confederation of european environmental engineering …€¦ · thermal shocks sine vibration...
TRANSCRIPT
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
ROSE
Robustness Specification for Environmental tests.
Faster testing & more information on only one page?
More information on less paper?
DEVELOPERSCONSERVATIVES
INNOVATORS PIONEERS
sing
leSI
M
SIM
.pr
ogr.
com
b.te
stin
g
sing
le.
RO
SE
RO
SE.
prog
r.
HA
LT.
prog
r.
HA
SS.
prog
r.
PoF
&
RC
A.
VIRTUAL
FMECA
MTBF
LCA
physics
theo
retic
s
…
… .
Reliability Maturity Model
© www.reliability-test.nl
ROSE
• 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
TESTS
REQ: BASE (QUALITY LEVEL)
LEV
ELS
PASS
FAIL
MO
NIT
OR
ING
RELIABILITY GROWTH
Step-Stress
Tendency• p/fà
Levels• single test à
system thinking• simulation à
robustness• test à
customer• standards à
tailoredEND USER
time
Technique
SingleSimulation
SimulationPrograms
CombinedTesting
SingleRobustness
Robustness Programs
HALTHASS
production testing
Simulation Robustness Acceleration
?
Virtual??
ROSE
ESS
Q $tR
p
PoF
Environmental Stress
UseUse
Envi
ronm
enta
l Stre
ss
Timedy mth yrswk
low
med.
high
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
?????
SIM
ROSE
HALT
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
ee F
all/d
rop
Ele
ctro
Sta
tic D
isch
arge
s
PR
V/R
LLD
T
Dam
age
boun
dary
Cor
rosi
on
Dam
p he
at c
yclicFI
OT
Life
test
s (s
witc
hing
+SC
T)
Ther
mal
sho
ck
Dam
p he
at s
tead
y
Col
d
Oth
er
85/8
5
Dry
hea
t
Com
bine
d te
st
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
RemarksValuesVariablesTest
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
RemarksValuesVariablesTest
Step-stress approach
+20ºC
-20ºC
-10ºC
+55ºC
+70ºC
+85ºC
+95%RH
+85%RH
+85%RH
0ºC
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
RemarksValuesVariablesTest
ROSE
• 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
EXAMPLEValues are edited for reference and education
Hippix TP series W609
EXAMPLEValues are edited for reference and education
Hippix RTL series W628
EXAMPLEValues are edited for reference and education
Robustness Growth
130%
88%
114%
136%
0%
20%
40%
60%
80%
100%
120%
140%
160%
180%
200%
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
100%
Fail costs
Total QualityCosts
Traditional
AB
Quality Costs by Juran
Quality level
Cost level
Inspection and Prevention Costs
100%
Fail costs
Total QualityCosts
Zero defectsLearning organisation
A B
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
2007
Ivan Malfait
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
Situating HALT in Barco
Ruimte voor uw eigen logo!
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.
0
9
4
6 7
8
5
32
1
10 11
12
14
1317
15
16
18
10
0 1
15 6 987 1514131211 2019181716 24232221 5432
70°C
25°C
-10°C
-20°C
2
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.
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
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
BCR
PDR
CDR
FQR
POR
CLR
SRR
EOL
1. Board level assemblies
2. Electronic/ Electromechanical Equipment
Test objects
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
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.
Ruimte voor uw eigen logo!
The HALT Procedure1. Thermal Step Stress Test
LTOL
UTOL
Start at ambient Temperature
10 °C Decrements
Minimum dwell time = 12minutes
FMT 45 °C/min
Ruimte voor uw eigen logo!
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
FMT
Ruimte voor uw eigen logo!
The HALT Procedure4. Combined Testing
5 Temperature cycles
UTOL – 5°C
LTOL + 5°C
Target is 5 complete combined test cycles.
FMT
FMT
Vibration step = (VOL – 5 gRMS) / 5
Ruimte voor uw eigen logo!
The HALT Procedure5. Thermal Step Stress Test (if no Destructive Limits are encountered
during previous testing)
LTDL
UTDL
Ruimte voor uw eigen logo!
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).
Ruimte voor uw eigen logo!
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.
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
Operating Cost• Renting of the LN2 tank • LN2 Consumption• Electricity / Compressed air• Operator (Almost Full Time)
LN2 Consumption
0
1000
2000
3000
4000
5000
6000
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 [-]
LN2
Con
sum
tion
/ day
Re
fill V
olum
e [L
]
0
500
1000
1500
2000
2500
3000
Refill Quanti ty LN2 Consum ption / day Linear (LN2 Consumption / day) Linear (Refill Quantity)
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
Other HALTs
• Rapid Voltage variation (determination of the Voltage margin)
• Power (On/Off) cycling (5000 cycles and perform FMT after each 500 cycles)
• …
Ruimte voor uw eigen logo!
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
BCR
PDR
CDR
FQR
POR
CLR
SRR
EOL
Ruimte voor uw eigen logo!
HASS
UTOL+20°C
UTOL-10°C
Ambient temperature
LTOL+10°CLTOL
Time [min.]
Temperature[°C]
0.8 * VOL
0
Vibration [Grms]
Time [min.]
VOL
5
12 12 12 3015 305 5 1515
0 60 120 180
15 30 305 5 30
Power to EUT
On
OffTime [min.]
Power interruption of 2 min.
72
Start of Screening
End of Screening
+45°C/min.
-45°C/min.
72 115
Precipitation Phase
Detection Phase
Steepness depends on UTOL and
LTOL
UTOL
12 1212
72
Closed EUTOpen EUT
Ruimte voor uw eigen logo!
End
Ruimte voor uw eigen logo!
Telemetry
Ruimte voor uw eigen logo!
Pipes Inside
LN2Exhaust
Ruimte voor uw eigen logo!
Chamber itself
Model
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
Ruimte voor uw eigen logo!
Liquid Nitrogen Tank
Ruimte voor uw eigen logo!
Pipes External
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
EUT
To increase thermal transition rate
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
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.
x
z
y
Z
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.
ØEven vibration distribution in each axis
Vibration input uniformity across a laminate table
ØEven vibration distribution in each axis
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
ØEven vibration distribution in each axis
• 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
-150
-100
-50
0
50
100
150
200
250
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
TIME
TEM
PER
ATU
RE
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
-60
-40
-20
0
20
40
60
80
100
120
1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91
time (hrs)
Var
ious
( )
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
Dust
Vibration Power Cycling
Voltage Margining Frequency Margining
Thermal Shock
Thermal
Humidity
Corrosive
Dust
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+
Vibe
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
Ruimte voor uw eigen logo!
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
Ruimte voor uw eigen logo!
• 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...
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 %
vehicle vehicle application application sheetsheet
FAD FAD withwithstandard life:standard life:540.000 km540.000 km
0.01 0.1 1 10 100 1000aantal wisselingen (genormeerd op 100 bochten)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
'snel
le b
ocht
en'
(gen
orm
eerd
e be
last
ing)
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
vehicle vehicle application application sheetsheet
FAD FAD withwithstandard life:standard life:540.000 km540.000 km
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..
vehicle vehicle application application sheetsheet
FAD FAD withwithstandard life:standard life:540.000 km540.000 km
• 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
0.001
0.01
0.1
1
2 3 4 5 6 7 8 9k-factor
equi
vale
ntie
fact
or t.
o.v.
pav
e
1:Cabine2:Wieloph.3:Chassis4:Assen5:Comp.6:Motoroude w aarde
'slecht beton'
0.0000001
0.000001
0.00001
0.0001
0.001
0.01
0.1
1
2 3 4 5 6 7 8 9k-factor
equi
vale
ntie
fact
or t.
o.v.
pav
e
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
1.31.3 TheThe influenceinfluence of componentof component locationlocation andand material surfacematerial surface((WöhlerWöhler constantconstant k)k)
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!
1.31.3 TheThe influenceinfluence of componentof component locationlocation andand material surfacematerial surface((WöhlerWöhler constantconstant k)k)
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)
0
100
200
300
400
500
600
700
800
rek
lang
slig
ger i
n de
kni
k (a
mpl
itude
) [u
rek]
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 )
0
1
2
3
4
5
6
7
8
accu
bak
uitla
atde
mpe
r
bran
dsto
ftank
band
elie
r
lang
slig
ger /
AV
VV
cabi
ne
stab
ilisa
tor
cabi
ne d
war
s
Mot
or v
oor
verti
caal
mot
or a
chte
r
dwar
s
mot
or v
oor
dwar
s
fatig
ue d
amag
e so
lutio
n / t
arge
tk = 3k = 5k = 7
overtested
undertested
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
0
1
2
3
4
5
6
7
8
accu
bak
uitla
atde
mpe
r
bran
dsto
ftank
band
elie
r
lang
slig
ger /
AV
VV cabi
ne
stab
ilisa
tor
cabi
ne d
war
s
Mot
or v
oor
verti
caal
mot
or a
chte
r
dwar
s
mot
or v
oor
dwar
s
fatig
ue d
amag
e so
lutio
n / t
arge
t
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: 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
1
Shock robustness of Mobile devices with HDD
Jan Ruigrok
2
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
3
1: Problem
• When do the HDD’sof MP3-players fail?– Testing current
designs– Improving designs
4
1: Displacement, velocity and acceleration during drop
Important parameters• Drop height• Impact velocity• Velocity change• Pulse time• Maximum
acceleration• Contact stiffness
5
1: Hard Disc Drive
• Operating: 200 G, 1ms Half sine pulse• Non-operating: 2000 G, 1ms half sine pulse
6
1: Current protection
• Buffer HDD with ‘snubbers’– Rubber cushioning
which can deform during impact
– Pulse time increases– Maximum acceleration
decreases
7
2: Robustness test
• Drop height• 6 orientations• Performance tests
– Turn on/off test– Data transfer rate
8
2: new test method
• Measure drop height• Control on impact
orientation• Acceleration
measurements– Evaluate different
protections– Analyse effect of
shock on HDD
9
2: Test resultsacceleration velocity
10
2: Test resultsacceleration velocity
e=1
e=0
11
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
12
2: 3-D results
• Maximum acceleration on flat impacts
• Behaviour around flat impacts important
13
2: Results
14
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
15
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.
16
3: Damage Boundary Curve
DBC• Critical velocity
change• Critical acceleration• Failure if both values
are higher.
17
3: Fixture
• HDD placed inside fixture
• Pendulum is used to impact fixture
• Impact time can be changed by changing contact area
18
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
19
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
20
3: Effect of buffering
• A small buffer does not reduce the acceleration below the critical acceleration
• Only possible with a deformation of several mm
21
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
22
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
23
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