`
`RELIABILITY OF MOBILE PHONES
`
`Dr U Daya Perera, Nokia Mobile Phones (UK) Ltd, Ashwood House, Pembroke
`Broadway, Camberley
`
`Key Words: Mobile phones, Reliability demonstration,reliability growth, Strife Testing,
`Thermal Imaging
`
`SUMMARY AND CONCLUSIONS
`
`The paper addresses reliability aspects of mobile
`phones, also known as cellular phones. Mobile phones
`have beenin use for about a decade. Early models were
`low in technology and usage waslimited. However, the
`technology of phones has been improved significantly
`during the last few years. The market demand has been
`increasing in a similar manneroverthe last few years.It
`is anticipated to increase this trend in the foreseeable
`future.
`in this highly competitive market, more and more
`products are released to satisfy the demand.
`In this
`situation, the product developmentcycle is getting shorter
`and the product costs need to be set at competitive
`levels. Also the productfailure rate criteria is decreasing
`for successive products. Hence there is a continuous
`need to review reliability assurance and demonstration
`processes at product and component
`levels. Primary
`areas of concern are, reduction of test time, sample size
`andverification of corrective actions as soon as possible.
`The current
`reliability assurance process for new
`products basically consist of testing a sample of about 50
`phones at an elevated temperature with power cycling
`and radio transmitting at power on periods. This is a
`traditional reliability demonstration process where product
`reliability is demonstration with the assumption that the
`failure rate is constant.
`One solution to achieve the above objectives was to
`introduce Test Analyze And Fix (TAAF) process(reliability
`growth testing). and to model test datato reliability growth
`models. Another solution being addressed is to assess
`the effectiveness of applying a Sirife type of test. In Strife
`testing the aim is to search for potential failures by
`subjecting the product to multiple (cyclic) stresses and
`introduce corrective actions to eliminate recurrence
`throughfailure analysis, rather than attempting to quantify
`reliability estimates.
`It is shown that application of reliability growth models
`is effective to demonstrate the product
`reliability. The
`effect of early life failures and corrective actions can be
`quantified,
`test time can be reduced and used more
`effectively. In order to achieve wider requirements, it is
`anticipated that Strife testing can be effectively used to
`reduces the test time further, by assessing the deviation
`of parametric measures with test time.
`
`1.
`
`INTRODUCTION
`
`This paper describesthereliability aspects of mobile
`phones. Aspects include those considered at different
`stages
`of
`the
`product development;
`design
`and
`development,
`reliability demonstration, environmental
`testing. Also considered is commercial issues that effect
`the productreliability.
`is
`It
`There is a growing market for mobile phones.
`anticipated that it will grow over the next decade and the
`household acceptance would reach a similar or better
`level to the video recorder market.
`In this competitive
`market
`there is a continuing need to satisfy varying
`customer needs. It is apparent that the customer needs
`vary for different continents. In this area reliability related
`aspects play a significant factor.
`With the rapidly
`advancing technology and market requirements in mobile
`phones in a competitive market, new products are
`developed more frequently to keep customers updated
`with new technology. Consequently the product cycle
`time is getting shorter.
`fn this situation there is an
`increasing demand to reduce the product development
`and test times.
`The reliability demonstration of new products is carried
`out through the traditional testing of about 50 phonesat
`an elevated temperatures for about six weeks. In these
`tests constant failure rate is assumed. In this situation,
`test data could be easily misinterpreted. TAAF together
`with reliability growth modelling would be a moreeffective
`solution. Observed failures are subjected to failure
`analysis and introduction of corrective actions to eliminate
`recurring failures.
`Due to the decreasing failure rate criteria for new
`models, for traditional reliability demonstration tests, test
`time and test sample size increases for successive
`products. Hence, there is a continuous need to seek and
`experiment with new techniques that can reducethe test
`time. The Strife technique is considered, where test
`samples are subjected to power cycling and temperature
`cycling beyond the product specifications. The objective
`is to 'search' for potential failures and introduce corrective
`actions,
`rather
`than attempting to demonstrate the
`product
`failure
`rate.
`It
`is also attempted whether
`deviations
`in parametric measures with accelerated
`ageing can be used to assess the productreliability
`measures.
`
`0149-144X/95/$4.00 ©1995 IEEE
`1995 PROCEEDINGS Annua! RELIABILITY and MAINTAINABILITY Symposium
`
`1
`
`33
`
`Exhibit 1019
`Apple v. Qualcomm
`IPR2018-01315
`
`Exhibit 1019
`Apple v. Qualcomm
`IPR2018-01315
`
`1
`
`
`
`Another proactive measure taken on newproductsis
`to introduce online stress screening processes. The
`objective is to protect customers from early manufacturing
`related problems, particularly during the production ramp
`up periods of new products.
`
`is also interesting to note that customer
`It
`phones.
`requirements based on market surveys vary markedly
`from one country to another. For example the main
`
`:
`
`
`
`
`j
`
`
`
`
`Australia
`
`Switzerland
`
`‘
`
`UK
`
`ieee
`
`
`
`
`
`
`
`
`
`
`Taiwan
`
`Japan
`
`ee
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Malaysia
`France SKorea
`Tailland
`Maxico
`Spain
`ee
`
`
`
`aly Germany
`
`ee
`SSIISSIRSTRS
`Countries
`
`Sweden
`
`Denmark
`
`ch -eS
`
`Penatration Level (%)
`r
`
`
`
`2] >
`z\EE16his
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`pet
`fs
`
`2. PRODUCTS - MOBILE PHONES
`
`12
`
`— Qa
`
`The mobile phone is a relatively new concept of
`personnel communication, originated about twenty years
`ago. During the early period very high costs were
`involved in purchase, installation and running compared
`to the traditional hard wired phones. in this era the mobile
`phone was mainly limited to business use. However with
`advances in network and mobile phone technology there
`has been a significant accepiance of the mobile phone
`over the last five years, not only in business use but in
`the consumer market as well. In parallel, the cost factors
`involved in the use of mobile phones are decreasing
`significantly. Detailed aspects of mobile phones can be
`found in many journals relating to mobile communication,
`such asref 7.
`A key measure of acceptance of a generic consumer
`products is
`the penetration level,
`the percentage of
`products per population at any given time. As wouid be
`expected, for mobile phones this value varies from one
`country to another. Scandinavian countries are leading (at
`about 8%), (Ref 2) while the penetration level in US is
`about 5%.
`In other European countries and continents
`this varies
`from around 0.4 to 4%. However
`the
`significant factor is that penetration levels are increasing
`rapidly. This implies that there is a significant potential
`world market for a number of years. A league table of
`mobile phone penetration levels in different countries is
`illustrated in figure 1. It is anticipated that during the next
`decade, penetration level would reach similar figures to
`that experienced for video systems, around 50% in
`western countries. This implies that there is a significant
`potential demand for mobile phones in the near future.
`The world subscriber mobile phone population during
`June 1893 was approximately 25 million.
`Since introduction, the mobile phone technology has
`being increasing rapidly. Most
`recent
`technological
`improvement was the introduction of digital phones in
`1993 (ref 3), superseding the analog technology. Main
`advantages of digital over analog phones are security,
`wider coverage,
`in particular
`international calls, and
`clearer voice. As digital
`technology is relatively new,
`these phones are more expensive than analog.
`It
`is
`forecast that by end of
`the decade the analog phone
`population would be replaced bydigital technology. There
`are many different versions of the above technologies.
`These aspects are well outside the scope of this paper.
`As would be expected,
`initial mobile phones were
`much bigger than current versions. As the technology
`improves the tendency is to make smaller and slimmer
`
`Figure 1
`
`- Mobile Phone Penetration Levels (Ref 2)
`
`requirement for one country can be size and weight while
`that
`for another country can be cost, warranty or
`reliability. Therefore manufacturers need to consider
`different market
`requirements
`in
`this
`competitive
`business. However,
`to be competitive,
`it
`is vital
`that
`mobile phones satisfy competitive reliability requirements
`as specified during the product developmentperiod.
`
`2.1 Reliability and Environmental Issues
`
`The reliability requirements of mobile phones are
`challenging when
`compared to many commercial
`products. This requirementis primarily due to the need to
`satisfy vastly varying environmental! conditions. This is
`due to fact that a mobile phone can be used in; indoor,
`outdoor and in a motor vehicle. The indoor and outdoor
`environmental conditions can vary significant from one
`couniry or continent to another. Also the environmental
`conditions within a vehicle can significantly vary in
`different countries. These can be temperature and or
`humidity
`extremes
`and
`dusty conditions. Typical
`environmental conditions in which a mobile phone needs
`to operate are:
`Temperature range -30 to +60 deg C
`Other environmental factors that need to be addressed
`include:
`Transportation and storage conditions:
`Temperature
`Rate change of temperature
`Vibration
`Shock
`Fragility of plastic housing:
`Due to accidental dropping
`Protection against ingress of:
`
`34
`
`1995 PROCEEDINGS Annual RELIABILITY and MAINTAINABILITY Symposium
`
`2
`
`2
`
`
`
`Dust and Liquids
`Mechanical wear/fatigue aspects of:
`Key mat
`Printing
`Cables/connectors
`Life characteristics of field replaceable units; eg. battery
`and antenna.
`Inaddition, different accessories are required to be used
`with a mobile phone. The required accessories can vary
`based on where it
`is used; at home, office, car or a
`combination of these. Accessories include; car holder, car
`cradle, handsfree, cigarette lighter charger, desk holder,
`mains charger, battery, etc;
`It
`is estimated that on
`average there are approximately four accessories used
`with a mobile phone. Even though all the accessories are
`not used at the same time, there is an increasing demand
`to improve and maintain the reliability of accessories at
`high levels. Phones and associated accessories are
`subjected to a series of
`reliability and environmental
`tests. The paper discusses only some reliability activities.
`
`3. RELIABILITY DEMONSTRATION
`
`Thereliability demonstration process includes testing a
`sample (about 50) of phones at a constant high
`temperature with
`power
`cycling
`(power
`on
`and
`transmitting) at a prescribed frequency. Both, the sample
`size and test duration are a function of the productfailure
`rate criteria. Typical test duration is about 6 weeks. The
`test acceleration is a function of both, temperature and
`power on/off and radio transmitting level and time. During
`the test process over thirty parametric measurements are
`taken at
`regular
`intervals. These measures have
`predefined criteria. In the event of a test measure outside
`the criteria being observed, the test unit is subjected to
`failure analysis and corrective action. Also in the event of
`a hard failure, the unit is subjected to failure analysis and
`corrective actions are introduced to eliminate recurrence.
`In analysis of
`failure data, constant
`fallure rate is
`assumed. This assumption could be misleading; earlylife
`failure can be masked by more mature unit test hours.
`The effect of early life failure in the field and the effect of
`modifications could not be easily assessed.
`Based on the observed failure rate trend it was found
`that
`in most applications data is most suitable for
`reliability growth models and in concurrent engineering,
`the reliability demonstration process can be replaced by
`a reliability growth process with a Test Analyze and Fix
`(TAAF) process. A Duane reliability growth plot (ref 4)
`applied to a mobile phone isillustrated in figure 2.
`Apart from applying reliability growth testing, as the
`generic product failure rate decreases the test sample
`and or test
`time required to demonstrate the failure
`increases. Therefore there is a need to look for more
`effective methods which can reduce the test time and
`sample size with the decreasing product development
`
`cycle times. In this area it is necessary to exploit various
`aspects through experiments.
`In general, temperature cyclic tests are favoured
`induce
`potential
`failures more
`than
`constant
`to
`temperature testing. Also it is necessary to research and
`evaluate reliability test methods being applied else where
`and how these could be best utilised.
`
`FIR (%fVr)
`
`
`
`CUM AMB TIME
`
`Figure 2 - Reliability Growth (Duane) plot of a Mobile
`Phone
`
`few years
`Test methods developed over the last
`include; HALT (ref 5), HAST (ref 6), step stress tests (ref
`7) and Strife (ref 8). The first two methods are very
`similar and limited to applications at component level.
`This is primarily due to the capability of applying very
`high
`stresses
`at
`component
`level
`and achieving
`correspondingly high accelerating factors. Thesetests are
`very effective in reducing the test time at component
`level. The step stress and Strife tests, again very similar
`and are applicable at system level. Strife (Stress + life)
`tests originated by Hewlett-Packard in early 1980s.
`Published studies to date are on computers systems and
`reported to be very successful.
`to induce potential
`The objective of Strife tests is
`failures by the application of stresses, preferably multiple
`cyclic stresses. Often tests would commenceat product
`specification extremes. Based on test results, stress
`levels can be extended beyond the specifications,
`provided no untypical failure modes are induced. The
`primary aim is to ‘look’ for potential failures and introduce
`corrective actions based onfailure analysis, to make the
`product robust. It is understood that an improved failure
`precipitation can be achieved with cyclic stresses
`compared to constant level stresses. This is contrary to
`expectations in traditional reliability demonstration tests,
`where no failures are wanted. Another aspect in Strife
`tests is
`that no emphasis is placed on estimating
`reliability measures, but
`to make the product more
`
`1995 PROCEEDINGS Annuai/ RELIABILITY and MAINTAINABILITY Symposium
`
`35
`
`3
`
`
`
`reliable. Also based on published information, it has been
`demonsirated that both the sample size and test duration
`for Strife tests is significantly less than that required for
`traditional
`reliability demonstration tests at constant
`temperature,
`
`4.0 STRIFE TESTING OF PHONES
`
`in order to assess the effectiveness of Strife tests a
`sample of phones were subjected to temperature and
`power cycling,
`transmitting maximum level of
`radio
`frequency during the power on states. The test process
`was:
`
`(1) Initially take (up to 30) parametric measures.
`(2) Temperature cycling between -30 and +60 deg C, two
`hour cycle with one power on/off cycle per hour with
`transmitting. The
`dwell
`time
`at
`temperature was
`approximately half an hour.
`(3) After five days, take parametric measures.
`(4) Subject test sample to random vibration at 2g RMS,
`between the frequency range 20 to 500 Hz for one hour.
`(5) Take parametric measures. Based on results, lift the
`upper temperature by 5 deg C.
`(6) Repeat (2) to (5) above.
`
` 1
`
`5
`
`6
`
`1.4
`1.2
`
`-0.6
`
`2
`
`3
`
`4
`
`Fig 3 - Transmission Frequency Error vs Test Level
`
`63.5
`
`63
`
`62.5
`
`62
`
`61.5
`61
`
`60.5
`
`69
`
`59.5
`
`
` 59
`
`The test is being carried out on 12 phones which
`werepreviously used in the reliability demonstration tests.
`No hard failures were observed, the upper temperature
`was
`increased by 5 deg C eachtime, from the second
`entry in to thermal cycling. A limited number of parametric
`measurementsare illustrated. Box-Whisker plots of four
`parametric measuresareillustrates below. Figures 3 to 6
`illustrates; Transmission frequency error,
`transmission
`mute
`error,
`receiver
`signal
`strength indicator and
`transmission signal to noise ratio, respectively. With the
`limited numberof points (six) no trend can be observed
`with any parametric measure. This indicates that
`the
`applied stresses had no effect on these parameters
`during the considered test period. This can be due to
`reasons such as:
`(a) Robust Parameters, specially as these were subjected
`to reliability tests previously.
`(b) Low severity of Applied stress levels.
`(c)
`Insufficient test duration.
`In order to ascertain which of above reasons is
`applicable, it is planned to continue the test programme.
`Thermal cycling and powercycling with transmitting
`primarily would address electronic componentsandjoints.
`As mechanical parts are not activated, Strife tests would
`
`1.9
`
`1.85
`
`1.8
`
`1.75
`
`17
`
`1.65
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`Fig 5-Receiver Signal Strength Indicator vs Test Level
`76
`
`74
`
`73
`
`1.6 1.55
`72
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`Fig 4 - Transmission Mute Level vs Test Level
`
`Fig 6 - Transmission Signal/Noise Ratio vs Test Level
`
`36
`
`1995 PROCEEDINGS Annual RELIABILITY and MAINTAINABILITY Symposium
`
`4
`
`4
`
`
`
`not address mechanical problems. These include;
`antenna, base connector, keymat, key contact actuators,
`cable, etc.
`It is necessary to carryout separate tests to
`assesstheir reliability.
`
`5.0 ASSOCIATED RELIABILITY
`ASSURANCEACTIVITIES
`
`During various development stages products are
`subjected to different test and evaluation activities. These
`include thermal evaluation using infra red imaging, and
`evaluation of components such as; key mat printing,
`keymat switch activators, base connector, antenna; eic;
`at development stages. Also stress screening is applied
`to precipitate early life failures, primarily during the initial
`product ramp-up periods. A number of these aspects are
`considered below: in particular at
`the early
`
`5.1 Thermal Imaging
`
`imaging is carried out to evaluate
`Infra red thermal
`thermal distribution of circuit boards. The primarily
`objective is to ascertain whether there are hot spots
`during the phone usage. Thermal imaging is carried out
`at all operational modes. Although no hot spots should be
`present if component derating factors have been kept at
`low levels as specified, often hot spots are observed
`during the early product development stages. This can be
`due to unforseen factors such as; timing errors, electrical
`over stress, etc. When the temperature of a component
`exceeds expected acceptable levels,
`the circuit
`is
`analyzed and corrective actions are introduced to
`alleviated the temperature.
`
`in a 25 deg C ambient temperature. The junction
`deg C,
`temperature can be in the order of 115 deg C. At higher
`operating temperatures, say 70 deg C,
`the junction
`temperature would rise to about 150 deg C. Cleanly this
`is not acceptable. Failure analysis revealed that the high
`temperature was caused by a timing error, a software
`modification alleviated the hot spot.
`
`5.2 Component Evaluation
`
`In component evaluation, components are subjected to
`various tests to evaluate reliability and durability and in
`certain instances to evaluate process reiated problems.
`Primarily the components tested are mechanical parts
`which are not tested or exercisedin reliability tests at the
`phone level.
`
`5.3 Stress Screening
`
`Certain phones are subjected to stress screening to
`production and ramp up periods. The primary objective is
`to screen as many phones as possible at
`the early
`production stages and to introduce corrective actions
`through failure analysis before
`the full production
`ramp-up. When satisfied with the early life performance,
`an Ongoing Reliability Test
`(ORT)
`is
`introduced to
`monitor the manufacturing process stability. A typical
`stress screen profile would include temperature cycling
`between -30 and +85 deg C, power cycling with
`transmission during power on periods.
`
`5.4 Commercial issues
`
`There are a number of commercial challenges mobile
`phones are facing in the current competitive market.
`These include; extended warranty, smaller and lighter
`phones,high availability (improvedfield service), etc.It is
`anticipated that technological improvements in progress
`over the last few years will continue to address current
`and future challenges.
`
`6.0 CONCLUSIONS
`
`4A&47é83
`
`NwwhanaSoPONheoymeHee
`
`(1) The Mobile phone is subjected to very harsh
`environmental conditions compared to many commercial
`products.
`eeAenea
`
`
`
`[WS_NHTP2.InG
`94-01-11 16:03359.85
`AGEHA Infrared Systens AB
`
`Figure 7 - An Infrared Thermal Image of a PCB
`
`Figure 7 illustrates a thermal image of a phone circuit
`board.
`lt can be observed that the case temperature of
`the diode is high, the temperature wasin the order of 100
`
`(2) The traditional reliability demonstration process can be
`inefficient and misleading and needsto be replaced by a
`reliability growth tests with TAAF activities.
`
`(3) Strife tests have not degraded test samples. This may
`be due tc a reasons such as; applied stress levels not
`severe enough, test duration is inadequate or phones are
`robust. Particularly so as the test sample was subjected
`to reliability tests previously.
`
`1995 PROCEEDINGS Annual RELIABILITY and MAINTAINABILITY Symposium
`
`37
`
`5
`
`
`
`(4) To assess whether environmental stresses affect
`parametric measurements the Strife test should
`be
`continued.
`
`7.0 REFERENCES
`
`Machine Tools Ltd. Daya is a memberofthe reliability
`committee of
`the Institution of Mechanical Engineers,
`corporate member of
`the Institution of Society of
`Environmental Engineers and the IEEE reliability society.
`
`(1) Pan European Mobile Communications, Summer 90
`(2) MTA-EMCI Communication consultants, 20 Top
`Cellular Countries', Cellular Marketing, June 1993, P58-
`59.
`(3) P Quigley,'Early days for GSM in the UK and Europe’,
`Mobile and Cellular, March 1994, PP15-17.
`(4) U D Perera, Reliability Demonstration of a High
`Reliability Disk Drive’, Proceedings
`of
`the Annual
`Reliability and Maintainability Symposium, 1991, pp30-38.
`(5) R Confer, et.al, 'Use of Highly Accelerated Life Tests
`(HALT) to Determine Reliability of Multilayer Ceramic
`Capacitors’, 41st Electronic Components and Technology
`conference, 1991, pp1320-322.
`(6) N M Troop,
`,et.al,
`'The Comparative Reliability of
`Resister types under HAST’, 6th European Capacitor and
`Resistor Technology Symposium, 1992, pp145-150.
`(7) K E Chesney,
`'Step Stress Analysis of a Printer’,
`Proceedings of the Annual Reliability and Maintainability
`Symposium, 1986, pp26-27.
`(8) A Bailey and R A Gilbert, ‘STRIFE Testing’, Quality,
`November 1992, p53-55.
`
`BIOGRAPHY
`
`, M Sc, B Sc, C Eng, MIEEE
`
`U Daya Perera, Ph D,
`Reliability Specialist
`Nokia Mobile Phones (UK) Ltd
`Ashwood House
`Pembroke Broadway
`Camberley
`Surrey GU15 3SP,
`UNITED KINGDOM
`
`Dr Daya Perera joined Nokia Mobile Phones (UK)
`Limited as Reliability Specialist a year ago,
`is primarily
`involved with reliability aspects
`of products during
`development stages and conducting educational courses
`at different Nokia sites. Prior to joining Nokia, Daya has
`been working
`18 years
`in
`the
`field
`of
`reliability
`engineering. Last eight years as the site reliability
`specialist at IBM (UK) Ltd, Havant. During this period he
`was responsible for reliability assurance and systems
`availability of products and assemblies from development
`stages to manufacturing and field performance. Daya has
`presented ten
`technical
`papers
`at
`national
`and
`international conferences on various
`topics;
`stress
`screening,
`reliability demonstration,
`reliability growth,
`systems reliability and availability.
`Prior to joining IBM, he worked in the field of reliability
`engineering at British Gas Corporation and Alfred Herbert
`
`38
`
`1995 PROCEEDINGS Annual RELIABILITY and MAINTAINABILITY Symposium
`
`6
`
`