On this page
A ceramic circuit may receive a withstand test at the component supplier, another after attachment, and another at incoming inspection. Those events act on the same physical insulation even when different organizations own the records. A retest decision therefore needs a joined exposure history. The latest passing result cannot erase an earlier trip, an interrupted ramp or an unexplained change in leakage. The useful output is an authorized next action tied to the actual specimen and test boundary.
Measurement purpose
Determine whether a proposed additional withstand exposure is justified and preserve the meaning of every earlier result on the same circuit.
Specimens and conditions
- Identity chain
- Connect panel, circuit and assembly identifiers; retain rework and exception histories.
- Electrical boundary
- Document nodes, attached components, fixture and insulation construction for each stage.
- Condition before exposure
- Record temperature, moisture treatment, cleaning, recovery time and any intervening change.
Equipment and records required
- Withstand test system: Use the approved rated instrument and protected fixture; retain actual settings, status and available waveform records.
- Safe-state verification: Apply the laboratory's approved discharge and verification procedure after normal, failed and interrupted events.
Method sequence
- Reconstruct
Join all prior events to the specimen and classify their termination states.
Record: Chronological ledger with original source records.
- Review
Identify the unanswered requirement and the reason another exposure is necessary.
Record: Specific authorization or hold instruction.
- Evaluate
Perform only an approved sequence, then retain achieved conditions and observations without overwriting earlier results.
Record: New linked event and final disposition.
Decision and uncertainty
A new pass is interpreted with the earlier history and governing retest rule; it does not cancel an unresolved failure.
Missing exposure records, changed boundaries and incomplete waveforms limit comparison and must remain explicit.
The responsible electrical-safety and acceptance authorities authorize retesting and product disposition.
Traceable outputs
| Record | Required contents |
|---|---|
| Exposure ledger | Identity, waveform, ramp, hold, achieved stress, termination, recovery and intervening handling. |
| Retest decision | Purpose, supporting evidence, required controls, authority and unresolved prior events. |
Method review decisions
- Reconstruct every relevant prior exposure before scheduling another high-voltage test.
- Separate a measurement-system invalidation from a specimen failure and from an incomplete exposure.
- Keep actual waveform and electrical end state with the decision; do not invent an equivalent cumulative test time.
Join component and assembly histories by identity
Start with the physical circuit identifier, panel position where available, and any identifier added during assembly. Record which earlier identifier it replaced. A batch certificate showing a test was performed does not establish how many times this particular circuit was exposed, especially after rework or a line restart. Request exception records as well as normal passing records.
Follow changes in the tested boundary. A bare heater dielectric, the same heater bonded to a support, and the completed assembly with connected electronics can present different capacitance, leakage paths and potential distribution. Link the events chronologically without treating their results as interchangeable observations of one unchanged circuit.
Record an exposure as an event, not a pass counter
Each event needs the method revision, connected nodes, waveform, polarity, voltage convention, ramp, hold, termination and recovery interval. Preserve the achieved voltage and current history when available. A programmed setting is not proof that the source reached it or maintained it throughout the hold. Retain instrument range changes, compliance flags and trip messages.
Include the condition immediately before testing: temperature, moisture conditioning, surface cleaning, handling and elapsed time after an earlier exposure. These are not excuses to dismiss a failure. They identify which physical state was evaluated and whether a proposed repetition would answer the same question or introduce a different one.
Distinguish a failed specimen from an incomplete test
A protective interlock opening, a communications interruption and an upper-current trip should not all become a generic fail code. Preserve the original event and determine whether the required stress had been applied. A source that never reached its target has not demonstrated the requested withstand condition, but its partial ramp may still have exposed the specimen.
Invalidate an observation only when the record supports a specific measurement or setup problem. If a loose connection is found, document how that connection affected the measured boundary and why the original judgment cannot be used. Replacing the fixture and obtaining a pass does not by itself establish that the first failure was fictitious.
Do not turn cumulative seconds into an equivalence claim
A ledger can sum time spent at a named test condition, but that sum is only descriptive. Two separated holds include different ramp and recovery histories from one uninterrupted hold. Different voltages, polarities, waveforms or material conditions cannot be converted into one equivalent exposure using an arbitrary multiplier. Such a conversion would require a validated model and the governing acceptance authority.
For example, three hypothetical ten-second holds at the same setting produce thirty recorded hold-seconds and three ramp events. They do not prove compliance with a separately specified thirty-second continuous hold. Keep those counts in separate fields. Likewise, a trip after four seconds is neither a full ten-second pass nor an event that can simply disappear from the cumulative history.
Authorize the next action from the event evidence
A predefined decision table makes supplier, assembly and incoming teams handle the same event consistently. It should identify who may authorize another exposure and which evidence must survive that decision. Neither an operator preference nor an automatic retry setting should determine how many failures are tolerated before a pass is retained.
| Recorded event | Evidence needed before another test | Result that must remain visible |
|---|---|---|
| Complete required exposure with a current trip | Specimen and fixture investigation; approval under the applicable requirement | Original failing waveform and specimen condition |
| Interlock opens during the ramp | Reason for interruption, achieved stress and verified safe end state | Incomplete exposure, not a completed pass |
| Confirmed acquisition configuration error | Specific error and demonstration of its effect on the judgment | Invalidated observation with its reason |
| Assembly stage adds a different insulation boundary | Updated node diagram and authorized stage requirement | Both component and assembly records |
| Prior exposure history unavailable | Documented risk review and instruction from the acceptance owner | Unknown history; never an invented zero count |
Verify recovery without erasing the failure condition
Follow the instrument and approved laboratory procedure for discharge, electrical verification and handling after every termination state. Automatic completion on a display is not an independent demonstration that all connected capacitance is safe to touch. Required enclosures, interlocks and qualified personnel remain necessary during diagnosis; do not bypass them to watch a suspected discharge more closely.
Record any rest, drying, cleaning or disassembly performed before another measurement. These actions can change the evidence. If leakage decreases after drying, preserve the original moisture-conditioned result and describe the new state. A changed state can support a mechanism investigation without qualifying the original state as acceptable.
Look for changes across comparable exposures
Compare current histories only where waveform, nodes, timing and specimen condition are sufficiently matched. A consistent shift in a settled region differs from a larger charging transient caused by fixture capacitance. Keep the full trace beside any summary statistic so a changed shape is not hidden by one terminal value.
A progressive change is a reason to investigate, not a universal damage index. Conversely, a flat sequence of passing readings does not establish unlimited safe repetition. The laboratory should identify what additional information another exposure would provide and whether a less intrusive inspection could answer that question while retaining the as-found specimen.
Transfer the exposure ledger with the part
The next organization should receive specimen identity, test boundaries, event chronology, achieved stress, judgments, interruptions and authorized rework. A compact cover sheet can point to the waveform files, but it must not replace them with a total number of passes. Where commercial confidentiality limits sharing, agree the minimum condition and exception information needed for safe acceptance.
Freeze the retest instruction with the governing requirement and responsible authority. Reopen it when insulation construction, connected components, fixture, test waveform or production sequence changes. This review controls the meaning of repeated tests; it supplies no universal maximum test count, dielectric-life prediction or company certification.
Provide the withstand exposure history
Send the existing events and the requirement that a proposed retest must answer.
- Circuit and assembly identifiers with tested-node drawings.
- Supplier, assembly and incoming test records including exceptions.
- Actual voltage/current histories and the governing retest instruction.
- Any drying, cleaning, repair or change between exposures.
The drawing-upload form loads as you reach this section.