Test Coverage Analysis by Fault Class

Test coverage is usually reported as a percentage, and the percentage on its own says very little. What it measures and what it excludes decide whether the figure describes the risk.

What Coverage Measures

The common measure is the proportion of nets or of pins that can be exercised by the test. It counts access rather than detection.

A net that is probed is not the same as a fault that is detected, because a fault that does not change a measurable value at the accessible node is not found by that node. Our test notes describe how the access is provided.

Structural and Functional Coverage

Structural test checks that the connections exist and that the values are correct. Functional test checks that the assembly does what it is supposed to do.

A structural test with high coverage can miss a wrong component that measures the same, and a functional test can miss a marginal joint that passes at the test temperature. The two are complementary. Our test coupon notes describe the process evidence that supports them.

Coverage by Fault Class

The useful analysis lists the fault classes and states which of them the test detects: open, short, wrong value, reversed orientation, missing component, out of tolerance.

A single percentage hides the class that is not covered, which is usually the one that reaches the customer. Our yield notes describe the defect classes that matter.

Nodes That Cannot Be Reached

Some nodes cannot be probed because a component covers them, because the pad is too small, or because the net is sensitive to the capacitance the probe adds.

Those nodes should be identified and the detection moved elsewhere, or the risk accepted explicitly. Our test point notes describe how the access is designed in.

Test Limits and Guard Bands

A test detects an out of tolerance part only if the limit is tighter than the spread of the measurement. A limit set at the datasheet value detects only gross faults.

The limit should account for the measurement system, the temperature and the unit to unit spread. Our derating notes describe the margins involved.

Coverage That Is Not Needed

Coverage costs time and it can cost access. A net whose failure would be detected immediately by a functional test does not also need a structural test.

The analysis should compare the cost of the additional coverage with the cost of the escape it prevents. Our process control notes describe how the trade is made.

Verification

The coverage claim should be verified by injecting faults on a sample assembly and confirming that the test detects them. A claim that has not been proven by injection is a claim about access.

The injection should be performed on the classes that matter rather than on the easiest ones. Our failure analysis notes describe how the faults are produced and recorded.

Process Control and Verification

Where a value sits close to a process limit, the drawing should say so, since the shop can then open the process window rather than working to a nominal figure that carries no tolerance. A stack-up that is drawn rather than described removes most of the ambiguity from a quotation, and it lets the fabricator price the board against the dielectric and copper weights that will actually be used.

Where the requirement is not written down, the shop supplies its own default, and the default is chosen for the process rather than for the design. The measurements that matter are the repeatable ones: conductor width and spacing, annular ring, finished hole size, plating thickness and surface finish are all verifiable on a coupon that travels with the panel.

Running a first article through the same checks as the production panel confirms that the two agree, and that comparison is the cheapest form of process control available at prototype stage. Keeping a sample from the panel turns a dispute into a measurement, because the same coupon can be re-examined by both parties without rebuilding the batch.

Reviewing the design before the data is released is cheaper than correcting it after the panel is in the tank, because every step downstream inherits the decision made at the front end.

Checks Before Release

The checks that matter are the ones performed on the product rather than on a sample kept for the purpose, because a coupon that travels with the panel is the only evidence about that panel. Where a requirement can be measured, it should be measured at the point of manufacture and recorded against the board or the lot it applies to.

A parameter that is set once and never re verified drifts, and the drift is usually discovered by a defect rather than by the record. The tooling, the material and the profile form one system, and a change to any of them should be assessed against the other two before it is released.

Where the process window is narrow, the measurement resolution has to be better than the window, or the data cannot distinguish a good part from a marginal one. A record that identifies the operator, the date and the settings is worth more than a record that identifies only the result.

The acceptance criteria should be written before the work starts, so that the decision is made by the specification rather than by the person inspecting.

Verification and Records

Where an operation cannot be verified afterwards, it has to be controlled during the operation, and that control has to be visible in the record. Sampling is a compromise between cost and confidence, and the sample size should follow from the failure rate that has to be detected.

The first article confirms that the setup matches the intent, and it is the cheapest point at which a wrong setup can still be corrected. Handling between operations is part of the process, and the damage it causes is often attributed to the operation that preceded it.

FAQ

Is ninety nine percent coverage good? It depends on which one percent is not covered and on what happens when that net fails. The class matters more than the number.

Can functional test replace structural test? It cannot detect everything structural test does, and structural test cannot detect everything functional test does.

What does gopcb provide for coverage analysis? We provide coverage stated by fault class rather than as one figure, identification of unreachable nodes with a disposition, limits set with a guard band against the measurement spread, fault injection on a sample to verify the claim, and a cost comparison for coverage that is not needed.