Test activities arise all along component life cycle in different ways:
testability analysis during design stage, then for both production
and maintenance stages, design and generation of test data sets,
application of these test sets with ATE (
Automatic Test Equipment) and last
interpretation of the test results obtained.
The design stage is well dominated for both digital, analog and
mixed-signal boards, thanks to mature CAD tools available in industry.
But, except in the digital case, testing activities in maintenance
stage are very difficult and costly to achieve. We focus on such kinds
of mixed-signal or analog boards testing problems during maintenance
phases. Specificities of the maintenance stage lead us to use a
functional-based approach for the board test data generation, coupled
with a functional modeling of the DUT (
Device
Under Test). We have defined a uniform framework, not only
allowing both analog and digital specifications, but also enabling the
modeling of both parts interactions and the test plan to be achieved.
This modeling work of both the whole hardware and the testing process
is the first step for the automation of the test procedure for
complete mixed-signal boards.
Our work results from an industrial partnership with
ISIS-MPP society.
We are currently developping a tool for manipulating these models and
applying them to industrial test cases, in order to validate our
approach.
Results have been published in international workshops adressing the
test domain. A PhD work is also in progress. We are currently working
on timing aspects and testing strategies.