On this page
A resistor can show an acceptable temperature coefficient during an excursion yet return with a different resistance at the starting temperature. These are different observations and should lead to different decisions. Keep the original baseline, the in-cycle readings and the recovered reference-state readings together. Renormalizing every stage to its newest value can hide a retained change that matters to the customer's circuit.
Measurement purpose
Determine whether a thermal excursion produced a reversible resistance response, a direction-dependent response, or a retained shift under matched reference conditions.
Specimens and conditions
- Identified starting state
- Retain product identity, trim and protection stage, earlier exposures and the initial conditioned resistance.
- Matched return state
- Compare actual temperature, excitation, support and contact boundary before interpreting a return difference as product change.
Equipment and records required
- Time-ordered thermal and electrical record: Associate every resistance result with temperature, direction, elapsed time and method status.
- Independent measurement checks: Use appropriate check artifacts or an independent route to investigate station drift without overwriting product history.
Method sequence
- Record the baseline
Acquire the defined reference-state resistance before the excursion and retain its acquisition conditions.
Record: Initial baseline and thermal state.
- Observe the excursion
Collect specified heating and cooling points without merging branches or resetting the baseline.
Record: Chronological paired observations.
- Evaluate recovery
Acquire justified return-state observations and compare with the preserved initial state.
Record: Retained difference, recovery timing and diagnostic evidence.
Decision and uncertainty
Temperature-coefficient acceptance and retained-shift acceptance are evaluated as separate requirements. A favorable result in one does not cancel an unfavorable result in the other.
Temperature mismatch, contact changes and station movement can imitate a retained shift; a final offset does not identify when the change occurred.
The product and test owners define the separate limits and recovery conditions before the excursion.
Traceable outputs
| Record | Required contents |
|---|---|
| Thermal chronology | Initial, in-cycle and return observations with immutable baseline identity. |
| Separate dispositions | Coefficient result, branch comparison, retained-shift result and unresolved cause where applicable. |
Method review decisions
- Preserve the starting reference resistance even when later calculations use another explicitly identified denominator.
- Compare return readings only after confirming equivalent temperature, excitation and contact conditions.
- Do not reconstruct the timing of retained drift from a before-and-after difference alone.
Ask three questions of the same sequence
The first question is how resistance varies with temperature over the specified interval. The second is whether heating and cooling give the same resistance at matched temperatures. The third is whether the specimen returns to its original resistance after the prescribed recovery conditions. These questions are related but not interchangeable.
A small endpoint coefficient does not demonstrate negligible retained drift. Likewise, a zero final offset does not prove that heating and cooling followed one reversible curve. Define the needed outputs before choosing temperature points so the test does not collect only the data convenient for a single summary number.
Retain a numerical sequence without hiding the offset
In an illustrative sequence, a resistor measures 1,000.00 ohms at the initial 25 °C state, 1,002.00 ohms at 75 °C and 1,000.30 ohms after returning to the matched 25 °C state. The initial-to-hot endpoint calculation gives 40 ppm/K. The return-state shift is separately 300 ppm relative to the initial resistance.
If the hot value is instead normalized using the final 1,000.30-ohm reading, the resulting coefficient changes to approximately 33.99 ppm/K. That arithmetic is possible, but it answers a differently defined question. It must not silently replace the original coefficient or erase the recorded 0.30-ohm retained difference.
Use the sequence record to preserve physical meaning
A concise chronology makes the report auditable. Add actual temperatures and measurement uncertainties to the working record rather than assuming the nominal values in this illustrative layout.
| Observation | Illustrative resistance | What it establishes |
|---|---|---|
| Initial reference state at 25 °C | 1000.00 Ω | Starting denominator and baseline identity |
| Hot state at 75 °C | 1002.00 Ω | Temperature-associated response at that time |
| Return state at matched 25 °C | 1000.30 Ω | 300 ppm retained difference from initial state |
| Later recovery observation | Measure; do not assume | Whether the return shift continues to change |
| Matched heating and cooling point | Two separately retained readings | Direction dependence at a common temperature |
Confirm that the reference state was actually restored
Returning the chamber controller to its starting setpoint is not sufficient. Verify specimen-associated temperature and the electrical observation window. A small remaining temperature difference multiplied by the local resistance sensitivity can look like a retained resistance change.
Also preserve contact geometry and excitation. Reconnecting the part between initial and final readings introduces a different comparison from leaving it connected throughout. Where reconnecting is unavoidable, include representative contact variability and inspect for altered pad surfaces. Do not assign every observed difference to the fired film before these method effects are assessed.
Separate branch difference from incomplete equilibration
A difference between heating and cooling readings at the same indicated temperature can arise from state dependence, but it can also arise because the specimen or temperature probe has not equilibrated. Retain direction, time and local temperature histories to distinguish these explanations.
Compare branches under a thermal-state criterion suitable for the measurement requirement. Increasing dwell may help reveal a lag-related effect, but it should be a documented diagnostic comparison, not a selective effort to obtain the preferred value. If the difference persists, report it as observed branch behavior without asserting a microscopic mechanism unsupported by the evidence.
Do not subtract an assumed drift trajectory
The initial and final reference measurements reveal a net change, not its timing. The change may have occurred during heating, at the hot plateau, during cooling or gradually throughout the cycle. Subtracting a linear interpolation of the final offset from every intermediate reading imposes an unverified history.
Such a correction is justified only when an appropriate independently tested model supports it and the resulting uncertainty is retained. Otherwise preserve the raw coefficient and the separate return shift. Additional intermediate reference observations may answer the question in a redesigned experiment, but they also change the exposure sequence and must be planned accordingly.
Define recovery observations before seeing the outcome
If the requirement includes a recovery period, specify the relevant temperature, humidity, electrical loading and timing. A later reading is a different observation, not a replacement for the first return reading. Record both when recovery behavior is part of the interpretation.
Avoid repeating reference measurements until one falls inside the acceptance limit and retaining only that point. Define the retest purpose and allowable sequence in advance. An evolving result may be useful evidence of relaxation or a measurement disturbance, but selecting the most favorable point prevents a reproducible qualification decision.
Report coefficient and retained shift as separate outputs
The final record should identify the denominator used for each coefficient and each retained-shift calculation. Include the exact temperature interval, heating or cooling direction, recovery state and uncertainty. If one requirement passes and another does not, retain both dispositions and the associated responsibility for further investigation.
For a customer circuit, the allowed in-service resistance movement and the allowed permanent offset may consume different parts of the system error budget. Request both requirements rather than assuming a TCR limit is a complete stability specification. This keeps thermal characterization connected to the actual drawing and circuit function.
Define thermal response and retained-shift requirements
Send the required excursion and the separate resistance limits at temperature and after recovery.
- Initial process stage, thermal history and preserved reference resistance.
- Temperature sequence, actual specimen temperatures and electrical conditions.
- Required TCR definition, branch comparison and retained-shift limit.
- Recovery timing, retest rule and any independent measurement checks.
The drawing-upload form loads as you reach this section.