Resistor End Overlap: Registration Margin Versus Active-Length Control

1. State the two competing outcomes

Capture means the resistor film continuously meets the intended conductor despite relative print displacement. Repeatable active length means the current-carrying body can be bounded consistently after firing. These outcomes can conflict: increasing nominal overlap may improve geometric capture yet move more of the resistor beneath the conductor, where current transfer is not described by a simple rectangular body.

Start with the drawing function and failure consequences. An open or weak transition, an out-of-window total resistance and an unstable model are different failures. State which one controls the design and who owns its criterion. This page addresses the overlap-versus-length choice; it does not replace broad resistor geometry or contact-material selection.

2. Describe overlap with four boundaries

Record the fired resistor start edge, conductor end edge, intended body boundary and probe or sense boundary at each termination. “Overlap” without those definitions is ambiguous. Measure left and right ends separately because a common registration shift can increase one overlap while decreasing the other. Retain signed coordinates in a substrate datum system.

Inspect edge shape, conductor encroachment and local narrowing rather than reducing each end to one scalar. A ragged resistor edge may satisfy nominal overlap at one point but fail continuous capture across width. A broad conductor edge can also change current entry. Scaled images tied to specimen identity allow the electrical data to be interpreted later.

3. Build the registration capture budget

For each print axis, combine artwork placement, screen-to-substrate registration, feature-width change and measurement allowance using the drawing owner’s approved tolerance method. The available capture margin is the smallest expected fired overlap after those directional effects. Do not mix worst-case and statistical terms without an explicit rationale.

If nominal overlap is 0.40 mm and the evaluated relative closing error is 0.15 mm, the simple remaining margin is 0.25 mm. That arithmetic does not establish sufficiency; it only locates a geometry input for the controlled study. No process capability or universal minimum is implied.

4. Measure active length on fired parts

Define active length between reproducible fired transition landmarks and quantify the operator or algorithm repeatability of that definition. Compare optical active length with the length inferred from a same-width resistor series. Disagreement is evidence that the end field or chosen landmark needs attention, not permission to tune the sheet value.

Short bodies deserve particular scrutiny because the two transition regions occupy more of the path. Report the ratio of uncertain end length to body length. A fixed absolute ambiguity that is negligible on a long coupon can dominate a miniature resistor and can make nominal number-of-squares calculations misleading.

5. Separate body resistance from end behavior

Use same-lot witnesses or long bodies to establish applicable sheet behavior, then compare total resistance across controlled overlaps. A four-wire connection can reduce fixture and lead terms but does not remove the printed conductor-resistor transition. Keep measurement temperature, excitation, settling time and probe position common.

Plot measured minus body-predicted resistance against left overlap, right overlap, active length and panel position. A shift correlated with only one end suggests asymmetric current entry. A panel-position shift suggests process state. A constant offset across overlap may be an end contribution, but it requires independent replication before use.

6. Run a controlled overlap ladder

Create at least three overlap levels that retain the same body width, nominal active span, conductor material, firing route and terminal pad. Interleave variants across independent panels instead of assigning one level to each position. Measure fired geometry before grouping results by nominal design.

The ladder should extend far enough to challenge both capture and active-length repeatability without deliberately producing unsafe specimens. Stop if a variant has incomplete physical contact, damaged edges or an unmeasurable landmark. Preserve exclusions with photographs and reasons; silently discarding low-overlap failures biases the trade-off.

7. Choose from evidence, not nominal size

A useful decision table separates outcomes: incomplete capture routes to more margin or tighter registration; adequate capture with length-dependent residuals routes to a refined end model; stable resistance but poor optical repeatability routes to a better landmark or metrology method. Do not collapse these into one average score.

Select the overlap only after uncertainty bands for capture, fired active length and electrical contrast are visible. If two levels are electrically indistinguishable, prefer based on documented robustness and inspectability rather than claiming a performance difference the study cannot resolve. Acceptance remains by drawing.

8. Freeze the transferable construction

Release the nominal overlap, fired-overlap measurement definition, allowable registration state, conductor/resistor material sequence, width range and electrical test state together. A new conductor termination, print order, body aspect ratio or probe boundary reopens transfer.

Archive coordinate data, images, raw resistance, exclusions and the comparison model. Validation fails when overlap is known only from artwork, variants are confounded with panel position, or active length cannot be repeated. The RFQ should request these missing inputs rather than promise a standard overlap.

For change control, compare the new fired transition map with the released map before relying on resistance alone. Two constructions can reach the same total value through different body and end contributions, leaving different sensitivity to future registration movement.