Week 14

Reliability Testing

06/12/2022 1 www.iuli.ac.id
 Reliability Testing

▪ Reliability and quality

▪ Design reliability and quality
reliability
▪ Examination following quality
cycling tests
▪ Reliability test economics

www.iuli.ac.id
 Reliability
The probability that a product, system, or service will
perform its intended function adequately for a
specified period of time, or will operate in a defined
environment without failure.

• Probability: the likelihood of mission success

• Intended function: to light, cut, rotate, or heat
• Satisfactory: perform according to a specification, with
an acceptable degree of compliance
• Specific period of time: minutes, days, months, or
number of cycles
• Specified conditions: temperature, speed, or pressure

www.iuli.ac.id
Components of reliability

Examples of product
reliability statements
or guarantees:

"This car is under

warranty for 40,000
miles or 3 years,
whichever comes
first.“

"This mower has a

lifetime guarantee."

www.iuli.ac.id
 Quality Vs. Reliability
Quality shows how well an object performs its
proper function

Reliability shows how well the object maintains

its original level of quality over time

Example
A car that is safe, fuel efficient, and easy to operate
may be considered high quality. If the car continues
to meet the criteria (safe, fuel efficient and easy to
operate) for several years, it may be considered
reliable.

www.iuli.ac.id
 Design reliability and
quality reliability
▪ Design reliability - verify a satisfactory
standard (by quality activities) originated by
design.

▪ Quality reliability - the fidelity of batch

production and manufacture such that the
product quality consistently attains the design
specification and conformance requirements.

www.iuli.ac.id
 Design reliability and
quality reliability
▪ The reliability of manufacturing quality is assured
by the quality activities. If all the controls are
operated, the product quality and reliability
should be attained.

▪ Complete quality coverage must include regular

reliability test monitoring, which aren’t proving
original designs and not testing for life of the
product, or for long-term failure.

www.iuli.ac.id
 Infant mortality failures
▪ Based on statistics and experience, faults in
reliability of manufacture usually show up very
early in the life of a product.

▪ This situation is depicted by the 'bath-tub' curve

shown in Fig. 8.1, which shows a product with a
designed life of five years.

▪ They occur after a very short period of use and

subsequent failures occur at much larger
intervals until the product settles down to a
minimal failure rate for most of its designed life.
www.iuli.ac.id
Bathtub curve

www.iuli.ac.id
 Infant mortality failures
▪ Assuming an accurately estimated design life,
failures due to wearing out, lack of maintenance,
etc., start to increase towards the end of the
expected life period and can mount rapidly
again, as shown in Fig. 8.1.

▪ Usually designs have a useful margin and the

onset of wear-out failures can be much later
than design nominal life. Products often attain
twice or three times the planned life figure.

www.iuli.ac.id
 Infant mortality failures
Fig. 8.1 is very arbitrary for the following reasons:

▪ A poor design will not settle down as shown.

▪ Faults may develop on a large scale after a
period caused by factors: poor manufacture,
material fatigue, lack of good maintenance, etc.
▪ Type of use may change.
▪ Grade of user may deteriorate.
▪ Consumables used with the product may
introduce trouble.

www.iuli.ac.id
 Quality monitoring tests
▪ Infant mortality faults can usually be revealed with
relatively limited amounts of cycling testing, termed
product quality reliability monitoring tests or simple
quality tests.
▪ Infant mortality faults will usually show up as early
as 200 cycles of a component which is required to
complete 5 million in normal use.
▪ Failures will occur later than 200 cycles but at
rapidly diminishing frequency. Quality monitoring
tests are adequate if performed for 1,000 cycles.
Experience shows that a lesser figure, of say 500
cycles, would be sufficient for particular products.
www.iuli.ac.id
Aspects of quality monitoring tests
▪ They should simulate exactly the component
application intended.
▪ The attitude of the component during test must be
exactly as its attitude in normal use, e.g. horizontal,
vertical, inclined.
▪ The test should include vibratory effects. Electrically
energized 'limpet' type vibrators are marketed which
are very convenient for vibration tests.
▪ The component should be loaded with the torque,
resistance or electrical load normally carried.

www.iuli.ac.id
13
Aspects of quality monitoring tests
▪ Such load should be switched on and off for the
number of cycles specified, and again simulating
normal application but emphatically not using an
exactly equal on/off pulse.
▪ The latter is most easy to arrange on a test rig
(bench), but equal pulses rarely occur normally.
▪ Equal pulses are far less likely to reveal faults than
unequal pulses, e.g. components build up
temperature, magnetic effects, etc., from longer on
than off pulses.

www.iuli.ac.id
Aspects of quality monitoring tests
▪ The rig should have self-indicating fault displays and
should not need to be constantly watched.
▪ Multiple station rigs are worthwhile in enabling
coverage of several components simultaneously.
▪ Soak-type tests whereby components are left under
constant load for long periods are frequently used in
the electrical component industry. Much shorter-
period cycling tests are far more preferable.
▪ Tests should, where required, be applied at other
than ambient temperature and humidity conditions.

www.iuli.ac.id
Electrical test bench

www.iuli.ac.id
Electrical test bench

www.iuli.ac.id
Diesel test bench

www.iuli.ac.id
Vehicle powertrain test rig

www.iuli.ac.id
 Examination after the tests
Quality reliability monitoring tests should be followed
by strip and examination of the components
concerned, for two reasons:
1. The quality and cleanliness of the unseen internal
areas and parts can be verified.
2. Signs of undue stress, heat, wear, etc. may be
observed.
Because the infant mortality tests are carried out over
a small number of cycles, the components concerned
are still satisfactory for normal use.

www.iuli.ac.id
 Examination after the tests
The effects of infant mortality failures can be
catastrophic but the precautions to obviate them
can be easily established.
Once determined, they usually have assignable
causes that can be quickly put right before
product distribution.
The remedial action cost of defective products
reaching customers can be enormous and
company’s public image can be seriously
impaired.

www.iuli.ac.id
 Examination after the tests
Design reliability and life tests carried out on similar
rigs require close cooperation between the
departments for the use of test rigs and facilities.
Design engineering coordinates the assurance of
continued product adherence to design specs.
Quality monitoring should be done successively on
samples from each week of a continuous production
run, or even daily, if practicable.
Such procedure will provide early warning of faults
and possible corrective action to avoid production
of similar faulty product.

www.iuli.ac.id
 Tests by vendors

▪ Supplies from bought-out sources should have such

reliability monitoring tests carried out by the vendor
concerned and certified with delivery paperwork.
▪ Requirements for such supplies must be stated on the
contract and verification of such procedures must be
part of vendor appraisal and quality liaison visits.
▪ Both design and quality reliability tests have the effect
of continuously improving designs, including proprietary
products used on particular applications.

www.iuli.ac.id
 Reliability test economics
▪ Product design improvement.
▪ Performance improvements.
▪ Infant mortality failure reduction.
▪ Customer appreciation, fidelity, public image.
▪ Reduction in service costs and travelling time.
▪ Minimising of warranty and guarantee claims.
▪ Fewer spares requirements and holdings.
▪ Less product 'down-time' at the customer.
▪ Sales and marketing advantages.

www.iuli.ac.id
 Personnel ratios

▪ As stated earlier, it is logical to ensure that bought-out

supplies are adequately quality assurance tested
before delivery.
▪ This factor and optimum liaison with design and
engineering in connection with their reliability testing
activities, should mean that normally only 1 to 2 per
cent of quality personnel are engaged full time on
quality assurance reliability monitoring test activities.

www.iuli.ac.id
 Today’s videos
▪ Should you Buy a New or Used Car? - The Bathtub Curve
https://www.youtube.com/watch?v=Q5KmSRWp-8I

▪ Nord-Lock Wedge-Locking Washers - Junker Vibration Test

https://www.youtube.com/watch?v=IKwWu2w1gGk

▪ Fully automated end-of-line test bench for BMW eDrive

https://www.youtube.com/watch?v=Kh4kbFyNI6U

www.iuli.ac.id