|
Background
A reliability group in a semiconductor company is planning an
accelerated test for an electronic device. 30 test units will be employed
for the test and the test is planned to last for 600 hours. Temperature
and voltage have been determined to be the main factors affecting the
reliability of the device. The normal use conditions of the devices are
300K for temperature and 4V for voltage. The purpose of the experiment is
to estimate the B10 life of the device. The reliability engineer wants to
use a three-level optimum plan because it will be easier to manage than a
five-level test plan.
Existing Information
Based on engineering knowledge, data from a previous design
and data from a pilot test, the following information is available:
- An Arrhenius model is assumed for the life-stress relationship
associated with temperature and is accurate up to 360K.
- A power model is assumed for the life-stress relationship associated
with voltage and is accurate up to 10V.
- The beta parameter for the underlying Weibull distribution is
estimated to be 3.
- The probabilities of failure for the product at the end of the test
are estimated as follows:
- Stress 1 = Usage Stress, Stress 2 = Usage Stress: 0.02
- Stress 1 = Highest Stress, Stress 2 = Usage Stress: 0.9
- Stress 1 = Usage Stress, Stress 2 = Highest Stress: 0.6
Analysis
Step 1: Using ALTA 7 or ALTA 7 PRO, the analyst adds a
Test Plan to a project. The known parameters of the test are entered in
the Setup Window as shown in Figure 1.
Step 2: The analyst generates the Test Plan, as shown in
Figure 2.
The Test Plan can be evaluated by
solving for any one of three criteria (confidence level, bounds ratio
or sample size) given the two other criteria. The bounds ratio is the
ratio of the upper confidence bound to the lower confidence bound on
Tp (the estimate of the time when the specified percentage of the
units will have failed at use stress).
In Figure 3, a confidence level of 0.9
and a sample size of 30 units are specified. The bounds ratio is
calculated to be 1.92.
In Figure 4, a bounds ratio of 1.5 and a
confidence level of 0.9 are specified. The sample size is calculated to be
78 units.
In Figure 5, a sample size of 30 units
and a bounds ratio of 1.5 are specified. The confidence level is
calculated to be 69.34%.
|
ALTA
7 Case Studies
Version
7
