A source-measure unit can report a plausible numeric reading while failing to establish the programmed stimulus because it reached compliance. If the test records only the measured value, an open connection or unexpected load may be mistaken for the intended operating condition. Record source state with the reading.
Tektronix’s compliance explanation defines compliance voltage as the voltage limit a current source can reach while attempting to produce its programmed current. Consult the actual instrument documentation for status reporting and for the complementary behavior in voltage-source mode.
Define the intended stimulus
Record source mode, programmed level, compliance setting, measurement range, and connection arrangement. State which quantity the test intends to control and which response it intends to measure. This prevents a current-source test from being interpreted as though the programmed voltage were the actual stimulus.
Read back the actual source and measurement information supported by the instrument. Preserve status bits or flags associated with the same acquisition. A compliance indicator read much later may describe another step rather than the reading under review. Verify whether the reported status is instantaneous, latched, or associated with a stored reading.
Investigate compliance events
| Possible condition | Discriminating check |
|---|---|
| Open or poor connection | Verify the intended force and sense path |
| Unexpected product load | Compare board state and measured response |
| Incorrect limit | Compare configuration with the approved method |
| Transient event | Review timing and time-resolved behavior where needed |
Do not automatically increase compliance to obtain a passing result. The limit may be an intentional part of the product test. First determine why the intended operating condition was not established and whether the proposed change is appropriate for the device and fixture.
Classify the result explicitly
- Capture the numeric value and associated source status.
- Determine whether the required stimulus was achieved.
- Apply the recipe’s defined interpretation for compliance.
- Keep infrastructure or connection errors distinguishable from product failures.
- Retain the evidence before any retry changes the state.
Some tests deliberately seek a limiting condition. In that case, define what constitutes the expected transition and what data are needed to evaluate it. Compliance is not universally a failure; it is a state whose meaning depends on the test objective.
Validate the classification with representative expected and unexpected conditions. The released method should show that it can distinguish a real product response from a missing stimulus. A single numeric limit comparison cannot make that distinction when the instrument’s source state is absent from the record.
Move from reading to a build-ready project
Use the fabrication and assembly checklists to prepare drawings, production files and test requirements.
Open the DFM checklist