On this page
A dark line in a microscope image does not establish that laser trimming has removed the intended resistive film. Shadow, redeposited material and unresolved edges can look similar at one illumination setting. Conversely, electrical continuity across a trimmed resistor is normally expected because current must still pass around the end of the cut. Kerf inspection needs the intended current path, a resolved geometric observation and an electrical result that answer different questions.
Measurement purpose
Classify trim-kerf observations without confusing image contrast, geometry and electrical continuity.
Specimens and conditions
- Trim state
- Known element, trajectory and before/after processing state.
- Protective layer
- Record whether the observed surface is resistor film or overglaze.
Equipment and records required
- Optical inspection: Validated scale, contrast and repeat illumination at the specimen plane.
- Electrical correlation: Controlled resistor connections and excitation with retained pre-trim and settled final values.
Method sequence
- Locate
Mark intended neck and suspect residue separately.
Record: Registered overview and close-up.
- Correlate
Compare resolved dimensions with electrical history.
Record: Separate geometric and electrical conclusions.
- Resolve layer condition
Choose profiling or authorized sectioning only for the remaining material question.
Record: Analysis location and preserved pre-intervention state.
Decision and uncertainty
Apply the agreed feature definition and uncertainty rule; unresolved material is not an automatic visual pass.
Edge location and layer visibility limit conclusions.
Trim-process and drawing owners.
Traceable outputs
| Record | Required contents |
|---|---|
| Kerf inspection | Coordinates, raw images, edge bounds, electrical state and disposition. |
| Trim correlation | Trajectory, intended current path, resistance progression and unresolved material observations. |
Method review decisions
- Distinguish the intended conducting neck around the trim tip from an unintended bridge across the cut.
- Keep optical classification separate from final resistance acceptance.
- Escalate unresolved film condition without repeating laser passes merely to improve image contrast.
Mark the intended remaining current path
Start with the untrimmed resistor outline, conductor contacts and approved trim trajectory. Identify where the laser enters the film, where it stops and which region carries current after adjustment. An L-cut, plunge cut and multiple-cut pattern do not leave the same current distribution. The inspection drawing should therefore mark the local feature being assessed rather than applying one generic clean-line criterion to every geometry.
A narrow conducting region around the cut tip can be intentional. Material spanning the width of a supposedly cleared section is a different observation. Calling both features a bridge creates conflicting acceptance decisions. Use separate names such as remaining neck and possible kerf residue, and preserve their positions on an overview image. Neither the darkness nor the brightness of that region alone identifies its electrical function.
Retain the image before changing the specimen
Record the resistor identifier, printed-face orientation and trim coordinates before cleaning, probing or additional processing. Keep an overview that links the close-up to the correct element in a network. Where pre-trim images exist, use stable conductor or substrate landmarks to register them with the post-trim image; do not align solely on edges altered by the laser.
Preserve raw or minimally processed captures alongside any display adjustments. Sharpening can create dark borders, while compression can merge a narrow gap into a single line. Record magnification, image scale, illumination, focus position and exposure. A screenshot pasted into a report without its scale and original image cannot support a dimensional decision about a small kerf.
Resolve both edges, not just a dark centerline
Evaluate whether the imaging system separates the two boundaries relevant to the measurement. Pixel spacing is only one part of that evaluation: optical contrast, focus, surface texture and sampling all influence edge visibility. Increasing digital zoom enlarges the same information and cannot establish a previously unresolved cut. Check resolution at the actual specimen plane and capture settings.
Use a stable, calibrated scale and avoid choosing a measurement line through obvious glare or debris. For an irregular kerf, specify whether the reported width is a local minimum, a median over a defined length or another drawing-defined quantity. A single favorable cross-section does not represent the narrowest region. The general object-space sampling method remains separate from this kerf-specific interpretation.
Change illumination before declaring material absence
Repeat the same field with a controlled change in illumination direction or contrast mode. A shadow may reverse or move with the light while a geometric boundary remains registered to the resistor. Focus through the relevant surface height without moving the measurement datum. These comparisons help distinguish an imaging artifact from a persistent feature; they do not establish chemical composition.
Where a translucent overglaze covers the resistor, distinguish the visible coating surface from the underlying resistive layer. Reflections or a surface mark in the glass can coincide with the projected kerf. Do not scrape the coating merely to make the image easier to interpret. Document the layer being observed and obtain approval for any method that changes the protective surface.
Apply uncertainty to the remaining neck
Consider an inspection example in which the apparent remaining neck is 60 µm wide. Suppose the method review assigns a conservative edge-location bound of ±3 µm to each of its two boundaries. Adding the opposing edge contributions gives a width interval of 54 to 66 µm. This is a worst-case interval for that example, not a statistical confidence interval or a laser capability specification.
If the drawing requires at least 55 µm of remaining neck, a displayed 60 µm does not by itself establish a clear pass under that conservative rule. The lower bound extends below the requirement. Improve the validated measurement or use the agreed decision rule; do not round the reading upward. The purpose is to prevent a visually plausible margin from concealing unresolved geometry.
| Observation | What it supports | What remains unresolved |
|---|---|---|
| Dark line at one lighting angle | Location for further inspection | Film removal and kerf width |
| Two stable boundaries under repeat capture | A geometric width estimate | Material state beneath a coating |
| Resistance reaches its target | Electrical endpoint at the test condition | Local residue and current concentration |
| Neck estimate 60 µm with ±6 µm bound | Possible width 54–66 µm | Compliance with a 55 µm minimum |
Use resistance as correlation, not a photograph substitute
Compare the pre-trim value, trim progression and settled final resistance under the declared measurement conditions. An unexpected small change can motivate a review of cut location or residual paths, but it is not unique evidence of incomplete ablation. Different sheet resistance, contact position, parallel network paths or measurement connections can produce similar electrical behavior.
A correct final value also does not prove every local feature is acceptable. Two geometries can have similar terminal resistance while distributing current differently. Do not apply extra laser passes solely to eliminate a dark spot after the resistor has reached its target: material removal can change resistance irreversibly and reduce the remaining neck. Endpoint control belongs with the trim method, not an improvised visual repair.
Choose an independent confirmation method deliberately
For a persistent unresolved indication, select the next method according to the disputed fact. Surface profiling may clarify topography; a carefully located section may reveal layer continuity; an electrical study can investigate the accessible current path. Each method has limitations, and preparation can remove or smear the very residue under investigation.
Reserve destructive analysis for identified specimens under an approved plan. Record the exact section location, neighboring retained area and pre-analysis electrical state. Do not imply that a clean section at one position proves the full length of the kerf is clear. If the available method cannot distinguish an acceptable remaining neck from unintended material, classify the result as unresolved and obtain a technical disposition.
Make the kerf decision reproducible
The inspection record should connect the trim trajectory, raw registered images, dimensional definition, uncertainty and final electrical result. State whether the conclusion concerns resolved geometry, material continuity or the electrical endpoint. This prevents a later reviewer from interpreting a visual pass as proof of long-term stability or a resistance pass as proof of film removal.
For a custom laser-trimmed thick-film resistor enquiry, provide the element layout, intended cut style, final resistance requirement and available inspection access. ChipSimple can review the drawing-specific relationship between trim geometry and inspection needs. Any minimum neck, kerf or film-condition criterion must come from the agreed material and process qualification, not from the illustrative dimensions above.
Share the trim geometry
Send the resistor layout and the actual unresolved feature so the inspection method addresses the right current path.
- Resistor and conductor geometry with trim entry, endpoint and allowed neck.
- Raw scaled images with illumination and coating state.
- Pre-trim and final resistance values with measurement connections.
- Drawing limits and permission boundaries for additional analysis.
The drawing-upload form loads as you reach this section.