1.

The Rule of Ten

A useful indicator that is quoted in order to demonstrate the importance of test and
discovering faults in an electronic circuit once it has been fabricated and before it
is used, is referred to as the “Rule of Ten” see Fig. 1.2. In this, the cost
multiplies by ten every time a faulty item is not detected but is used to form a large
electronic circuit or system. Whilst this is a generalisation, it identifies the cost
escalation of failing to detect faults whenever an electronic circuit/system is
manufactured/fabricated. Here, if the cost of detecting a faulty device (IC) when it
is produced is one unit, the cost to detect that faulty device when used at the board
level (PCB) is ten times that. The cost to detect that faulty board when inserted in
its system is ten times the cost of detecting the faulty board at production, and so
on. Hence, there is a cost escalation that must be avoided.

2. The Evolution of Test Engineering

Over the last few years, the evolution of test engineering has brought it closer to
design, bridging the “traditional” gap between design and test. Here, where once
the design and test activities were separate and distant, the gap has been bridged by
developing and adopting the “Design for Testability” (DfT) – or sometimes
referred to as “Design for Test” - approach, see Fig. 1.3. Here, test is now closer to
design and may influence specific details on how a design is created in order to
4 Integrated Circuit Test Engineering: Modern Techniques
identify testability issues and improve test access to specific circuitry within the
design. In this scenario, specialist engineers in design and test are supported with a
generalist DfT engineer - with the ability to bridge the gap, but not necessarily
required to be a specialist in either field – the need for specialists is based on the
need for in-depth knowledge of specific design and test issues, roles which a single
person could not realistically be expected to undertake. The exact structure of any
DfT activities would ultimately be specific to the particular company.
3. Test Engineer Activities
It could be argued that, given a perfect design and a perfect fabrication process,
there is no need for test. This would be true if the world was perfect, but given that
even the most minor defect in either the design and/or fabrication process can
cause a circuit fault, the importance of test cannot be underestimated.
In order for a design to be tested, there is a requirement to ensure that the design is
both controllable and observable:
Controllability: - The ability to control specific parts of a design in order to set particular values at specific
points. In a digital logic design, this would be a particular logic value. In an analogue circuit, this would be a
particular voltage or current level.
Controllability: - particular logic value. In an analogue circuit, this would be a particular voltage or current
level.
The need for access to parts of the design by addressing the controllability and
observability is fundamental to test, but it must be addressed in the most cost effective means possible.
Adequate circuit test coverage must be provided at the
lowest cost possible. This is a non-trivial problem to solve. For example, the
number of nodes within a circuit is increasing with increases in design complexity.
This is causing additional problems for test in the ability to access these nodes
within the circuit that cannot have signals directly applied to them. A traditional
approach to test is that once a design has been completed, it is given over to the
test engineer to identify and implement a test procedure (the procedure will be
implemented as a software program that will control the target tester hardware that
then physically interface to the device under test (DUT)). At this stage, any
testability issues relating to the ability to develop an adequate and minimal cost test
procedure, are only identified. However, at this stage, if it is found that the design
cannot be adequately tested, it may be too late to rectify the problem and a greater
potential for faulty devices to be passed onto the customer exists.
engineer alongside the design engineer in order to assess test and design testability
issues at an early stage in the design process – starting at the design specification
stage and prior to detailed design activities. Potential problems can be identified
 and, if possible, removed or their effects minimised. In this, the test engineer
would be involved with design at a Test Development stage and fabrication
(process) at the Production Test stage. A coherent test strategy would be
developed and implemented, see Fig. 1.4. In this:
 The Test Development stage occurs at an early stage in the product
development process and requires a basis in design. At this stage, the test
strategy is developed, providing a “high-level” description for the needs
to provide test access and the addition of any additional test circuitry.
 The Production Test stage implements the test during high volume
production of the product. This is to be as cost-effective as possible in
order to aid the minimisation of the overall device cost.

The steps between “Test Development” and “Production Test” stages involve the
translation of high-level strategies into the hardware and software infrastructure
that is required to undertake design test in high volume production. The test
program development, in a suitable programming language (commonly based on
C), targets the required tester hardware system. In production test, a suitable
Automatic Test Equipment (ATE) system would be identified and resourced.
Such a system would be required to test the device to the required quality levels
but within the shortest time as possible.