000 hours of “effective” screening yields a factor of 1 (the scale is logarithmic).

93


♦ Using Reliability Prediction results to determine how much screening is necessary, continued

• Effective screening allows for accelerants such as temperature and temperature cycling.

• HASS offers the best acceleration of any known screen. Therefore, HASS is the perfect vehicle for helping to keep this factor low in a reliability prediction.

94

HASS: Using the Results of HALT to Develop a HASS Profile

♦ Using the HALT Results, we then run a HASS Development process• The process must prove there is significant life left

in the product• The process must prove that it is effective at finding

defects.

95

HASS: Linking the Repair Depot with HASS by Sending “NTF” hardware back through HASS♦ During the repair process, we may identify a large

number of “No Trouble Founds” or NTFs. HASS is the perfect vehicle for identifying if these NTFs are truly intermittent hardware problems or due to something else. Using HASS to assist with the “No Trouble Found (NTF)” issue at the Repair Depot.

Ø Difficult to implement without impacting productionØ Expensive to implement across many CM’s.Ø Difficult to cost-justify

HASA Solves All These Issues

HASS Dilemma

97

HIGHLY ACCELERATEDSTRESS AUDITING

(HASA)

98

What is HASA

♦ HASA is a Highly Accelerated Stress Audit♦ HASA is an effective audit process for

manufacturing.

99

HASA Advantages

♦ HASA combines the best screening tools with the best auditing tools.

♦ Better than ORT because it leverages off of HALT and HASS to apply a screen tailored to the product

♦ Better than HASS because it is much cheaper and easier to implement and “almost” as effective.

100

Steps to HASA

♦ HALT♦ HASS Development♦ Pilot HASS♦ HASA

101

What is HASA

♦ HASA is an audit or sampling procedure♦ HASA is intended for high volume applications in

which the emphasis is not on catching every defect but rather detecting process shifts

102

Advantage of switching from HASS to HASA?

♦ To achieve cost efficiency for High Volume production, reducing manpower, equipment, utilities, and space costs.

Risks in switching from HASS to HASA?

♦ Some defective units will be shipped to the field.♦ Corrective actions must be fast and accurate.

103

Review of HASA

♦ When can HASA be used?• Design and processes are control• HASS failure rate has become acceptable

♦ What is the advantage of a HASA program?• Cost

♦ What is the risk of a HASA• Statistical confidence

104

When should HASA be considered as an option?

Only if:HALT is completed

HASS Development is completedHASS is successful

105

Is HALT Completed?

If so:Design defects have been eliminated

Margins are knownMargins are largeDesign is robust

106

Is HASS Development completed?

If so:Screen is effective

Screen is safe

107

Is HASS implementation complete?

If so:Failure rates are acceptable

Manufacturing processes are under control

108

HASA PlanGoal is to catch shifts in processes

1. Detect degradations in process quality control2. Compare with pre-established thresholds3. Empower Corrective Action Team

109

HASA Example

Example from HP Vancouver

# units shipped per day = 1000# units tested per day = 64

90% probability of detecting a rate shift from 1% to 3% by sampling 112 units in just under 2 days

110

HASA Choices

Acceptable risk of allowing a “bad lot” to shipProbability of detecting a process shift of some amount

SAMPLING PLAN♦ Sample Size

♦ Decision Limits♦ Commitment to Action

111

HASA Summary

♦ For high volume production, HASA is the best process monitoring tool

112

HASS/HASA Flow ChartReliability - Highly Accelerated Stress Screening (HASS) Flow

Rel

iabi

lity

YesYes

No

Yes

No

No

No

YesDevelop HASSSampling Plan and

Implement

Send Failuresto FailureAnalysisProcess

Perform SampleHASS and Collect

Data

Do ResultsWarrant Staying

with SampleHASS?

Has ProductUndergone a Change

that Could AffectPerformance?

Develop a HASSProfile that Matches

ProductPerformanceCapabilities

Prove Profile UsingIterative Process ofIncreasing Stress

to MaximumPossible without

Weakening Product Perform HASS andCollect Data

Do Results/Volumes Warrant

Moving toSample?

Has ProductUndergone aChange thatCould AffectPerformance

Data from HALT

Analyze RepairData to Determine

if HASS ProfileNeeds to BeStrengthened

Perform HASSon Materialfrom Repair

Center

113

ON-GOING RELIABILITY TESTING (ORT)

114

On-Going Reliability Testing (ORT)

♦ ORT is a process of taking a sample of products off a production line and testing them for a period of time, adding the cumulative test time to achieve a reliability target. The samples are rotated on a periodic basis to:• get an on-going indication of the reliability• assure that the samples are not wearing too much

(because after the ORT is complete, the samples are shipped).

115

ORT vs RDT

♦ ORT is a very similar test to the Reliability Demonstration Test (RDT) except that the RDT is usually performed once just prior to release of the product, whereas the ORT is an on-going test rotating in samples from the manufacturing line.

♦ An ORT consists of a Planning stage and a Testing and Continual Monitoring stage. The inputs from the customer are the number of units allocated to the test, the duration that each set of units will be in the test before being cycled through, and the stress factors to be applied.

116

ORT Parameters

♦ Just as in a RDT, we must choose a goal, sample size, acceleration factors, and confidence.

♦ In addition, we must choose length of time each sample will be in ORT. Because these are shippable units, we cannot risk taking significant life out.

117

ORT Goal

♦ The goal of an ORT is to:• Ensure that the defined reliability specification,

including the MTBF, are met throughout the manufacturing life of the product.

• Verify that infant mortalities have been removed during the standard manufacturing process.

118

ORT Goal, a closer look


including the MTBF, are met throughout the manufacturing life of the product.Does it do this?

119



including the MTBF, are met throughout the manufacturing life of the product.Does it do this?The answer is “yes”, but a better question is: Is this really the most effective method of doing this?

120



including the MTBF, are met throughout the manufacturing life of the product.Does it do this?The answer is yes, but a better question is: Is this really the most effective method of doing this?Probably Not. Wouldn’t a HALT/RDT be much more effective? HALT will make the product more robust, and then RDT will measure the reliability after that. Then we perform periodic HALT’s to assure the product remains robust.

121




• Verify that infant mortalities have been removed during the standard manufacturing process.Does it do this?

122




• Verify that infant mortalities have been removed during the standard manufacturing process.Does it do this?The answer is “NO, IT DOES NOT!”ORT is ineffective for process monitoring becausea) ORT rarely finds problems because acceleration factors are typically not aggressive enough.

123




• Verify that infant mortalities have been removed during the standard manufacturing process.Does it do this?The answer is “NO, IT DOES NOT!”ORT is ineffective for process monitoring becausea) ORT rarely finds problems because acceleration factors are typically not aggressive enough.b) when problems that are found, it may be weeks later and products from that lot have already shipped.

124

Comparison Between ORT and HASA

♦ ORT Benefits over HASA • You can measure reliability at any given time

♦ HASA Benefits over ORT• Effective process monitoring tool due to ability to

find failures and to timely corrective actions• Don’t need to measure on-going reliability because

reliability measurement was already done once in RDT. Also, periodic HALT is a much better vehicle for continuously monitoring reliability over time after it has been baselined.

125

REPAIR DEPOT SETUP

126

Repair Depot Setup

♦ A Repair Depot facility must be set up with the proper testing in place to reproduce the failures and to assure that the product has enough life left to be shipped back into the field.

♦ But more importantly it must be set up in such a way as to learn from the failures and make changes to the design and manufacturing processes to assure the failures are not repeated.

127

Repair Depot Setup

♦ Repair Depot Plan♦ Lowest Replaceable Unit (LRU) Level Analysis♦ Repair Depot Location Strategy – In-house vs. 3rd

Party♦ Repair Process

128

Repair Depot Setup

♦ Repair Depot Plan• The plan will outline the Repair Depot process from

beginning to end and the decisions that have to be made along the way, including identifying the Lowest Replaceable Units (LRUs), the location of the Repair Depot, and the Repair Process itself.

129

Repair Depot Setup

♦ Lowest Replaceable Unit (LRU) Level Analysis• Through the use of a reliability prediction and a

maintainability prediction, we shall help identify the LRU’s for the product. The choice of which subassemblies will be LRU’s is based on how often the subassembly will fail, how easy it is for the user to identify and replace the LRU, and how safe the operation is for the user and the product.

130

Repair Depot Setup

♦ Repair Depot Location Strategy – In-house vs. 3rd Party• Depending on the complexity of the product, it may

make sense have a 3rd Party vendor act as the Repair Depot. We shall help perform this analysis and evaluate the needs vs. the capabilities to determine the strategy that makes the most sense.

131

Repair Depot Setup

♦ Repair Process • Integrate with HALT results• Integrate with HASS results• Using HASS for “No Problem Found” issues• “Three Strikes” Process• Set up to feed data to the Field Failure Tracking

System

132

Repair Depot Setup

♦ Integrating the Repair Depot Center with HALT results• If we find design issues in the field and we can

duplicate these issues in the Repair Depot Center, we must feed them back into the HALT process.

• Therefore, the Repair Depot must be set up to easily feed information back to the HALT process when issues like this arise.

• The Corrective Action System is the perfect vehicle for linking these two together.

133

Repair Depot Setup

♦ Integrate the Repair Depot Center with HALT results• Once the issue is fed back into the HALT process,

we will review the type of failure and why HALT was not able to find the problem.• Under what conditions did the failure occur?• Was this something that HALT found but for which

corrective action was not implemented?• Perhaps we stopped short of reaching the

fundamental limit of technology.• Perhaps additional stresses are required.

• This will then be used for future HALT’s so that the process continually learns and adapts.

134

Repair Depot Setup

♦ Integrate the Repair Depot Center with HASS results• If we find process issues in the field and we can

duplicate these issues in the Repair Depot Center, we must feed them back into the HASS process.

• Therefore, the Repair Depot must be set up to easily feed information back to the HASS process when issues like this arise.

• The Corrective Action System is the perfect vehicle for linking these two together.

135

Repair Depot Setup

♦ Integrate the Repair Depot Center with HASS results• Once the issue is fed back into the HASS process,

we will review the type of failure and why HASS was not able to find the problem.• Was the issue a DOA or did it occur into the life of

the product? This will help determine why HASS did not catch.

• Perhaps the screen limits need to be adjusted.• Perhaps additional stresses are required.

• This information will then be used for future HASS’es so that the process continually learns and adapts.

136

Repair Depot Setup

♦ Using HASS for “No Problem Found” issues • We shall help determine how to treat “No Problem

Founds” (NPFs). • We shall also help integrate the Repair Depot

process with the HASS process so that “No Problem Found” items are sent through the HASS process so that intermittent failures can be discovered and repaired as well.

137

Repair Depot Setup

♦ “Three Strikes” Process • We shall help identify how often to allow failed

products to be sent back to the field. • A typical rule of thumb is to apply a “three strikes”

policy in which a failed product may not be returned to the field if it is has been returned 3 times.

138

Repair Depot Setup

♦ Set up the Repair Depot System to feed data to the Field Failure Tracking System • The Repair Depot Center retests products returned

from the field to confirm failures and determine root cause.

• The confirmation is then fed back to the Field Failure Tracking System so that it can be properly categorized for reliability data reporting.

139

FIELD FAILURE TRACKING SYSTEM

140

Field Failure Tracking System♦ The purpose of the Field Failure Tracking System

is to provide a system for evaluating a product’s performance in the field and for quickly identifying trends.

141

Field Failure Tracking System♦ When setting up a Field Failure Tracking System,

we must • Identify key parameters to monitor• Develop failure codes• Choose appropriate Tracking System• Implementation of the Tracking System• Integration of the Tracking System• Track trends• Educate others as to use

142

Field Failure Tracking System♦ Identify key parameters to monitor

• Before setting up a system, we must first identify what key parameters need to be monitored – Date of Shipment, Date of Return, Failure Code, Further Actions, etc. as well as what key metrics need to be calculated (DOA rate, MTBF, etc.)

143

Field Failure Tracking System♦ Failure Code Definitions

• Failure code definitions are key to the tracking system because proper coding of failures is the only way to easily identify trends.

144

Field Failure Tracking System♦ Choice of Tracking Systems

• Once we have identified all of the key parameters to capture, we will help choose the best tracking system for your particular application. Anything from Excel to Oracle can be used. Some level of customization may be needed. Special modules are also available from reliability software houses such as ReliaSoft and Relex.

• This system must link well with the Failure Analysis System.

145

Field Failure Tracking System♦ Implementation of Tracking System

• Once we have assured data integrity, then we can use the Field Failure Tracking System to • track trends• calculate reliability of the product on-going,

including

146

Field Failure Tracking System♦ Implementation of the System - Tracking Trends

• The most important step in the process is to evaluate the data, spot any trends that are developing, and to provide corrective action as needed. We will show you proper techniques for identifying trends from the given data and how to follow through the entire process.

147

Field Failure Tracking System♦ Implementation of the System - Metrics

• There are numerous metrics we can track, including • DOA rates• MTBF• Warranty returns• End-of-Life Issues (see EOL Assessment for more details)

148

Field Failure Tracking System♦ Set up to easily collect data to calculate Field MTBF

• All of these reliability calculations can be presented using• total over time• point estimates• rolling averages

• 3 month rolling average• 12 month rolling average

149

Field Failure Tracking System♦ Set up to easily collect data to calculate Field MTBF

• Rolling averages are typically the best because • they show the reliability trend• they make for an easy comparison from the time a

product starts shipping to present to show reliability growth

• they can show effectiveness of a corrective action by comparing from the time a failure is discovered to after a corrective action has been implemented.

150

Field Failure Tracking System♦ Setting up the Field Failure Tracking System to

integrate with sales support and customer service to assure data integrity • Data Integrity is key to a good Field Failure

Tracking System. We must be able to accurately determine• Date product was put into service• Date of failure• Circumstances around failure and solution

151

Field Failure Tracking System♦ Data Integrity – Date product was put into service

• This is not the same as the date of shipment• Often we have an “adder” to the ship date (e.g.

ship date + 30 days) but we must verify this is accurate• Sales support will help define this since they

know the customer installation process best• From this we need to be able to accurately

determine DOAs from products that failed soon after installation.

• We need to come up with “Definition of DOA”, or how many days after installation is a failure considered a DOA.

152

Field Failure Tracking System♦ Data Integrity – Date of failure

• The failure date is equally as critical and not always easy to determine because the customer sometimes only indicates the date the product was returned and not the date it failed• Customer service can help work with the

customers to assure that we get accurate information here

153

Field Failure Tracking System♦ Data Integrity – Circumstances around failure and

solution• The key to each field issue is the actual failure itself

and how it was solved.• Tags on product with failure information is essential• RMA numbers should also be on these tags and

these RMA numbers should tie back to database• The database should be linked to the Field Failure

Tracking Database so that all of the circumstances around the failure are known before trying to repair.• Often we will find that the true problem was

identified after hardware was returned• Sometimes numerous pieces of hardware are

pulled to solve a single problem – “Shotgunning”

154

Field Failure Tracking System♦ Integrating the Field Failure Tracking System with

the Repair Depot Center• Failed products from the field are returned to the

Repair Depot Center for confirm and to determine root cause.

• The confirmation is then fed back to the Field Failure Tracking System so that it can be properly categorized for reliability data reporting.

155

RELIABILITYPERFORMANCE

REPORTING

156

Reliability Performance Reporting♦ Reliability Performance Reporting in its simplest form

is just reporting back how we are doing against our plan. In this report, we must capture• how we are doing against our goals and against our

schedule to meet our goals ?• how well we are integrating each tool together ?• what modifications we may need to make to our plan ?

♦ In the report, we can also add information on specific issues, progress on failure analyses, and paretos and trend charts

157

Reliability Performance Reporting♦ How we are doing against our goals and against our

schedule to meet our goals ?• After collecting the field data, we then compare with our

goals and estimate how we are doing.• If we are achieving a specific goal element, we explain

what pieces are working and the steps we are going to take to assure that this continues

• If we are not achieving a specific goal element, we must understand what contributed to this and what steps we are going to take to change this• As part of this, we must understand the major

contributors to each goal element through trend plotting and failure analyses

158

Reliability Performance Reporting♦ How well we are integrating each tool together ?

• As part of an understanding the effectiveness of our reliability program, we must look at the overall program

• For example, if we stated in the plan that we were going to use the results of the prediction as input to HALT, we must describe here how we accomplished this• This can help explain the effectiveness of the HALT

so that its results can be repeated• This can help explain how the HALT can be more

effective in future programs if we overlooked or skipped some of the integration

• This will serve as documentation for future programs

159

Reliability Performance Reporting♦ What modifications we may need to make to our plan ?

• Occasionally, we may need to modify the plan• Goals may change due to new customer/marketing

requirements• We may have discovered new tools or new

approaches to using existing tools based on research• We may have developed new methods of integration

based on experimentation and research• Schedule may have changed

160

Reliability Performance Reporting♦ What modifications we may need to make to our plan ?

• If this occurs, we need to• Re-write the plan• Summarize the changes in our Reliability

Performance Report so that we can accurately capture these new elements going forward

161

END-OF-LIFEASSESSMENT

162

End-of-Life (EOL) Assessment♦ We Perform End-of-Life Assessments to

• Determine when a product is starting to wear out in case product needs to be discontinued

• Monitor preventive maintenance strategy and modify as needed

• Monitor spares requirements to determine if a change in allocation is necessary

• Tie back to End-of-Life Analysis done in the Design Phase to determine accuracy of analysis

163

End-of-Life (EOL) Assessment♦ Determining when a product is starting to wear

out in case product needs to be discontinued• In most market segments, customers don’t expect

products to last forever and would gladly replace if technological advances dictate it

• Therefore, discontinuing a product and offering an upgrade/replacement is common

• If we can calculate when a product is starting to reach end-of-life, this will help provide the cost justification for both us and our customer

164

End-of-Life (EOL) Assessment♦ Monitor preventive maintenance strategy and

modify as needed• If we decide to continue supporting the product, we

may determine our preventive maintenance strategy needs to be modified• Perhaps we didn’t anticipate a wearout

mechanism that now needs to be dealt with• Perhaps we estimated incorrectly the length of

time before wearout would begin

165

End-of-Life (EOL) Assessment♦ Monitor spares requirements to determine if a

change in allocation is necessary• Our spares allocation is partly based on our initial

Reliability Prediction and End-of-Life (EOL) Analysis.

• If our EOL Assessment is showing something that we did not predict in our EOL Analysis, then our spares allocation will need to be adjusted as a result.

166

End-of-Life (EOL) Assessment♦ Tie back to End-of-Life Analysis done in the

Design Phase to determine accuracy of analysis• For future programs, this comparison between

EOL analysis and EOL assessment is critical to understand how to modify.

• In some cases, our EOL analysis may differ because of analytical techniques. If this is the case, we can develop an adjustment factor between EOL analysis and EOL assessment and carry forward to new programs.

167

End-of-Life (EOL) Assessment♦ So now that we understand how to use EOL

Assessments, how to we actually perform one ?• An EOL Assessment uses Weibull-plotting

techniques to determine where on the “bathtub”curve we are

168

End-of-Life (EOL) Assessment

Failure Rate

Time

Infant Mortality level driven by amount of screening in mfg./characterized using a special factor in prediction

Ideal reliability at

time ofship

Steady State Reliability Level described by prediction

Onset of end-of-life (EOL)

♦ A review of the “bathtub” curve

169

End-of-Life (EOL) Assessment♦ To figure out where we are, we plot the field data

• We must “scrub” the data to • accurately determine the number of days in use

before failure• properly categorize the failure

• We must be careful and plot data by assembly type, especially if different assemblies have different wearout mechanisms. Otherwise, it will be impossible to determine a pattern

170

End-of-Life (EOL) Assessment

0

0.10

0.02

0.04

0.06

0.08

0 200.0040.00 80.00 120.00 160.00

ReliaSoft's Weibull++ 6.0 - www.Weibull.com

Failure Rate vs Time Plot

Time, (t)

Failu

re R

ate,

f(t)/

R(t)

5/2/2004 07:58CompanyMike Silverman

WeibullSince Jan 28 - (NTF-knwnissues)

W2 RRX - SRM MEDF=49 / S=0

β=2.9032, η=60.9188, ρ=0.8154

171

In this class, we learned about:• The four phases of a reliability program

• Concept• Design• Prototype• Manufacturing

• We learned about the reliability tools used in each phase and how to integrate all of the tools together

• We learned about HALT and HASS and their role in an overall reliability program

Presentation Summary

172

In the Prototype Phase, we learned about:• Highly Accelerated Life Testing (HALT)• Failure Reporting, Analysis and Corrective Action

System (FRACAS)• Reliability Demonstration Testing

and how to integrate these together and with tools from the other phases


173

In the Manufacturing Phase, we learned about:• Highly Accelerated Stress Screening (HASS)• Highly Accelerated Stress Screen Audit (HASA)• On-Going Reliability Testing (ORT)• Repair Depot Setup• Field Failure Tracking System Setup• Reliability Performance Reporting• End-of-Life Assessment

and how to integrate these together and with tools from the other phases


174

In Summary we have learned:• the power of developing realistic reliability goals

early, planning an implementation strategy, and then executing the strategy, and...

the power of integration !!


175

WHAT ARE YOUR QUESTIONS ?


176

Further Education

• For a more In-depth view of this topic and more, Mike will be teaching at:• May 18th-20th: Applied Reliability Symposium,

San Diego• Ops A La Carte is a proud sponsor of the 2005 ARS at the

Catamaran Resort on Mission Bay in San Diego, CA.

• In addition to sponsoring, we shall be giving a presentation on “Reliability Integration Across the Product Life Cycle”

• October 20th and 21st: “Essential Reliability Tools: A look at the best reliability tools being used”

• Go to www.opsalacarte.com/pages/news/news_events.htm for more details

177

For more information...

♦Contact Ops A La Carte (www.opsalacarte.com)

• Mike Silverman• (408) 472-3889 • [email protected]

Documents

HALT to HASS to HASA