Confederation of European Environmental Engineering …€¦ · Thermal shocks Sine vibration...

Environmental Testing of the Future

Reality in reliability, virtual testing & trends in physical testing Date: September 27, 2007

Location: ESA/ESTEC space centre Noordwijk, Holland Organisation: FHI, federation of technology branches / PLOT, Platform Omgevingstechnologie, Dutch

Association for Environmental Engineering

Parallel Session B - Trends in Physical Testing

ROSE Robustness Specification For Environmental Tests By Harry Roossien (21 mb)

Highly Accelerated Life Testing HALT In BARCO By Ivan Malfait, (1.3 mb)

HALT/HASS How Does It Work? By Keith Barber (600 kb)

MEOST Multiple Environment Over Stress Testing By Jan Eite Bullema, (160 kb)

Hands On HALT And HASS By Chris Peterson, (290 Kb)

Structural Durability Testing Of Commercial Vehicles Parts 1 and 2 By Simon de Cock, (660 kb & 620 kb)

Hydra Shaker In Non-Space Testing By Renato Salle, (70 kb)

Shock Robustness Of Mobile Devices With HDD By Jan Ruigrok, (1 mb)

Confederation of European Environmental Engineering Societies

Robustness Specification for Environmental tests.

Faster testing & more information on only one page?

More information on less paper?

DEVELOPERSCONSERVATIVES

INNOVATORS PIONEERS

VIRTUAL

physics

Reliability Maturity Model

• Position in Reliability Roadmap• Pass/fail vs. quality levels and reliability growth• One page overview, graphical lay out• Different stressors (levels)• Examples

– BT headset– watch

• Conclusions, lessons learned and continuation

ROSE CONCEPT

REQ: BASE (QUALITY LEVEL)

RELIABILITY GROWTH

Step-Stress

Tendency• p/fà

Levels• single test à

system thinking• simulation à

robustness• test à

customer• standards à

tailoredEND USER

Technique

SingleSimulation

SimulationPrograms

CombinedTesting

SingleRobustness

Robustness Programs

HALTHASS

production testing

Simulation Robustness Acceleration

Virtual??

Environmental Stress

UseUse

l Stre

Timedy mth yrswk

veryhigh

Simulation testingSimulation testing

Robustness testingRobustness testing

Reliable simulationHigh temperatureLow temperatureDamp heat steady stateDamp heat cyclicThermal shocksSine vibrationRandom vibrationBump & shockLife tests (bending,

switching etc.)Click ratioDust testPerspiration testDrop test etc.

Testtime reductionStress crackingReliability stress screenDamage boundary85/85 testHT drop/LT dropStep stress free fallHigh temp. vibration

Margins testing6 axes omnidirectional vibr.Humidity RSSDamage boundaryCombined testing

Comparison test philosophies

– fast testing, – levels, – “new”– no standards– destructive (TTF)– RCA always necessary– new test– combined stress

– fast testing, – levels, – “new”– based on standards– destructive (TTF)– RCA always necessary– existing tests– single stress

– longer testing,– pass/fail,– proven, – standards,– non-destructive– RCA when necessary– existing tests– Sinle stress– good simulation– field experiences,

HALT testingROSE testingSimulation testing

Goals of ROSE

• more insight in product quality and robustness• quantification of product quality & comparisons

(progress/competitors)• time reduction: faster testing and “better” results• cost reduction: prevent overkill• mmt summary: one page overview

What is ROSE

• Method to determine product robustness

– Presentation - One page overview– Destruction – Upper operating/destruct limits– Based on standard tests and equipment

• No HALT/HASS– ROSE is in between– HALT/HASS is 6 axis omnidirectional vibration– HALT/HASS is forced temperature changes

What is ROSE

Effectiveness of environmental stress at SEMCFr

yclicFI

What is ROSE

UDL/UOL

• Upper Destructive/Operating Limit– 1. time (continued testing)– 2. stress (increase level)– 3. stressor (different test parameters)

Stressors – Bump test

¨ 3 msþ 6 ms¨ 11 ms

¨ Pulse duration

¨ 15 g þ 25 gþ 30 g? necessary?¨ 40 g

¨ Impact level

1/3 or 1/6 bumps in each direction

þ 3000 bumps¨ 4500 bumps¨ 6000 bumps¨ 9000 bumps

þ Number of bumpsBump test

RemarksValuesVariablesTest

Pulse effects

à Time

Level Most energy?

3 ms 6 ms 12 ms

Stressors – Free fall test

þ room ¨ -20°C¨ +70°C

¨ Temperature

¨ rubber ¨ woodþ steel/concrete

¨ Surface

¨ 1.00 meterþ 1.50 meter¨ 1.80 meter

¨ Impact level

¨ 10 fallsþ 20 falls¨ 40 falls

þ Number of dropsDrop test

Step stress (other format)

1,5 meter 1,65 meter 1,80 meter

50 dropsç robust

30 dropsbase / robust

20 dropså base

Stressors – High temp/hum.

þ not specified¨ 5°C/min

¨ Slope

¨ noneþ 85% RH¨ 95% RH

¨ Humidity

¨ 0, -10, -20°Cþ +55, 70, 85°C

þ TemperatureTemphumtest

Step-stress approach

+20ºC

-20ºC

-10ºC

+55ºC

+70ºC

+85ºC

+95%RH

+85%RH

Stressors – Random vibration

* Use max. product temp.

þ 55°C*¨ 70°C

¨ Temperature

¨ singleþ combined with temp

¨ Type

þ fixed¨ pseudo

¨ Fixation

¨ 0,96 g rmsþ 3,13 g rms¨ tbd

¨ Impact level

¨ 3x 0,5 hoursþ 3x 2,0 hours¨ tbd

þ DurationRandom vibration test

• Between simulation tests and HALT• Understanding of failure mechanisms

• 1 page overview • Insight in quality levels• Insight in failure mechanisms

Example• HBH-DS970 • BT voice streaming headset (with display and cord/cables)

Observations

EXAMPLEValues are edited for reference and education

Hippix ES series W604

Hippix TP series W609

Hippix RTL series W628

Robustness Growth

Target ES W604 TP W609 PP W628 TTC Wxx

HIPPIX ROBUSTNESS PROGRESS

Vanguard

• Bluetooth watch• Duration:

Simulation: 3 weeksRobustness: 1 week

Original – Pass/Fail

ROSE - Robustness

For Study

And in new designs?

ROSE - Robustness

First time right ….?

Competitor

Development and application

Practice• Duration:

– Simulation: > 3 weeks (BSM)– Robustness: 3 days

• >80 % of fail modes simulation discovered– Clippix > 90%

• Levels and focus areas clear/assigned• 3/5 main field problems anticipated (ev. group)• Typical failure mechanisms determined

– design inputs

Development/continuation

• First projects: technical focus– feasibility ROSE testing

• Latter projects: time focus– from min. 3 weeks to max. 1 week

• Current projects: cost focus– ongoing (10 k to 3 k?)

Costs and pay-back

• Juran• Schneiderman

Quality level

Cost level

Inspection and Prevention Costs

Fail costs

Total QualityCosts

Traditional

Quality Costs by Juran

Quality level

Cost level

Inspection and Prevention Costs

Fail costs

Total QualityCosts

Zero defectsLearning organisation

Quality Costs by Scheiderman

Conclusions/summary

• Clear: one page overview product quality– gives insight in strong and weak parts, as well as relationships– enables comparisons (previous revisions and competitors)

• Levels: quantification (KPI) for reliability level – reliability progress measurable– quantifyable control over all product revisions

• Insight: no pass/fail, but margin to fail– knowing how far from base quality level gives insight in work to be done– no need to do all tests again, concentrate on low robust/improvement areas (focus)– shows and prevents overkill in design

• Effective: selected/limited tests• more frequent tests gives more insight in progress and focus areas • more effective testing, insight in failure mechanisms• overall shorter leadtime

• cultural fit• concrete vs discrete: no p/f, but levels (digital vs analog)

- Insight- Focus- Time/Costs

Lessons learned• Root cause analysis very important and will be more detailed• Monitoring very important• One page overview works very fine• Due time significantly reduced• Difficult to address right failure mechanisms• Additional to simulation testing (share between partners)• Other way of thinking, paying back in costs (Schneiderman)• Test labs need to be prepared (equipment/competence!!)• Predecessor of HALT• Management commitment/ filosophy to find failures (e.g. China)

Todays request

when testing, in your daily work:

• think once about what will happen with extended test• think about what will happen at changed conditions• keep an eye on ROSE and HALT, but do not forget your

current position and capabilities

• Be aware that you are using it already….

Thanks

Ruimte voor uw eigen logo!

Highly accelerated Life Testing Highly accelerated Life Testing HALT in BarcoHALT in Barco

Ivan Malfait

HALT in BarcoTopics

• Situating HALT in Barco

• How it started

• Current situation: The HALT Installation itself

• Situating HALT in the Design Process ?

• The HALT Procedure

• Practical Tips

• Operating Cost

• Other HALTs

• HASS

Situating HALT in Barco

How it started in … with “Long Term” HALTTemperature cycling only on Qualified units

Target 1: 500 hours before release of design

Target 2: 3500 hours in total

manufacturing release if:

- 500 h

- If all malfunctions have a root cause analysis and a corrective action is implemented.

15 6 987 1514131211 2019181716 24232221 5432

-10°C

-20°C

It needs to be taken into account that condensation occurs after one hour at 25°C.

HIGH ACCELERATED LIFE TEST PROFILE

POWER OFF

POWER ON

-10ºC to 70ºC: 2cycles/day.70ºC to -10ºC: 3 cycles/day.21 hours operational/day.

Current situation Short Term” HALT The installation itself

• In the lab• Chamber itself ……………• Pipes Inside ………...……..… • Safety aspect

• Outside the lab• Pipes External ……………

• Thermal isolation by Vacuum

•Liquid Nitrogen Tank …........…• 5000 L (4500 L useful)

•Telemetry …………………………•Never empty

Situating HALT in the Design Process•For new or derived products

• During the Design and before final qualification.• To compare same product of different competitors

•For finalized products• To simulate field failures

Business Case Evaluation

PreliminaryDesign

Prototype/Alpha stage

Pre-series/Beta stage

Series

Phase-out

Pre-Study

1. Board level assemblies

2. Electronic/ Electromechanical Equipment

Test objects

The HALT Procedure1. Define a HALT team with multiple disciplines (Project Leader,

Mech. resp, Elect. resp, Test executer)

Mech. Resp. and Elect. Resp: looking for root causes and implement corrective actions

2. Define the EUT and how to test (closed, open, cabling, …), fixation

3. Get the Operating Temperature and Vibration specifications (from Marketing/Sales or Customer) ………………………………………..

4. Define the Target Operational Limits for Temperature and Vibration ……………………………………………………………………………………..

5. Make an FMT (Functional Monitoring Test) and define the coverage (must cove at least the major functionality) ………….…

6. Check LN2-Volume left before extensive testing.

7. Fix the EUT in the chamber ……………………………………………………….Vibration Jig: stiff (transferring vibration energy at all frequencies)

low thermal inertion (cooling and heating rate)

open structure (air flow)

low weight (G level)

minimum number of resonances

8. Connect the accelerometers and the thermocouples ……………….

9. Run at least one FMT cycle before start of HALT

Before HALT

The HALT ProcedureDuring HALT

1. Continuously perform the FMT (it is not a humidity test !)

2. Monitor all HALT Parameters as a function of time (Temp, Vibr, FMT,…)

3. List the deviations, if any

What should be corrected / what not (LCD clearance,

Deformation of plastic housings, …) ?

• Perform root cause analysis• Implement a corrective action

• Cost of corrective action (time and material) ?

• Delay in product release ?

• Risk of non-implementation ?

• Benefit for other products ?

• Fundamental limit of technology ?

• Perform a verification HALT

Note: Do not “explain issues away” as this does not improve the reliability. It is a repeated “Stress – Fail – Fix” process.

The HALT Procedure1. Thermal Step Stress Test

Start at ambient Temperature

10 °C Decrements

Minimum dwell time = 12minutes

FMT 45 °C/min

The HALT Procedure2. Rapid Temperature Cycling

3. Vibration Step Testing (VOL)

Minimum 3 thermal cycles

UTOL – 5°C

LTOL + 5°C

Minimum dwell time = 10 minutes

Start at 5 gRMS

5 gRMS Increments

Minimum dwell time = 12minutes

The HALT Procedure4. Combined Testing

5 Temperature cycles

UTOL – 5°C

LTOL + 5°C

Target is 5 complete combined test cycles.

Vibration step = (VOL – 5 gRMS) / 5

The HALT Procedure5. Thermal Step Stress Test (if no Destructive Limits are encountered

during previous testing)

The HALT ProcedureAfter HALT

1. Make report (Report should be finished as Test is finished)

• Product identification

• Description of product fixture

• Location of response sensors

• Deviations from the standard HALT process

• FMT

• Detail of occurrences of unit degradation

• Root cause analysis + Corrective actions implemented

(Resulting in Engineering Changes)

• Summary of reached levels

2. Store all HALT data (so that the test sequence can be reproduced afterwards).

3. Clean up.

4. Keep test unit (if possible).

The HALT Procedure

• HALT in Barco is considered as being successful when:

• target limits are reached,• when failures occur, the failures are understood,• corrective actions are taken,• the product limits are clearly defined and pushed

as far as possible.

Practical Tips

• have knowledge of the equipment• LN2 Storage • LN2 Pressure stabilization (how it works)

• Max cooling performance versus max efficiency• Air pressure & Flow rate• fixation of EUT & pipes

Operating Cost• Renting of the LN2 tank • LN2 Consumption• Electricity / Compressed air• Operator (Almost Full Time)

LN2 Consumption

29/Dec/2004

17/Feb/2005

8/Apr/2005

28/May/2005

17/Jul/2005

5/Sep/2005

25/Oct/2005

14/Dec/2005

2/Feb/2006

24/Mar/2006

13/May/2006

time [-]

fill V

Refill Quanti ty LN2 Consum ption / day Linear (LN2 Consumption / day) Linear (Refill Quantity)

HALT Projects doneAVIONICS

1. Unit 1: Underfill Simulation of BGA failure in the field.

Broken connections underneath QUASAR chip

2. Unit 2:

3. Unit 3:

• 3 Ethernet failures have occurred during the test:• Abnormal flickering image on the display

• The unit reboots automatically when this failure occurs.• Combo card is resetting at random (this reset is typically triggered

by the main processor board)• Front of the EUT appears to reboot at random.

•The bottom plate of keyboard tablet has become loose during the Vibration Operation Limit (VOL) test

•Coil L6 (part of 3V3 switching regulator) has become loose from the plastic footprint with a broken inductor wire finally

HALT Projects doneDEFENCE

• Unit 1:

• Unit 2:

• Light leak on top of the unit that has become worse during the Short Term HALT.• Bad contact in the LCD connector (J5) on the PDB-Board of the Unit.• Electrolytic capacitor C388 on the VPB-Board of the VCM-1102 that has shortened.• Isolation of the VECTORLINK cable on the PM-side does become loose.

• one or more backlight lamps are not functioning anymore.• Bad contacts electrolytic capacitors

HALT Projects done

MEDICAL• Unit 1: Backlight (test of cracks in Light Guide)

• Unit 2: (with witnessing)

• A vertical tab is broken after about 3 minutes • light-leaks are visible at the bottom of the display • 5 electrolytic capacitors have broken off • Optical link does not function anymore• USB link does not function anymore

Other HALTs

• Rapid Voltage variation (determination of the Voltage margin)

• Power (On/Off) cycling (5000 cycles and perform FMT after each 500 cycles)

• …

HASS•During manufacturing, based on a “finished design”• Replaces the current Temp. Screening (Burn-in)• The limits discovered during HALT are used as the basis for

setting the HASS parameters•Detect weaknesses that are possibly introduced during

manufacturing (done on a finished product).•Shorter screening time (5 hours instead of 24 or 48 hours)•Based on same equipment (with or without vibration).•Check differences between chambers (T° and Vibration

capabilities).

Business Case Evaluation

PreliminaryDesign

Prototype/Alpha stage

Pre-series/Beta stage

Series

Phase-out

Pre-Study

UTOL+20°C

UTOL-10°C

Ambient temperature

LTOL+10°CLTOL

Time [min.]

Temperature[°C]

0.8 * VOL

Vibration [Grms]

Time [min.]

12 12 12 3015 305 5 1515

0 60 120 180

15 30 305 5 30

Power to EUT

OffTime [min.]

Power interruption of 2 min.

Start of Screening

End of Screening

+45°C/min.

-45°C/min.

72 115

Precipitation Phase

Detection Phase

Steepness depends on UTOL and

12 1212

Closed EUTOpen EUT

Telemetry

Pipes Inside

LN2Exhaust

Chamber itself

Typhoon 3.0

Vibration

6 degree of freedom (3 translations and 3 rotations).

10 Hz – 5 kHz

50 GRMS min. (no load)

Thermal

-100 °C à +200 °C

Max. 70 °C/min.

Useful Volume

(91 x 91 x 89) cm³

Liquid Nitrogen Tank

Pipes External

Mechanical Fixation

VCM 1102

PM 1131

Front LCD

Back of the PM: electronics boards

are not covered

Pipe that blows underneath the panel module

2 pipes for the EUT

Both parts are not completely closed

To increase thermal transition rate

Defining the Target Operational Limits• No standard known, only guidelines, product dependent

For Barco Defined Products, the specification is the basis. • Target UTOL = Spec + 48°C• Target LTOL = Spec – 48°C• Target VOL = 20 GRMS + 12 GRMS à 35 GRMS

The Target Operational Limits• Low margins indicate poor performance (short life), • High margins indicate good performance (longer life).

Typical Medical Products, • Target UOL = +45 °C + 48 °C = + 95 °C• Target LOL = 0 °C – 48 °C = - 48 °C• Target VOL = 20 g + 12 g => 35 g

Typical Avionics Products,• Target UOL = +55 °C + 48 °C = + 103 °C• Target LOL = -25 °C – 48 °C = - 73 °C • Target VOL = 20 g + 12 g => 35 g

Typical Defence Products,• Target UOL = +63 °C + 48 °C = + 111 °C• Target LOL = -42 °C – 48 °C = - 90 °C• Target VOL = 20 g + 12 g => 35

Position of accelero’s

Measured level differs a lot from setpoint

Measured Level almost equal to setpoint

HALT HASS Multi Axis Vibration TablesWhat do we need from them?

Presented by: Keith Barber

ØVibration in 6 degrees of freedom

Degrees of freedom.In mechanical engineering it defines the number of directions that an object is free to move. For example if an object is only able to move up and down it would be described as having a single degree of freedom.

Electrodynamic vibration table

Laminate structureSteel base plate

Aluminium segmented top plate

Hammer

ØTypes of tables available

Laminate structure table

ØTypes of tables available

Cast aluminium construction table

ØEven vibration distribution in each axis

Cast aluminium tables respond almost like a flat square plate which exhibit a few dominate resonant plate bending modes, the frequencies of which are determined by the size of the particular table.Laminate tables are designed to smooth out some of the bending modes making vibration input more uniform across the table.

Vibration input uniformity across a laminate table

Typical input uniformity across tables of solid construction.

10kHz filter 2.5kHz FilterX1 = +40% of average grms X1 = +13% of averageY1 = -7% of average grms Y1 = 0% of average grmsZ1 = -5% of average grms Z1 = -2.5% of average grms

• Comparison of a laminate table PSD to an un-damped cast aluminium construction table.

Laminate Table Cast aluminium Tables

ØSmooth & repeatable vibration spectrum with a low Crest factor

In ESS, the vibration screening effect depends on displacement.The displacements at 3000 Hz & above are useless.

From Laws of Vibration:Displacement, X = g(3.13/F)2In vibration, the displacement is inversely proportional to the square of the frequency.

ØMaximum energy or grms in the right frequency bandwidth

A molecular measurement

Cast Tables

Laminate Table

Overlay of the two PSD’s shows how much more energy or grmsthe laminate table provides at lower frequencies. Also note the very high and potentially destructive Crest factor of the cast table.

ØMaximum energy or grms in the right frequency bandwidth

ØWhere is the vibration measured and more importantly what is being measured?

A table rated 50grms or 100grms means nothing unless it is accompanied by the following information. q Over what frequency bandwidth? q Where is the energy concentrated in that frequency band?q Where is it being measured?q Is it a Z axis measurement or an average of X, Y & Z?q Most importantly for a 6 DOF table. How much acceleration is there in the

Z axis compared to X & Y.

It would be very easy to produce a table to give 200grms but it would be very stiff, have a very high frequency band and very little displacement.

Unlike electrodynamic vibration tables whose random force is calculated against ISO 5344 there is currently no ISO standard against which to measure the grms or the force available from HALT HASS tables. Therefore it is difficult to compare like for like.

ØVibration measurement accuracy over a wide temperature range

Ideally the vibration measurement needs to take place on the vibration table where the Test Unit is located.

q ICP accelerometers cannot survive the -100°C to +200°C temperature range of the chamber due to the inbuilt electronics.

q Easy solution move it outside the chamber and measure the underside of the vibration table!?

q Transmissibility through the table to the product will probably mean that the Test Unit is receiving less vibration input than the control input or in a very resonant table it could be a lot more.

q Use charge type accelerometers that have been developed to withstand a greater temperature range and can be mounted on the table surface inside the chamber. These accelerometers can also be insulated to give them greater protection.

q Vibration response analysis with low mass accelerometers is usually carried out in ambient conditions or within the temperature range of this type of accelerometer.

Compression ModeCompression ModeShear ModeShear Mode

ØAccelerometer types

•Shear mode accelerometer offers the best solution because the crystal stack is better isolated from the base.

•ICP (integral electronics) accelerometers cannot survive the temperature range.

Diagrams courtesy of PCB piezotronics Inc

FAQ’s

vCan I change the PSD to reduce energy at certain frequencies?

vCan I select vibration in one axis at a time?

vCan I perform sinusoidal vibration?

vDoes a HALT HASS vibration table replace my electrodynamic table?

ØDo I need big solid expensive fixtures?

Large heavy fixtures will damp the damp the table resonances and reduce the Vibration input to the test unit.

ØDo I need big solid expensive fixtures?

Rate of Air Temperature Change +200°C to -100°C =3min =100°C/min +200°C to 0°C = 1min = 200°C/min -100°C to 0°C = approx 15 seconds = 400°C/min -100 to +200 = 3.5 min = 85°C/min.

Sigma800-38 ROC CHART

12:20 :27

12:20 :57

12:21:27

12:21 :57

12:22 :27

12:22 :57

12:23 :27

12:23 :57

12:24 :28

12:24 :58

12:25:28

12:25 :58

12:26 :28

12:26 :59

12:27 :29

12:27 :59

12:28 :29

12:28 :59

12:29:29

12:29 :59

12:30 :29

12:30 :59

ManChmTSP ChmProcT/C ChmProdT/C VibCtrlAvg

R ESET

OUTE XCEE DALM

RE SETOUTE XCEE DALM

Air Conditioning Plenum

Heater Elements

N2 Gas Exhaust

Chamber Working Area

Screening Systems Vibration Table

Air Circulation Fan Monitor

Keyboard

Isolation Air Bag Pneumatic Hammer

Control panel

Computer

Compressed Air LineDigital Valve Module

LN2 InletLN2 Jets

Accelerometer Accelerometer

Digital Valve Module Compressed Air Line

Monitor

Vibration Table

LN2 Inlet

Air Circulation Fan

Control PanelKeyboard

Computer

Chamber Working Area

N2 Exhaust

Heating Elements

Air Conditioning Plenum

LN2 Jets

Accelerometer Accelerometer

Screening Systems Vibration Table

Digital Valve Module

Isolation Air Bag Pneumatic Hammer

Compressed air lineCastors

Sigma 800

Services required for Sigma HALT/HASS chambers.Insulated N2 exhaust ducting

SIGMA Chamber

SIVL (Super Insulated Vacuum Line) supply line for LN2 services

VIT (Vacuum Insulated Tank) for bulk storage of LN2

Compressed air line Air compressor Internal or external

Power Typically 400V, 80-100Amps

Sigma 800-38

965mm x 660mm vibration table Acceleration 5-1,000Hz 14grms

5-2,000Hz 33grms5-3,000Hz 42grms5-5,000Hz 49grms

Measured on the table surface as an average of X, Y, Z axis.

Temperature Range: -100°C to +200°CTemp rate of change: 80°C/min

MEOSTMultiple Environment Over Stress Testing

Jan Eite BullemaTNO

Content

• What is MEOST?• MEOST Challenges• Practical Experience• Conclusion

MEOST the Mount Everest of Stress Testing

According to Keki R. Bhote in World Class Reliability, ISBN 0-8144-0792-7

Claims for MEOST

• Reliability levels of 10:1 to 100:1 over traditional field reliability.

• Reductions in design validation time from over 16 weeks to less than 2 days.

• Reductions in design test costs by factors of 5:1.

• Reductions in design sample sizes by factors of 10:1.

What is MEOST?

• In MEOST testing, it is not the objective to pass a product, but to fail it.

• It is only through failures that the weak links of a design can be ‘smoked out’

• Failures ‘paradoxically’ mean success.• A single stress/environment is not enough

to generate failures.

Challenges

• Definition of the Maximum Practical Over Stress Level (MPOSL)

• Combination of Stresses• Finding failure modes which are

representative for field failures• Limited number of test units (3 – 10)

Maximum Practical Over Stress Limit

200 % Destruct Stress170 % Maximum Practical Over Stress Level 130 % Operational Stress100 % Design Stress

0 % Room Ambient

Combination of Stresses

Example MEOST test plan

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91

time (hrs)

Time of DayRHTemp CyclVoltage CyclingLoad DumpField Decay

Stages of MEOST

• Single Stress Up to the Design Limit• Single Stress Up to the Maximum

Practical Over Stress Limit (MPOSL)• Prototype – Full MEOST to MPSOL)• Pilot Run• Mini-MEOST in Outgoing Production

Relative Influences of Various Stresses in Electronics

Thermal Shock

Thermal

Humidity

Corrosive

Vibration Power Cycling

Voltage Margining Frequency Margining

Thermal Shock

Thermal

Humidity

Corrosive

Vibration Power Cycling

Frequency Margining

Reasons for Solder Joint Failure

• Poor solder joint design • Poor solder joint processing • Solder material issues • Excessive stresses applied to solder joints

Practical Experience

• Accelerated Testing of Leadfree Soldered interconnects

• Traditional (Accelerated) Testing• Definition of MPOSL for soldered

interconnects

Conclusions

• Successful Application of MEOST requires experience with the application.

• Insight in Physic of Failure is required.• If applied correctly MEOST can reduce

development time. • MEOST is not sufficient for product

qualification.

Chris Peterson

Hands on HALT and HASS

Practical Advice for

Accelerated Testing

A Series of Tests is Necessary

Chris Peterson

Accelerated, But Not Instantaneous

• The following tests should be done as step stress

– Cold only

– Heat only

– Vibration only

Chris Peterson

The following guidelines apply– Change rates as fast as possible– Dwell or soak time 10 minutes or as long as it

takes to stabilize the product and run diagnostic tests

– Start at room temperature (typically about 25 °C)

– When a failure is found, go back to the previous, less stressful, step

Chris Peterson

Combined Tests• Thermal Swings• Single Environment

with 6DoF Vibration• Thermal Swings with

6DoF Vibration• Use data from single

environment tests

Chris Peterson

How Much Time Should It Take?

TEST Cold Hot Vibe Swings Heat or Cold + Vibe

Swings+

TIME 2’ 40” 3’ 35’ 3’ 40” 2’ 32” 1’ 30” 2’ 32”

Chris Peterson

Assumptions

• Cold test uses10° C steps• Hot test uses 5° C steps

– Can be 10° C throughout, or start at 10° C and change to 5° C as you get close to melting point

• Vibe test starts at 3 g’s and moves up 3 g’s each step

Chris Peterson

Further Assumptions

• 5 thermal swings for both the thermal swing test and thermal swings with 6 DoF

• Chamber high temperature goes to 200°C and -100°C

• Times assume no failure is found, which is rare. These figures would be worst case scenario

Chris Peterson

What If There Is a Failure?• In all but thermal swings tests, go back to

the previous step which will be less stressful

• If the failure is no longer evident, step back up to the step where failure occurred

• If the failure does not go away, bring product back to laboratory ambient conditions and do failure analysis

Chris Peterson

Common Mistakes• Building the prototype to “pass” the test in

a different configuration than the end product will be in

• Trying to protect the product by using padding on the vibration table –remember, scratches don’t matter because this product WILL NOT ship

• Ignoring a failure

Chris Peterson

Sensor Placement• The product thermocouple should be ON

the product NEAR or ON the most sensitive component (heat producing or one of concern)

• Accelerometer– Ideally, on product in Z axis– Secondarily, on the fixture– If first two choices are not possible, on table

top, never table bottomChris Peterson

More Cautions

• Remember that cables see the same temperature as the product – make sure that they are suitable for the testing temperatures

• If there is a fire, DO NOT open the door. Turn on cooling and the nitrogen will remove the extra oxygen, putting the fire out

Chris Peterson

Margins

Chris Peterson

HALT and HASS Comparison

Chris Peterson

Testing Levels

Chris Peterson

• Level Name • What it Means• UDL

– Upper Destruct Limit• UOL

– Upper Operating Limit• Product Specification• LOL

– Lower Operating Limit• LDL

– Lower Destruct Limit

ó Operating limit – the point at which a unit has a soft, or recoverable, failureó Destruct limit – the point at

which a unit has a hard, or unrecoverable, failureó Operating margin – the gap

between the spec and the soft failureó Destruct margin – the gap

between the spec and the hard failure

How to Calculate Temperatures

Chris Peterson

Two Styles of HASS Profile

Chris Peterson

Differences Between Screening Styles

Precipitation DetectionMore extreme temperatures, going beyond OL

Less extreme temperatures, staying within 80-85% of OL

Looking to change defects from latent (dormant) to patent (active)

Good for field returns where a latent defect is already known

Vibration levels at half of DL Tickle vibration, very low levelsStimulation SimulationHALT beneficial in advance Good field data most beneficial

Precipitation DetectionMore extreme temperatures, going beyond OL

Less extreme temperatures, staying within 80-85% of OL

Looking to change defects from latent (dormant) to patent (active)

Good for field returns where a latent defect is already known

Vibration levels at half of DL Tickle vibration, very low levelsStimulation SimulationHALT beneficial in advance Good field data most beneficial

More Differences

• For the Precipitation Screen, the ramp rate is as fast as possible– Up to 100°C/minute on air

• For the Detection Screen, a slower ramp– Typically 5° to 30°C/minute on air

• Different ramps can find different failures• Dwell time needs to be long enough for product

stabilization and for a diagnostic check before the next ramp begins

In Conclusion• Remember that HALT is a discovery

test, each one an experiment• Learn from the mistakes of others so

that you can avoid making them yourself

• HASS has double check measures so that you can verify the screening strength

Chris Peterson

Enjoy Yourself as You Test• You are making products that are:

– More reliable– Safer– Faster to market– Less expensive – Fully understood

• Each of us has a chance to make a difference

Chris Peterson

Structural durability Structural durability testing of commercial testing of commercial

vehiclesvehicles

S. de Cock

Sept 2007

11. . the DAFthe DAF Service Conditions ProtocolService Conditions Protocol

1.11.1 IntroductionIntroduction

1.21.2 derivation of a durability specification, the DAF protocolderivation of a durability specification, the DAF protocol

1.31.3 The effect ofThe effect of component location and surface treatment.component location and surface treatment.

1.41.4 Derivation of design criteria for a componentDerivation of design criteria for a component

1.51.5 The derivation of a test specification forThe derivation of a test specification for whole vehicleswhole vehicles

1.61.6 CorrelationCorrelation of durability tests with service conditions.of durability tests with service conditions.

2.2. durability test methodsdurability test methods

3. Conclusions3. Conclusions

1.1 Introductiontrucks come in many shapes and sizes...

Vehicle:(> 15 ton, global data)

6 cabin types3 engine types * 4 power ratingsGear box ratios, final drive ratio16 axle configurationsChassis: 4 wheelbases, 3 choices of load capacity2 maximum axle load ratings per axle40 tire types30 fuel tanks15 component layouts

Further options:fifth wheel, side skirts, mudguards, spare wheel,bumper/front underrun protection, retarder, …

1.1 Introduction the customer decides what his vehicle is going to look like:

1.1 Introduction derivation of design criteria

A frequently used method:• Marketing defines a notoriously severe vehicle application.• Marketing chooses a typical customer.• The Testing group performs extensive measurements.• vehicle loads are extracted for design and testing purposes.• Start the design and testing activity.

• However,…….. The DAF method is different!

1.2 derivation of a durability specification1.2 derivation of a durability specificationthe DAF protocol:the DAF protocol:

Objective:To State a durability criterion

in terms of uniquely defined tests.The criterion should be formulated in such a manner,

that vehicles which meet it will fulfill all the customners' claimed and expected requirements;

this without any degree of over-design.⇓

The Service Conditions Protocol:Translates all types of use and associated loads

to standard test maneuvers ⇓

- Design criterion for durability- Test criterion for durability

1.2 Derivation of a durability specification: 1.2 Derivation of a durability specification: the service conditionsthe service conditions•• standard vehicle life:standard vehicle life: the distance a vehicle must be able to go the distance a vehicle must be able to go

without serious defects; 90% of all vehicles will achieve thiswithout serious defects; 90% of all vehicles will achieve this mileage. mileage.

•• Country of use.Country of use.

•• gross Vehicle Weight/payloadgross Vehicle Weight/payload

•• road surface qualityroad surface quality

•• Uphill gradient (particularly for the drive line).Uphill gradient (particularly for the drive line).

•• Vehicle speed pattern (city traffic, motorway).Vehicle speed pattern (city traffic, motorway).

•• maneuver loads (Cornering, braking).maneuver loads (Cornering, braking).

•• Climate.Climate.

1.2 Derivation of a durability specification:1.2 Derivation of a durability specification: the process the process

•• Market researchMarket research ⇒⇒ vehicle conceptvehicle concept ⇒⇒ service loads/standard lifeservice loads/standard life

•• Define at leastDefine at least 1 1 “vehicle application”“vehicle application” per per

vehicle configuration/ axle layout.vehicle configuration/ axle layout. This in terms of:This in terms of:

-- Nr. of kmsNr. of kms per per road surface typeroad surface type main structuremain structure

-- Nr. ofNr. of maneuvresmaneuvres forfor

-- road gradient pattern.road gradient pattern. drive linedrive line

•• translate each road surface type to one of the tracks at the testranslate each road surface type to one of the tracks at the test site. t site.

Each of these tracks is rated in terms of vehicle damage by meEach of these tracks is rated in terms of vehicle damage by means of ans of

an equivalence number. These numbers allow the expression of an equivalence number. These numbers allow the expression of

vehicle damage in terms of an equivalent mileage on one referevehicle damage in terms of an equivalent mileage on one reference track:nce track:

“equivalent“equivalent--pave kilometerspave kilometers””

1.2 The derivation of a durability specification: the process1.2 The derivation of a durability specification: the process

•• from all vehicle applicationsfrom all vehicle applications (~ (~ 800800) ) now select the criterion:now select the criterion:

thethe application which is more severe for the component than 95application which is more severe for the component than 95% % of theseof these..•• PerPer component component determine required life anddetermine required life and the maximum chance of failure :the maximum chance of failure :

-- Safety critical: Safety critical: x % failures @ 2 x life (90 % confidence)x % failures @ 2 x life (90 % confidence)-- impaired serviceability: y % failures @ 1 x life (90 % conf.) impaired serviceability: y % failures @ 1 x life (90 % conf.) -- other: other: z % failures @ 1 x life (50 % confidence).z % failures @ 1 x life (50 % confidence).

•• translate this into a test specificationtranslate this into a test specification (nr. of samples tested, level,(nr. of samples tested, level,

nr. of load cycles)nr. of load cycles) using the standard load spectrausing the standard load spectra..

All types of road surface and maneuvers available at the test siAll types of road surface and maneuvers available at the test site.te.

The DAF test site at St. OedenrodeThe DAF test site at St. Oedenrode

1.2 Vehicle application sheetsummary of the principal elements in the service conditions.

Type: FT 75 CF

land: France

standard life:

800 000 km, 7 years

application:

regional distribution

number sold per year:100

1. vehicle weight (ton)1. vehicle weight (ton)

From To FocusGV front 4 8 7GV rear 8 10 9,5GCW 14 44 38

Varies

2. road gradient pattern2. road gradient pattern

From To %Flat 0% 2% 70Hilly 2% 6% 20Mountains 6% 10% 10

Varies

3. vehicle speed pattern3. vehicle speed pattern

From To %Local roads 0 10 15City 10 40 15Inter city 40 70 30Long distance 70 100 40

Varies

5. road surface type5. road surface type

From To Average

Bitumous type 11 10% 15% 12%type 12 12% 20% 16%type 13 20% 40% 32%

Concrete type 21 0% 0% 0%type 22 0% 0% 0%type 23 0% 0% 0%

Clinckers type 31 0% 0% 0%type 32 0% 0% 0%type 33 0% 0% 0%

Pave type 41 0% 0% 0%type 42 0% 0% 0%type 43 0% 0% 0%

0% 0% 0%etc. type xx 0% 0% 0%

0% 0% 0%Ground type 101 0% 0% 0%

type 102 0% 0% 0%type 103 0% 0% 0%

Varies

there are some 800 vehicle applications !!

P rintda tum 29 -2 -0 0T ijd 15:5 7

V oertu igtoep assin g 9P P 1910 3D a tum va n o p ste llen 01 -0 3-0 0 eq u iv ale ntiefa cto ren g elde n d v oo r:

gro e p: vo e rtu ig (ge m idd e ld )T y pe F A D 9 5 X F 35 5 2 W e g de k k-fa cto r: 3 ,5 0O m s ch rijv in g K ipp e r w eg typ e eq . % v a n eq .V o ertu ig n orm le v en sd u ur k m 5 40 00 0 fo to om sc hr ijving fak tor vo er t.n lv d pa v e-k mV o ertu ig n orm le v en sd u ur jare n 7 1 .1 a s fa lt g oe d 0,0 00 8 80 ,0 36 0L an d Is ra el 1 .2 a s fa lt m atig 0,0 06 1 15 ,0 49 6V e rko o pa an ta l p er jaa r 20 0 (ja a r 19 99 ) 1 .3 a s fa lt s le c ht 0,0 5 0 ,0 0

2 .1 b e to n g o ed 0,0 02 1 0 ,0 0A s ty p e 13 55 T 2 .2 b e to n m a tig 0,0 03 3 0 ,0 0

2 .3 b e to n s lec ht 0,0 36 8 0 ,0 01 Vo e rtu ig g ew ich t [to n ] 3 .1 k lin ke rw e g g oe d 0,0 03 0 ,0 0

v ariee rt 3 .2 k lin ke rw e g m atig 0,0 2 0 ,0 0v an to t zw pt 3 .3 k lin ke rw e g s le c ht 0,2 0 ,0 0

a sla st vo o r 0 0 18 4 .1 p a ve go ed 0,0 4 0 ,0 0a sla st ac h te r 0 0 26 4 .2 p a ve m a tig 0,1 0 ,0 0G C W 63 4 .3 p a ve slec h t 1 0 ,0 0G VW 44 5 .1 s tee n slag go e d 0,0 1 1 ,0 5 4

5 .2 s tee n slag m a tig 0,0 5 2 ,0 54 03 H ellin g sp a tro o n 5 .3 s tee n slag sle ch t 0,5 1 ,0 2 70 0

v ariee rt 6 .1 g ra ve l g oe d 0,0 1 0 ,0 0v an to t zw pt 6 .2 g ra ve l m atig 0,0 5 0 ,0 0

v la k 0 0 60 6 .3 g ra ve l s le c ht 0,5 0 ,0 0h eu v elac h tig 0 0 38 6 .4 w a s bo rd g o ed 10 0 ,0 0b erg a ch tig 0 0 2 6 .5 w a s bo rd s lec h t 3 0 ,0 0

T o taa l 1 00 % 7 .1 h a rd z an d g o ed 0,0 2 0 ,0 07 .2 h a rd z an d m atig 0,1 0 ,0 0

4 Sn e lh eid s p atro o n 7 .3 h a rd z an d s lec ht 0,5 0 ,0 0v ariee rt 8 .1 d ro og za n d g oe d 0,0 1 0 ,0 0v an to t zw pt 8 .2 d ro og za n d m a tig 0,0 5 0 ,0 0

lok a al w erk v erk e er 0 0 3 8 .3 d ro og za n d s le ch t 0,5 0 ,0 0s tad sv e rke e r in te rn 0 0 5 9 .1 n a t z an d g o ed 0,0 1 0 ,0 0s tad sv e rk. d oo rg aa nd 0 0 10 9 .2 n a t z an d m a tig 0,0 5 0 ,0 0inte rste de lijk v er ke er 0 0 70 9 .3 n a t z an d s lec h t 0,5 0 ,0 0lan g e a fs tan d ve rk ee r 0 0 12 10 .1 te rre in go ed 0,0 1 0 ,0 0

T o taa l 1 00 % 10 .2 te rre in m a tig 0,0 5 0 ,0 0R a n g ee rfa kto r 1 5 b o ch ten pe r 1 00 0 k m 10 .3 te rre in slec h t 0,5 0 ,0 0S tro e fh eid 1 0 0 % T o taa l e q . p av e k m 4 14 9

9 95 K lim a at g ee n s p ec ifiek e in fo

6 O p m erk in g en 1 0% ov e rbe lad in gF A D om bo u w na a r F T D (k o pp els ch o te l)

V errichtingen per vo ertu igno rm levensduuro u d (1 99 5 ) va n af 20 0 0, z ie [1 ]

S n elle stu urbe w e ging e n L /R 1 48 32 0 7 4 16 00 b ela d ing s g ra ad (a ls fu n ctie va n g ere d en k m 's):L an g za m e stu urbe w e ging e n L /R (10 % S S )14 83 2 7 41 60B o ch ten lo k aa l w e rk ve rk ee r L/R 1 62 00 pe rc en ta ge le eg : 50 %R a n ge ren L /R 81 00 pe rc en ta ge vo l: 25 %S turen in sta nd L/R 8 10 pe rc en ta ge ov erbe la d en : 25 %R e m m in g en ('m e tho d ie k 2 00 0 ') 3 43 03 5 4 5 92 70 1 00 % (tota al)E q . h e uv ela ch tig m o tor k m 's 4 2 12 00

E q . S ta d/b erg a ch tig pig no n k m 's 8 14 54

vehicle vehicle application application sheetsheet

FAD FAD withwithstandard life:standard life:540.000 km540.000 km

V o e r tu ig to e p a s s in g 9 P P 1 9 1 0 3D a tu m v a n o p s te lle n 0 1 -0 3 -0 0

T y p e F A D 9 5 X F 3 5 5O m s c h r ijv in g K ip p e rV o e r tu ig n o rm le v e n s d u u r k m 5 4 0 0 0 0V o e r tu ig n o rm le v e n s d u u r ja re n 7L a n d Is ra e lV e rk o o p a a n ta l p e r ja a r 2 0 0 ( ja a r 1 9 9 9 )

A s ty p e 1 3 5 5 T

1 V o e rtu ig g e w ic h t [ to n ]v a rie e r t v a n to t z w p t

a s la s t v o o r 0 0 1 8a s la s t a c h te r 0 0 2 6G C W 6 3G V W 4 4

3 H e ll in g s p a tr o o nv a rie e r t v a n to t z w p t

v la k 0 0 6 0h e u v e la c h t ig 0 0 3 8b e rg a c h t ig 0 0 2

T o ta a l 1 0 0 %

4 S n e lh e id s p a tr o o nv a rie e r t v a n to t z w p t

lo k a a l w e rk v e rk e e r 0 0 3s ta d s v e rk e e r in te rn 0 0 5s ta d s v e rk . d o o rg a a n d 0 0 1 0in te rs te d e lijk v e rk e e r 0 0 7 0la n g e a fs ta n d v e rk e e r 0 0 1 2

T o ta a l 1 0 0 %

0.01 0.1 1 10 100 1000aantal wisselingen (genormeerd op 100 bochten)

oud/huidig bochten kollektief

nieuw bochtenkollektief

schadeinhoud nieuw kollektief t.o.v. oud remkollektief

k=3 k=5 k=7 k=9met v-grens 77% 34% 20% 15%zonder v-grens 88% 35% 20% 15%

Example: the load spectrum for corneringExample: the load spectrum for cornering

T otaa l 100 %R an g eerfa kto r 15 bo chten pe r 100 0 kmS tro e fh e id 1 00 %

5 K lim a at ge en spe c ifie ke in fo

6 O p m erkin g en 10 % ov erb e lad in gF A D om b ouw na ar F T D (kopp e lsch ote l)

Verrich ting en pe r vo ertu ig n orm leven sd u u rou d (19 95) va naf 200 0, z ie [1 ]

S n elle s tuurbew eg ing en L /R 148 320 741 600La ngz am e s tuurbe w eg in gen L /R (1 0% S S )14 832 74 160B o chten lok aa l w erkverkee r L /R 16 200R an geren L /R 8 100S turen in s tan d L /R 810R em m in gen ( 'm ethod iek 200 0 ') 343 035 459 270E q . he uve la ch tig m otor km 's 421 200

E q . S tad /b erg ach tig p ig no n km 's 81 454

e q uiva le ntie fa cto re n g e ld e nd vo o r:g roe p : v oe rtu ig (ge m idd e ld )

2 W eg de k k -fa cto r: 3 ,5 0w e gty pe e q . % va n e q .

fo to om sc hr ijv in g fa kto r vo e rt.n lvd p a ve -km1.1 a sfa lt g oe d 0 ,00 0 8 8 0 ,0 36 01.2 a sfa lt m a tig 0 ,00 6 1 1 5 ,0 49 61.3 a sfa lt s le c ht 0 ,05 0 ,0 02.1 b e ton g oe d 0 ,00 2 1 0 ,0 02.2 b e ton m atig 0 ,00 3 3 0 ,0 02.3 b e ton s le c ht 0 ,03 6 8 0 ,0 03.1 k lin ke rw e g g o ed 0 ,00 3 0 ,0 03.2 k lin ke rw e g m a tig 0 ,02 0 ,0 03.3 k lin ke rw e g s lec h t 0 ,2 0 ,0 04.1 p a ve go e d 0 ,04 0 ,0 04.2 p a ve m a tig 0 ,1 0 ,0 04.3 p a ve s le ch t 1 0 ,0 05.1 s tee n sla g go e d 0 ,01 1 ,0 5 45.2 s tee n sla g m a tig 0 ,05 2 ,0 54 05.3 s tee n sla g s le ch t 0 ,5 1 ,0 2 70 06.1 g rav el g o e d 0 ,01 0 ,0 06.2 g rav el m a tig 0 ,05 0 ,0 06.3 g rav el s le ch t 0 ,5 0 ,0 06.4 w a sb ord go e d 1 0 0 ,0 06.5 w a sb ord s le ch t 3 0 ,0 07.1 h a rd za n d g o e d 0 ,02 0 ,0 07.2 h a rd za n d m a tig 0 ,1 0 ,0 07.3 h a rd za n d s le ch t 0 ,5 0 ,0 08.1 d roo g za n d g o ed 0 ,01 0 ,0 08.2 d roo g za n d m atig 0 ,05 0 ,0 08.3 d roo g za n d s lec ht 0 ,5 0 ,0 09.1 n a t za nd g oe d 0 ,01 0 ,0 09.2 n a t za nd m atig 0 ,05 0 ,0 09.3 n a t za nd s le c ht 0 ,5 0 ,0 0

1 0.1 te rre in go e d 0 ,01 0 ,0 01 0.2 te rre in m a tig 0 ,05 0 ,0 01 0.3 te rre in s le ch t 0 ,5 0 ,0 0

T o ta a l e q. p av e k m 4 14 99 9

15% 15% moderatemoderate asphaltasphalt ==0.15 x 450.000 km = 81.000 km 0.15 x 450.000 km = 81.000 km ≅≅81.000 x 0.0061 = 496 km pave 81.000 x 0.0061 = 496 km pave eqeq..

• standard life (kms)• percentage full/empty• equivalent pave km• equivalent nr. of corners• equivalent nr. of slow

corners• equiv. nr. of brake

applications

9595--percentilepercentile

vehicle applicationvehicle application

Equivalent km’s paveEquivalent km’s pave

Equivalent km paveEquivalent km pave

% % of annual of annual vehicle salesvehicle sales

95 %CumulativeCumulative % % of vehicles soldof vehicles sold

Application sheet

……………..

…………..

……………..

…………..

1.2 Derivation of a test specification; determination of the 951.2 Derivation of a test specification; determination of the 95--percentile percentile vehicle application vehicle application via translationvia translation

into equivalent pavinto equivalent pave.e.

Durability test spec: Durability test spec: NN km equivalent km equivalent pave (test site)pave (test site)

1.2 derivation of a test specification1.2 derivation of a test specificationdetermination of the 95determination of the 95--percentile vehicle application percentile vehicle application via translation into via translation into equivalent pavequivalent pavéé..

1.31.3 The effect of componentThe effect of component locationlocation andand surface treatmentsurface treatment

OneOne single testsingle test requirement forrequirement for thethe whole vehiclewhole vehicle isis impossibleimpossible::

•• ForFor differentdifferent componentscomponents the mostthe most severesevere serviceservice conditionsconditions arearesometimes found onsometimes found on differentdifferent vehicle configurationsvehicle configurations//applicationsapplications..

•• Components mayComponents may havehave veryvery differentdifferent sensitivity tosensitivity to thethe variousvarioustypes oftypes of loading conditionsloading conditions ((cornering vscornering vs.. road excitationroad excitation).).

•• In the case ofIn the case of road excitationroad excitation: the: the equivalence numbersequivalence numbers,, which which representrepresent equivalentequivalent vehicle damagevehicle damage per kmper km for eachfor each type oftype of roadroadsurfacesurface,, alsoalso turn out differentturn out different for variousfor various componentcomponent locationslocations..

•• TheThe translationtranslation of theof the variousvarious serviceservice loadload spectraspectra to anto an equivalentequivalentnumbernumber ofof cyclescycles atat one load level leads toone load level leads to a differenta different result forresult fordifferent types ofdifferent types of material surfacematerial surface finish. (finish. (WöhlerWöhler constantconstant k).k).

Slope “Slope “--k”k”

The S-n curve: the average result of a large number of fatigue tests

The SN curveThe SN curveindicates indicates Fatigue lifeFatigue lifeas aas a functionfunction ofofload cycleload cyclemagnitude.magnitude.

1.31.3 TheThe influenceinfluence of componentof component locationlocation andand material surfacematerial surface((WöhlerWöhler constantconstant k)k)

N = c.Sa-k

'goed asfalt'

0.0000001

0.000001

0.00001

0.0001

2 3 4 5 6 7 8 9k-factor

1:Cabine2:Wieloph.3:Chassis4:Assen5:Comp.6:Motoroude w aarde

'slecht beton'

0.0000001

0.000001

0.00001

0.0001

2 3 4 5 6 7 8 9k-factor

1:Cabine2:Wieloph.3:Chassis4:Assen5:Comp.6:Motoroude w aarde

The equivalence factors between track types at the test siteThe equivalence factors between track types at the test site

One single test requirement for the entire vehicle is impossibleOne single test requirement for the entire vehicle is impossible!!

Solution: perform a separate translation to the reference road Solution: perform a separate translation to the reference road surface:surface:

•• for each value of k (k = 3, 5, 7, 9) for each value of k (k = 3, 5, 7, 9) •• for each component location (6 component groups).for each component location (6 component groups).

This is less draconic than it looks: a designer is always workinThis is less draconic than it looks: a designer is always working on g on one component at a time!one component at a time!

1.4 Derivation of a design criterion for a component1.4 Derivation of a design criterion for a component

•• load spectrum is a load spectrum is a straight linestraight line(log(log--lin).lin).

•• Max. load onceMax. load onceper 10per 1066 cycles.cycles.

•• 300 cycles/km300 cycles/km(Störimpuls).(Störimpuls).

•• Max. amplitude taken Max. amplitude taken from the vehicle from the vehicle response on the response on the pothole track. pothole track.

The "European load spectrum" for poor pavement loadThe "European load spectrum" for poor pavement load

10 100 1000 10000 100000 1E+06 1E+07 1E+08 1E+09aantal wisselingen (geextrapoleerd naar 1.2 mln km's)

dimensioneren: Europees kollektief voor rechtuitrijden

beproeven: ontwerpmaatstaf (95%-punt: 6500 km pave)

The European load spectrum is more severe thanThe European load spectrum is more severe thanthe spectrum of the equivalent pave test (whole vehicle test)the spectrum of the equivalent pave test (whole vehicle test)

1.4 Derivation of a design criterion for a component1.4 Derivation of a design criterion for a component

One single test requirement for the whole vehicle is not possiblOne single test requirement for the whole vehicle is not possible. Instead a e. Instead a test specification is formulated which is more severe than 95% otest specification is formulated which is more severe than 95% of the vehicle f the vehicle population in a global sense. This type of test is used as a verpopulation in a global sense. This type of test is used as a verification on a ification on a driving vehicle to highlight things that might have been overloodriving vehicle to highlight things that might have been overlooked. ked.

Basic assumptions:Basic assumptions:•• equivalence factors are the average for the whole vehicleequivalence factors are the average for the whole vehicle•• k = 3.5 (k = 3.5 (Wöhler constantWöhler constant))•• Take the 95Take the 95--percentile point for the primary load (pave)percentile point for the primary load (pave)•• Take the 50Take the 50--percentile point for the other loads (cornering, etc).percentile point for the other loads (cornering, etc).•• percentage laden/unladen as in the 95percentage laden/unladen as in the 95--percentile vehicle applicationpercentile vehicle application

1.5 The derivation of test specifications for vehicles1.5 The derivation of test specifications for vehicles

1.6 Correlation1.6 Correlation of of durability durability tests tests with with service service conditionsconditions..In order to validate the methods as laid down in the Service ConIn order to validate the methods as laid down in the Service Conditions ditions Protocol, measurements are performed on a vehicle in the field (Protocol, measurements are performed on a vehicle in the field (Benelux, Benelux, Portugal, Poland, Russia, Finland etc.). In addition the responsPortugal, Poland, Russia, Finland etc.). In addition the response of the e of the same vehicle is measured on the proving ground test tracks at thsame vehicle is measured on the proving ground test tracks at the test site. e test site.

•• The load spectra as measured in the field are compared to thoseThe load spectra as measured in the field are compared to thoseresulting from the Service Conditions Protocol.resulting from the Service Conditions Protocol.

•• In the case of mismatches the protocol may be modified.In the case of mismatches the protocol may be modified.

Allowing a mix of several vehicle speeds and several pavement tyAllowing a mix of several vehicle speeds and several pavement types in pes in the test specification improves the match with field conditions.the test specification improves the match with field conditions.

t r a d i t i o n a l d u r a b i l i t y t e s t o n B e l g i u m b l o c k s ( 3 1 0 0 k m )

tk = 3k = 5k = 7

overtested

undertested

overtested

1.6 Test vs. Service conditions: the1.6 Test vs. Service conditions: the effect effect onon simulationsimulation qualityquality ofofallowingallowing severalseveralroadroad types types and and vehicle vehicle speeds speeds in the in the test. test.

vehicle vehicle speed is speed is ““untranslatable”untranslatable”Combitrack - optimal solution mix of testtracks

VV cabi

k=3k=5k=7

mix consisting of:369 km off-road track 25 km/h7970 km clinckerroad 30 km/h468 km clinckerroad 40 km/h199 km belgium blocks 50 km/h230 km bad asphalt 30 km/h

(no restrictions total testing time)

1.6 Correlation1.6 Correlation of of durability durability tests tests with with service service conditonsconditons..The The degree degree of of correlationcorrelation betweenbetween poor pavement poor pavement type type durability durability tests and tests and actual actual service service conditionsconditions benefitsbenefits greatlygreatly ifif::

•• Vehicle Vehicle speed speed distribution distribution in the test is made in the test is made toto ((coarselycoarsely) match) match thatthatin the field (3 in the field (3 speeds instead speeds instead of 1 of 1 for instancefor instance).).

•• More More than one pavement than one pavement type is type is allowed allowed in the test (4 in the test (4 instead instead of of just just pavépavé))

The The difference between difference between the test and thethe test and the actualactual service service conditions remains conditions remains the the single most important single most important source source of of errors errors in the in the entire process entire process of of safeguarding safeguarding structural durabilitystructural durability ((errorserrors in the order of factors)in the order of factors)

By comparison By comparison the the errors errors in the in the laboratory simulationlaboratory simulation of track tests of track tests may be may be called insignificant called insignificant ((errorserrors of of tenstens of of procentsprocents in in damagedamage).).

For many components the fatigue load is dominated by 1 or 2 prinFor many components the fatigue load is dominated by 1 or 2 principal cipal forces, in that case a component rig may be used to test the comforces, in that case a component rig may be used to test the component.ponent.

For some other components (chassis, cabin) the fatigue loading iFor some other components (chassis, cabin) the fatigue loading is a s a result of their own dynamic response to the global vehicle excitresult of their own dynamic response to the global vehicle excitation. ation. In that case a real time simulation on the whole vehicle (In that case a real time simulation on the whole vehicle (--combination!) combination!) is the only way to ensure a faithful reproduction of the serviceis the only way to ensure a faithful reproduction of the service loads. loads.

•• Test on a component rig whenever possible (simple, cheap, many Test on a component rig whenever possible (simple, cheap, many samples).samples).

•• Test on a whole vehicle rig where necessary (complex, expensiveTest on a whole vehicle rig where necessary (complex, expensive, 1 or, 1 or2 samples).2 samples).

We always test to failure.......We always test to failure.......

2. Durability test methods: rig types2. Durability test methods: rig types

rigs for whole vehicles or large subassemblies (real time simularigs for whole vehicles or large subassemblies (real time simulations)tions)•• Thee road simulator (simulation of poor road surfaces).Thee road simulator (simulation of poor road surfaces).•• The component vibration rig (simulation of poor road surfaces)The component vibration rig (simulation of poor road surfaces)•• The maneuvre load rig (cornering/braking/worksThe maneuvre load rig (cornering/braking/works-- and building site)and building site)•• the drive line rig (simulation hill climb/acceleration)the drive line rig (simulation hill climb/acceleration)

Component rigs (single amplitude load cycling, block programs)Component rigs (single amplitude load cycling, block programs)•• rigs mostly have 1 or 2 hydraulic actuators on air sprung bed prigs mostly have 1 or 2 hydraulic actuators on air sprung bed plates.lates.•• Meccano system of columns, beams, plates and ballMeccano system of columns, beams, plates and ball--/leaf joints./leaf joints.•• Hydraulic ring main, automatic pumps, 1800 l/min, 200 barHydraulic ring main, automatic pumps, 1800 l/min, 200 bar•• Mostly constantMostly constant--amplitude fatigue tests or block program tests,amplitude fatigue tests or block program tests,

gradually more “real time” excitation (“variable amplitude tegradually more “real time” excitation (“variable amplitude testing”)sting”)under computer control.under computer control.

2. Durability test methods: rig types2. Durability test methods: rig types

2. Component test on a leaf spring hand2. Component test on a leaf spring hand

Structurally nearly complete vehicle, excited by 4......8 hydrauStructurally nearly complete vehicle, excited by 4......8 hydrauliclic actuatorsactuators. .

3 turntable forces;3 turntable forces;

vertical force through vertical force through

2 actuators + X2 actuators + X--membermember

2 axle2 axle--constraintsconstraints

longit. + laterallongit. + lateral

RollRoll--excitation of frontexcitation of front-- & rear axles& rear axles

Lateral Lateral

excitation excitation

for cab for cab

front axlefront axle

2. Whole vehicle test: the maneuvre load rig

2. Whole vehicle test: the2. Whole vehicle test: the roadroad simulatorsimulator

The tires are placed The tires are placed

on wheel pans;on wheel pans;

verticalvertical excitationexcitation by by

means of hydraulicmeans of hydraulic

actuatorsactuators

2. The2. The roadroad simulator:simulator: tire excitationtire excitation

Complete combination, Semitrailer excited also (not shown here)

empty rig:6 wheel pans,semitrailer excited also

2. The2. The roadroad simulatorsimulatorexcitation-unit front tires

2. The2. The roadroad simulatorsimulator

3. conclusion3. conclusion

TheThe difference between difference between the test the test specification andspecification and thethe actualactual service service conditions conditions remains remains the single most important the single most important source source of of errors errors in the in the entire process entire process of of safeguarding structural durabilitysafeguarding structural durability ((errorserrors in the order of factors)in the order of factors)

By comparison By comparison the the errors errors in the in the laboratory simulationlaboratory simulation of track tests of track tests may be may be called insignificant called insignificant ((errorserrors of of tenstens of of procentsprocents in in damagedamage).).

by R. Salles/ETS B.V.

Summary

Vibration Testing Overview

Non-space Testing

Typical Space Test Scenarios

Hydra Shaker History

Conclusion

Shock robustness of Mobile devices with HDD

Jan Ruigrok

Content of this presentation

1. Introduction MP3-player2. Development of new test method (pendulum)

– Requirements– Solved problems

3. Robustness test of HDD– Current test– New test method– Results

4. Recommendations

1: Problem

• When do the HDD’sof MP3-players fail?– Testing current

designs– Improving designs

1: Displacement, velocity and acceleration during drop

Important parameters• Drop height• Impact velocity• Velocity change• Pulse time• Maximum

acceleration• Contact stiffness

1: Hard Disc Drive

• Operating: 200 G, 1ms Half sine pulse• Non-operating: 2000 G, 1ms half sine pulse

1: Current protection

• Buffer HDD with ‘snubbers’– Rubber cushioning

which can deform during impact

– Pulse time increases– Maximum acceleration

decreases

2: Robustness test

• Drop height• 6 orientations• Performance tests

– Turn on/off test– Data transfer rate

2: new test method

• Measure drop height• Control on impact

orientation• Acceleration

measurements– Evaluate different

protections– Analyse effect of

shock on HDD

2: Test resultsacceleration velocity

2: Problem with Pendulum• Maximum drop height of

pendulum lower than the drop height measured with the free fall test

• During the free fall little offsets occur due to rotations

• These offsets influence the maximum acceleration of the HDD

2: 3-D results

• Maximum acceleration on flat impacts

• Behaviour around flat impacts important

2: Results

2: Summary of pendulum results

• Pendulum makes it possible to have control on impact orientation

• Flat impact not always worst case

• Maximum acceleration can change due to offsets• First impact not always most severe impact • Velocity change decreases with an offset

• Applying an offset with pendulum gives the same result as the free fall test

3: Accelerations on HDD• I-pod mini

– Drop height: 50 cm– Measured: 5000 G, 0.2 ms– No failure– Spec:2000 G, 1 ms

Other players show same kind results. Much higher accelerations before failure of HDD.

3: Damage Boundary Curve

DBC• Critical velocity

change• Critical acceleration• Failure if both values

are higher.

3: Fixture

• HDD placed inside fixture

• Pendulum is used to impact fixture

• Impact time can be changed by changing contact area

3: Calibration of fixtureAcceleration measurements

on fixture• Every line represents one

contact area.• Increasing drop height,

larger acceleration• Increasing drop height,

larger velocity change• With short pulses and

high drop heights noise on acceleration measurements

3: DBC results• Supplier A

– Amax=2000 G, 1ms– ∆v =6.3 m/s

• Supplier B– Amax=1250 G, 1ms– ∆v =6.1 m/s

∆v more important for MP3-player

3: Effect of buffering

• A small buffer does not reduce the acceleration below the critical acceleration

• Only possible with a deformation of several mm

3: Effect of bouncing• Lower restitution

coefficient increases the drop height

• Drop height could be increased with factor 2.6 for Philips HDD084.

• Reducing velocity change does also reduce the maximum acceleration

3: DBC conclusions• Critical velocity can be measured with fixture• Critical acceleration can be measured if the impact

velocity of the pendulum is increased

• Velocity change more important for robustness MP3-player

• Reducing the bouncing could improve drop height

• Possible to evaluate effect of shock on HDD

4: Future research

• Higher impact velocities for pendulum. For both robustness test as for the DBC test.

• Influence of offset on velocity change• Influence of material on restitution

coefficient

Confederation of European Environmental Engineering …€¦ · Thermal shocks Sine vibration...

Documents

Transcript of Confederation of European Environmental Engineering …€¦ · Thermal shocks Sine vibration...

SOLVING STRUCTURAL VIBRATION PROBLEMS USING OPERATING ... · analysis (EMA) or “bump” testing, combined with finite element analysis (FEA). 85 SOLVING STRUCTURAL VIBRATION PROBLEMS

JCETLowCostHighDensityJCET Low Cost High Density ...m.semi.org.cn/download/Semicon China 2015 - JCET Low Cost Pkg S… · PKG Interconnection -Bump Solder Bump Cu Pillar Bump Bump

General ‘Bump’ Terms

ANSI Switchboard Meters - Crompton Canada...• Meets all the requirements of ANSI-C39.1 (1981) • Parallax error-free platform dials • Bump, shock and vibration proof • Customized

Jump, Don’t Bump!

HONEYWELL BUMP CAP

Bump to baby

Random Vibration and Transient Bump Analysis of Automotive Heavy Vehicle Levelling frame

Mr.big - Bump Ahead

Bump That Shit!

Something goes bump

Tutorial Bump Mapping

Before the bump

Bump babies

Bump 2.1.11

Brown bump stepped

Japanese culture bump

Machine Data Acquisition - Vibration Analysisazimadli.com/wp-content/uploads/040417r2-Azima-TRIO-Bro.pdf · Machine Data Acquisition ... • Bump Test o Equipment ON o Equipment OFF

Delphi Confidential Precision Bump Formation Mushroom type bump X-Sec laser ablate plated up bump Fabrication process.

Bump shirts