ISO 18926:2012

INTERNATIONAL ISO
STANDARD 18926
Second edition
2012-06-01
Imaging materials — Information stored on
magneto-optical (MO) discs — Method for
estimating the life expectancy based on the
effects of temperature and relative humidity
Matériaux pour l’image — Information stockée sur disques opto-
magnétiques (MO) — Méthode d’estimation de l’espérance de vie
basée sur les effets de la température et de l’humidité relative
Reference number
ISO 18926:2012(E)
©
ISO 2012

---------------------- Page: 1 ----------------------
ISO 18926:2012(E)
COPYRIGHT PROTECTED DOCUMENT
© ISO 2012
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized in any form or by any means,
electronic or mechanical, including photocopying and microfilm, without permission in writing from either ISO at the address below or ISO’s
member body in the country of the requester.
ISO copyright office
Case postale 56 • CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 749 09 47
E-mail copyright@iso.org
Web www.iso.org
Published in Switzerland
ii © ISO 2012 – All rights reserved

---------------------- Page: 2 ----------------------
ISO 18926:2012(E)
Contents Page
Foreword .iv
Introduction . v
1 Scope . 1
2 Purpose and assumptions . 1
2.1 Purpose . 1
2.2 Assumptions . 1
3 Normative references . 1
4 Terms and definitions . 2
5 Measurements . 3
5.1 Summary . 3
5.2 Byte error rate (BER) . 3
5.3 Test equipment . 4
5.4 Test specimen . 4
6 Accelerated stress test plan . 4
6.1 General . 4
6.2 Stress conditions . 5
6.3 Accelerated test cell sample population . 7
6.4 Time intervals . 7
7 Data evaluation . 7
7.1 Lognormal distribution model . 7
7.2 Eyring acceleration model . 8
7.3 Acceleration factor . 9
7.4 Survivor analysis . 9
8 Disclaimer .10
Annex A (normative) Ten-step analysis outline . 11
Annex B (informative) Example of a test plan and data analysis .12
Bibliography .21
© ISO 2012 – All rights reserved iii

---------------------- Page: 3 ----------------------
ISO 18926:2012(E)
Foreword
ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies
(ISO member bodies). The work of preparing International Standards is normally carried out through ISO
technical committees. Each member body interested in a subject for which a technical committee has been
established has the right to be represented on that committee. International organizations, governmental and
non-governmental, in liaison with ISO, also take part in the work. ISO collaborates closely with the International
Electrotechnical Commission (IEC) on all matters of electrotechnical standardization.
International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 2.
The main task of technical committees is to prepare International Standards. Draft International Standards
adopted by the technical committees are circulated to the member bodies for voting. Publication as an
International Standard requires approval by at least 75 % of the member bodies casting a vote.
Attention is drawn to the possibility that some of the elements of this document may be the subject of patent
rights. ISO shall not be held responsible for identifying any or all such patent rights.
ISO 18926 was prepared by Technical Committee ISO/TC 42, Photography.
This second edition cancels and replaces the first edition (ISO 18926:2006), of which it constitutes a minor
revision with the following changes:
— the original Annex A has been removed and the remaining annexes have been reidentified;
— in Clause 3, references to ISO/IEC 17346:2005, ISO/IEC 22092:2002 and ISO/IEC 22533:2005 have been
added;
— in 6.2.4, Table 2, the bottom line temperature has been changed from 25 °C to 23 °C;
— in 7.3, Formula (4), the temperature has been changed from 25 °C to 23 °C;
— in Annex B, the temperature in the first sentence of the paragraph above Table B.6 has been changed from
298,1 K to 296,1 K;
— in Annex B, the temperature in the second paragraph below Figure B.5 has been changed from 25 °C to 23 °C.
iv © ISO 2012 – All rights reserved

---------------------- Page: 4 ----------------------
ISO 18926:2012(E)
Introduction
This International Standard is one of a series of standards dealing with the physical properties and stability of
imaging materials.
© ISO 2012 – All rights reserved v

---------------------- Page: 5 ----------------------
INTERNATIONAL STANDARD ISO 18926:2012(E)
Imaging materials — Information stored on magneto-optical
(MO) discs — Method for estimating the life expectancy based
on the effects of temperature and relative humidity
1 Scope
This International Standard specifies a test method for estimating the life expectancy (LE) of information stored
on rewritable and write-once magneto-optical media. Only the effects of temperature and relative humidity on
the media are considered.
2 Purpose and assumptions
2.1 Purpose
The purpose of this International Standard is to establish a methodology for estimating the life expectancy of
information stored on magneto-optical discs. This methodology provides a technically and statistically sound
procedure for obtaining and evaluating accelerated test data.
2.2 Assumptions
The validity of the procedure defined by this International Standard relies on five assumptions:
— the failure mechanisms acting at the usage conditions are the same as those at the accelerated conditions;
— the linearity of the byte error rate (BER) estimated over the accelerated and design conditions is valid;
— all failure mechanisms have been accounted for and appropriately modelled;
— failure caused by reversible effects such as surface dust is not included;
— failure from repairable parts such as external cartridge components is not included.
3 Normative references
The following referenced documents are indispensable for the application of this document. For dated
references, only the edition cited applies. For undated references, the latest edition of the referenced document
(including any amendments) applies.
ISO/IEC 10089:1991, Information technology — 130 mm rewritable optical disk cartridge for information interchange
ISO/IEC 10090:1992, Information technology — 90 mm optical disk cartridges, rewritable and read only, for
data interchange
ISO/IEC 11560:1992, Information technology — Information interchange on 130 mm optical disk cartridges
using the magneto-optical effect, for write once, read multiple functionality
ISO/IEC 13549:1993, Information technology — Data interchange on 130 mm optical disk cartridges —
Capacity: 1,3 gigabytes per cartridge
ISO/IEC 13963:1995, Information technology — Data interchange on 90 mm optical disk cartridges — Capacity:
230 megabytes per cartridge
ISO/IEC 14517:1996, Information technology — 130 mm optical disk cartridges for information interchange —
Capacity: 2,6 Gbytes per cartridge
© ISO 2012 – All rights reserved 1

---------------------- Page: 6 ----------------------
ISO 18926:2012(E)
ISO/IEC 15041:1997, Information technology — Data interchange on 90 mm optical disk cartridges — Capacity:
640 Mbytes per cartridge
ISO/IEC 15286:1999, Information technology — 130 mm optical disk cartridges for information interchange —
Capacity: 5,2 Gbytes per cartridge
ISO/IEC 17346:2005, Information technology — Data interchange on 90 mm optical disk cartridges — Capacity:
1,3 Gbytes per cartridge
ISO/IEC 22092:2002, Information technology — Data interchange on 130 mm magneto-optical disk cartridges —
Capacity: 9,1 Gbytes per cartridge
ISO/IEC 22533:2005, Information technology — Data interchange on 90 mm optical disk cartridges — Capacity:
2,3 Gbytes per cartridge
Aitchison, J. and Brown, J.A.C., The Lognormal Distribution, Cambridge University Press, 1957
4 Terms and definitions
For the purposes of this document, the following terms and definitions apply.
4.1
baseline
condition representing the disc at time of manufacture
NOTE This is customarily the initial parameter measurement taken prior to any application of stress. The designation
is usually t = 0 for a stress time equal to zero hours.
4.2
byte error rate
BER
number of bytes in error divided by number of bytes tested
NOTE BER refers to the raw byte error rate, without benefit of any error correction or sector re-allocation.
4.3
censored data
time at which a specimen is removed from life testing due to any reason other than having reached end-of-life
4.4
end-of-life
occurrence of any loss of information
4.5
information
signal or image recorded using the system
4.6
F(t)
probability that a random unit drawn from the population fails by the time t, or the fraction of all units in the
population which fail by time t
4.7
life expectancy
LE
length of time that information is predicted to be retrievable in a system under extended-term storage conditions
4.7.1
standardized life expectancy
SLE
minimum life span, predicted with 95 % confidence, of 95 % of the product stored at a temperature not exceeding
23 °C and a relative humidity (RH) not exceeding 50 %
2 © ISO 2012 – All rights reserved

---------------------- Page: 7 ----------------------
ISO 18926:2012(E)
4.8
magneto-optical disc
any disc conforming to the ISO/IEC standards contained in Clause 3
NOTE Double-sided media are considered to be composed of two discs, one per side. In general, a magneto-optical
disc is one that uses thermo-magnetic properties for recording and opto-magnetic properties for reading.
4.9
R(t)
probability that a unit drawn from the population will survive at least time t, or the fraction of units in the
population that will survive at least time t
NOTE R(t) = 1 − F(t)
4.10
retrievability
ability to access information as recorded
4.11
stress
experimental variable to which the specimen is exposed for the duration of the test interval
NOTE In this International Standard, the stress variables are confined to temperature and relative humidity.
4.12
system
combination of recording medium, hardware, software and documentation necessary to retrieve information
4.13
test cell
device that controls the stress to which the specimen is exposed
4.14
test pattern
distribution of 1’s and 0’s within a sector
5 Measurements
5.1 Summary
A sampling of 80 discs is baseline tested for the BER, then divided into five groups according to a specified
plan. Each group of discs is subjected to one of five combinations of temperature and relative humidity (stress).
During the exposure to the stress condition, discs are periodically removed from the environmental test cell
according to a set plan. These discs are then retested for BER and subsequently returned to the test cell for
additional increments of exposure at the same stress.
−4
For each disc, the time to reach end-of-life (loss of any information or BER 5 × 10 ), is then determined or estimated.
For each stress condition, the resulting service life data are fitted to a lognormal distribution for that stress. These
five sets of parameters (lifetime, temperature and relative humidity) are regressed to fit an Eyring acceleration
model. This model is then used to estimate the distribution of lifetimes at a standardized set of conditions.
5.2 Byte error rate (BER)
The objective of measuring the BER is to establish a practical estimation of the system’s ability to read previously
written bits using a standard drive. This International Standard considers BER to be a reasonable estimate of
the performance of the system. A change in the BER in response to the time at the accelerated temperature
and humidity is the principal degradation parameter.
© ISO 2012 – All rights reserved 3

---------------------- Page: 8 ----------------------
ISO 18926:2012(E)
The true end-of-life for any data storage media is any loss of information. Ideally, each specimen is tested
until actual failures occur. The first occurrence of any disc degradation that results in uncorrectable errors is
considered to signal the actual end-of-life.
Realistically, testing until all discs have failed is impractical. For the purposes of this International Standard,
−4
the maximum average BER shall be 5,0 × 10 if actual failures do not occur during testing. This is very
system dependent and its use here is an arbitrary level chosen as a conservative prediction of the onset of
unacceptable errors and thereby the end of disc life. All BER measurements are made with the system error
correction switched off.
5.3 Test equipment
5.3.1 General
Any disc drive system that conforms to ISO/IEC standards (see Clause 3) may be used. The tester shall be
capable of reporting errors occurring prior to the implementation of error correction systems.
5.3.2 Calibration and repeatability
A control disc shall be maintained and measured before and after each data collection interval. For each test
drive, a control chart shall be maintained for this control disc with plus or minus three sigma action limits. The
mean and standard deviation of the control disc shall be established by collecting at least five measurements.
If any individual BER reading lies outside the action limits, the problem shall be corrected and all data collected
since the last valid control point shall be remeasured.
If it becomes necessary to replace the test drive, the new drive shall be calibrated using the control disc and
compared to the replaced drive. If a statistical difference exists between the control disc BER means, subtract
the new disc mean from the old disc mean and add this correction factor to all subsequent BER measurements
made with the new drive.
5.4 Test specimen
A test specimen is any disc that conforms to ISO/IEC specifications referenced in Clause 3 and contains
representative data written over 100 % of the user area. Representative data may be real data or random test data.
6 Accelerated stress test plan
6.1 General
A well manufactured magneto-optical disc should last several years or even decades. As such, it is not practical
to conduct life studies under normal usage conditions. It is then necessary to conduct accelerated aging
studies in order to determine the estimated potential for life of this medium. To be successful, these studies
shall be planned ahead of time in order to be of sound design both technically and statistically.
Many accelerated life test plans follow a rather traditional approach in sampling, experimentation and data
evaluation. These “traditional plans” share the following characteristics:
— the total number of specimens is evenly divided amongst all of the accelerated test cells;
— the specimen from each test cell is evaluated at the same increment of time;
— the Arrhenius relationship is used as the acceleration model;
— the Least Squares method is used for all regressions;
— the calculated life expectancy is for the mean or median life rather than for the first few failure percentiles.
4 © ISO 2012 – All rights reserved

---------------------- Page: 9 ----------------------
ISO 18926:2012(E)
Statisticians, on the other hand, have devoted considerable attention to developing “optimum test plans” for an
ideal situation. These plans have the following characteristics:
— two and only two acceleration levels for each stress;
— a large number of specimens distributed mostly amongst the lowest stress levels;
— the need to know the failure distribution, a priori, in order to develop the plan.
The maximum effectiveness of a plan can either be estimated before the test starts or determined after the
results have been obtained. As each MO system will have different characteristics, a specific detailed optimum
plan is impossible to forecast.
This test plan borrows from the optimum plan, the traditional plan, previous experience with the systems, test
equipment and accelerated test stresses to put together a “compromise test plan”. Modification of this plan is required
to design the best plan for other applications. The methodology shall be applicable to all MO media assessments.
6.2 Stress conditions
6.2.1 General
As mentioned in 6.1, an optimum test plan utilizes only two stress levels for each parameter evaluated, since in
an ideal case the relationship between changes in the parameter investigated and changes in stress are known.
The compromise test plan documented in this International Standard does not make such an assumption;
therefore, three different stress levels per parameter shall be used so that the linearity of the parameter function
versus the stress level may be demonstrated.
The test plan shall have the majority of test specimens placed at the lowest stress condition. This minimizes
the estimation error at this condition and results in the best estimate of the degradation rate at a level close to
the usage condition. The greater number of specimens at the lower stress also tends to equalize the number
of failures observed by test completion.
For implementing the test plan documented in this International Standard, five stress conditions shall be used.
The minimum distribution of specimens among the stress points that shall be used is shown in Table 1.
Table 1 — Summary of stress conditions
Test cell Test stress Number of Interval duration Minimum total time
number T /RH specimens h h
inc inc
1 80 °C/85 % RH 10 500 2 000
2 80 °C/70 % RH 10 500 2 000
3 80 °C/55 % RH 15 500 2 000
4 70 °C/85 % RH 15 750 3 000
5 60 °C/85 % RH 30 1 000 4 000
6.2.2 Temperature (T)
The temperature levels chosen for this test plan are based on the following.
— There shall be no change of phase within the test system over the test temperature range. This would
restrict the temperature to greater than 0 °C and less than 100 °C.
— The level of temperature shall not be so high that either plastic deformation or excessive softening of
thermoset adhesives occurs.
A common substrate material for magneto-optical discs is polycarbonate (glass transition temperature
approximately 150 °C). Experience with high temperature testing of MO discs indicates that an upper limit of
80 °C is practical for most applications.
© ISO 2012 – All rights reserved 5

---------------------- Page: 10 ----------------------
ISO 18926:2012(E)
6.2.3 Relative humidity (RH)
Practical experience shows that 85 % RH is the upper limit within most accelerated test cells. This is due to
the tendency for condensation to occur on cool sections of the chamber (such as observation windows, cable
ports, wiper handles, etc.). Droplets may become dislodged and entrained in the circulating air within the
chamber. If these droplets fall on the test specimen, false error signals could be produced.
6.2.4 Rate of stress change
The process, described in this International Standard, requires that temperature and relative humidity be
gradually changed (ramped) from permitted testing conditions to accelerated stress conditions and back again
a number of times during the course of testing. The ramp duration and conditions shall be chosen to allow
sufficient equilibration of absorbed substrate moisture.
Large departures from equilibrium conditions may result in the formation of liquid water droplets inside the
substrate or at its interface with the thin film layers. Gradients in the water concentration through the thickness
of the substrate shall also be limited. These gradients drive expansion gradients which can cause significant
disc deflection.
In order to minimize moisture concentration gradients, the ramp profile specified in Table 2 shall be used. The
objects of the profile are:
— to avoid any situation that may cause moisture condensation within the substrate;
— to minimize the time during which substantial moisture gradients exist in the substrate;
— to stay within specified rates of temperature and humidity change;
— to produce, at the end of the specified profile, a disc which is sufficiently equilibrated to proceed directly to
testing without delay.
Discs bonded with thermoplastic adhesives may be close to, or above, their softening temperatures. By
including a 2 h step at 50 °C/85 % RH, these adhesives have an opportunity to set before continuing the ramp
to ambient conditions.
Table 2 — Temperature and relative humidity transition (ramp) profile
Temperature Relative humidity Duration
Process step
°C % RH h
Start at T at RH —
amb amb
T, RH ramp to T to RH 0,1/°C
inc inc
Incubation at T at RH See Table 1
inc inc
T, RH ramp 50 85 0,1/°C
Adhesive set 50 85 2
RH ramp 50 35 5
T, RH ramp 23 50 2,5
6.2.5 Independent verification of chamber conditions
A system independent of the chamber control system shall be used to monitor temperature and humidity
conditions in the test chamber during the stress test.
6.2.6 Specimen placement
Fully assembled specimens (includes cartridge and shutter) shall be placed uncovered, either vertically or
horizontally, within the test chamber. Discs shall be aligned so that their surface is parallel to the chamber
6 © ISO 2012 – All rights reserved

---------------------- Page: 11 ----------------------
ISO 18926:2012(E)
airflow. A space of at least 5 mm shall be maintained between cartridges. Cartridged discs shall be stressed
with the shutter closed.
6.3 Accelerated test cell sample population
In order to estimate the log mean and log standard deviation of a lognormal distribution, (see Aitchison and
Brown in Clause 3) at least ten failures shall be observed. Observing at least ten failures may not be a problem
for a realistic test time at 80 °C/85 % RH but becomes more difficult at milder stress temperature and relative
humidity combinations. Assigning a larger percentage of the specimens to the milder stresses increases the
chance of observing the necessary number of failures within a practical time interval.
Specimens that have not failed at the end of the test duration shall be time censored. This is also known as
Type I censoring (see Reference [2], page 233).
If ten failures are not observed by the end of the test duration, then failures may be estimated. To compute
the estimated failure time for each disc, it is necessary to first determine a transformation of the BER, such as
ln(BER), that results in a linear time dependence. Standard linear regression techniques shall be used to find
the best fit to the transformed data. The failure time for each disc shall then be computed by interpolation or
extrapolation using each disc’s regression equation.
6.4 Time intervals
6.4.1 General
For a test plan where the “exact time-to-failure” is to be the result of extrapolated rate data, no fewer than five
time intervals for data collection are required. The baseline measurement (at t = 0) is one of these data points.
Within a stress condition, the intervals shall be constant.
As the stress conditions get milder, the intervals become longer. Longer time intervals provide the opportunity
for more failures to occur at the milder stress conditions.
6.4.2 Test plan
Table 1 specifies the temperatures, relative humidities, time intervals, minimum total time and specimen
distributions for each stress condition. A separate group of specimens is used for each stress condition. This
constitutes a “constant stress” test plan.
All temperatures have a permitted range of ± 2 °C; all relative humidities have a permitted range of ± 3 % RH.
The stress conditions tabulated in Table 1 offer sufficient combinations of temperature and relative humidity to
satisfy the mathematical requirements of the Eyring model (see 7.2), to demonstrate linearity of BER versus
time, and to produce a satisfactory confidence level to make meaningful conclusions.
6.4.3 Measurement conditions
Discs shall be equilibrated to the environment in which they will be tested. Foreign surface contaminants shall
be cleaned from the disc prior to testing.
7 Data evaluation
7.1 Lognormal distribution model
7.1.1 General
The lognormal distribution model shall be used for characterizing the failure rate distribution. The lognormal
distribution model has been found to be very flexible and to fit many applications in the corrosion of thin metal
films. It is likely to be the best distribution model for cases in which the dominant failure mechanism relies on
© ISO 2012 – All rights reserved 7

---------------------- Page: 12 ----------------------
ISO 18926:2012(E)
chemical reactions or diffusion. Experience has shown that the life distribution of MO discs may be modelled
by the lognormal distribution (see, for example, Reference [3]). The lognormal equation is:
1  log(x)−μ 
el
t

 
1 1
2 σ
2
 l 
F(t) =− e dx (1)
∫
σ x
2π
l
o
where
t
is the time;
σ is the log standard deviation;
l
x
is a variable representing specimen failure time;
µ is the log mean;
l
loge(x) is the natural logarithm of x.
7.1.2 Model validity
The accuracy of life estimates and confidence limits depend on how well a model fulfils a few basic assumptions.
One important assumption for the lognormal model is that the log standard deviation has the same value at all
stress levels. It is essential to verify this assumption.
One test method that is available with almost all life time data analysis computer packages is a comparison
of log standard deviation confidence limits. If the confidence interval for the log standard deviation at each
accelerated stress level overlaps the confidence interval at the usage stress levels, statistically the parameters
are not significantly different.
If a statistically significant difference exists among the stress level log standard deviation parameters, examine
the estimates and confidence limits for each scale parameter and determine how they differ. It may be
appropriate to edit data due to different failure modes, testing error or simple human error.
A listing of computer packages, along with their key features, which can be useful for life expectancy data
[2]
analysis is given by Nelson on pages 237 to 239. Equivalent software may be used.
7.2 Eyring acceleration model
The Eyring model has found broad application and shall be the model for estimating the life expectancies of
MO discs.
8 © ISO 2012 – All rights re
...