How to find wafer 2D code locations before inspection

A wafer identifier can be readable yet outside the region your inspection station searches. Diameter alone does not tell you which face carries the mark, where it lies relative to a notch or flat, or which drawing controls its position. This procedure helps you establish the permitted location, measure its presentation in your fixture, and select an image region before committing to hardware. Treat 200 mm and 300 mm wafers, and notch and flat references, as separate recipes until approved documentation establishes otherwise.

Establish which document controls the location

1. Collect the released wafer drawing and marking requirements. Obtain the applicable marking standard through its publisher and the approved drawing through the authorized wafer supplier or customer. Record document identifiers, revisions, wafer type, and any purchase specification that changes the marking requirement.

Do not use a remembered dimension or an image from another production line as an acceptance requirement. A standard may define a marking convention, while the released drawing specifies the requirement for your particular wafers. If they disagree, obtain a documented resolution from the responsible quality or design authority before configuring acceptance limits.

2. Extract the complete marking definition. Your setup sheet should identify:

• Wafer diameter and permitted dimensional variation.
• Notch or flat geometry and its reference direction.
• Marked face and the direction from which the drawing is viewed.
• Symbol type, marking method, and required data format.
• Location dimensions, tolerances, and orientation limits.
• Symbol envelope, required clear space, and applicable edge restrictions.
• Any approved alternate marking location.

An absent dimension is not permission to choose an arbitrary location. Ask for clarification instead. Confirm the symbol family before selecting decoder settings; the image-to-validated-data workflow explains why locating a pattern, decoding it, and accepting its contents are separate operations.

Define a wafer-referenced coordinate frame

3. Write down the coordinate convention before measuring. A practical internal convention places the origin at the wafer center and defines an axis from the approved notch or flat reference. This is your inspection convention, not a claim about any marking standard.

For a notch, specify exactly how its reference direction is derived. For a flat, specify whether the reference uses the flat midpoint, its normal, or another drawing-defined construction. These definitions are not interchangeable.

Record which face the camera sees. Viewing the opposite face can reverse the apparent relationship between a drawing and the image. Use a known reference sample to check the mapping rather than correcting it by intuition.

4. Map wafer coordinates into camera coordinates. For a planar setup with a calibrated scale and negligible perspective error, the following is a useful model:

This simplified model assumes consistent axis directions and the same viewed face. Account for image-axis reversal explicitly. If the camera is oblique or lens distortion is significant, use a calibrated mapping rather than forcing this model to fit.

Diameter remains a recipe parameter. Do not scale a location measured on one wafer diameter to another unless the governing drawing explicitly defines that relationship.

Confirm the location on representative wafers

5. Measure actual presentation without replacing the drawing. Inspect authorized samples from each relevant diameter, reference geometry, marked face, and production variant. Follow the site's wafer-handling and contamination-control procedures.

Begin with a view large enough to find the expected region. Record the wafer center, orientation reference, mark center, and complete symbol boundary. Save the original image alongside the recipe and document revision.

Measure these contributions separately:

Repeated images of one stationary wafer characterize imaging repeatability, not placement variation. Include reloads and representative production samples when evaluating the fixture.

Keep measured variation separate from specified limits. A small sample showing little movement does not justify narrowing the permitted search area below the documented tolerance envelope.

Calculate the search region and pixel requirement

A mark near the perimeter can move substantially under a small angular placement error. For small angles, its displacement is approximately the radial distance multiplied by the angular error in radians.

Add independent worst-case allowances when every permitted placement must remain visible. A statistical combination requires a justified process model and an agreed acceptance basis.

Worked example: a hypothetical search window

Suppose an approved application has a mark center 95 mm from the wafer center. Fixture rotation is limited to ±1°, and translation is limited to ±1 mm along each image axis. These are example application values, not standard wafer-marking dimensions.

The approximate angular displacement is:

Assume the maximum projected symbol envelope, including required clear space across the allowed orientations, is 4 mm wide. Allow another 1 mm on each side for optical and localization margin.

Using the same angular allowance in both axes is conservative here. A coordinate transformation can produce tighter axis-specific bounds if space or resolution demands them.

Next, suppose the smallest module is 0.10 mm wide. Choose an initial trial target of 5 pixels per module, then validate it with real marks. This is an engineering starting point, not a universal decoder requirement.

The trial image therefore needs at least 600 useful pixels across that field, with comparable vertical sampling. Blur, distortion, glare, and weak module contrast can still prevent reading.

Field coverage and sampling must be checked together: widening the view without adding useful pixels reduces the sampling of each module.

Choose a fixed reader or camera workflow

Use the fixed code readers category when the main task is identification and a candidate reader can demonstrate the required mark contrast, symbol support, and integration behavior. Use area scan cameras when you need greater control over image access, localization, or additional inspection.

The LS-CRD-0028 provides a catalog-listed global shutter and Fast Ethernet interface. Treat it as a candidate for sample evaluation, not as a confirmed wafer-marking solution. Verify the required working distance, symbol family, contrast, and access to setup images before selection.

The LS-CMR-0166 provides a catalog-listed USB3 interface and rolling shutter. Evaluate it with a stationary wafer during exposure; moving presentation needs a separate assessment of row-timing distortion and motion blur. Lens selection, illumination, decoder software, and host integration remain application decisions.

Use the camera and reader finder to compare interfaces and imaging options after the field requirement is known. For a separate camera workflow, the camera configuration starting point helps organize the remaining optical and integration choices.

Configure the recipe and verify acceptance

6. Configure location before optimizing decoding. Establish a repeatable image with the full permitted mark region visible. Adjust focus and illumination using representative marks, including difficult samples. A reflective surface may require changes in illumination angle; a brighter exposure alone can increase glare.

Set the initial localization region from the approved tolerance envelope and fixture allowances. Keep required clear space inside the image. Restrict the decoder to the documented symbol family where supported, but do not shrink the image region merely to make one sample read faster.

7. Verify both correct reading and correct rejection. Test marks near the permitted positional limits and repeat the test after unloading and reloading. Include the maximum allowed rotation, each approved wafer variant, and representative low-contrast marks.

Check that:

• The entire required region remains visible and acceptably focused.
• The decoded identifier belongs to the wafer currently presented.
• An absent or unreadable mark produces an explicit failure state.
• A neighboring identifier cannot satisfy the acceptance rule accidentally.
• Recipe changes cannot leave a previous identifier active.
• Archived images can be associated with the corresponding inspection event.

A successful decode does not establish compliance with marking-quality requirements. The distinction is explained in barcode reader verification for inspection. Apply the actual wafer-marking acceptance requirements, not a label-specific assumption.

Similarly, do not treat a decoder score as a universal probability of correctness. Use the checks described in judging barcode read confidence, together with expected length, permitted characters, and any defined identifier structure.

Recognize when the usual setup is wrong

A small fixed search region works only when the wafer pose and marking convention are controlled. If angular placement varies widely, first locate the wafer reference or use an approved orientation process. Simply enlarging the region may leave too few pixels per module.

If the mark is on the opposite face, software cannot recover an image that the optics never capture. Resolve the presentation and handling arrangement before purchasing the reader.

If one recipe must cover several approved marking conventions, define explicit branches and validation rules. Do not search indiscriminately across both faces, every orientation, and every symbol type without a documented reason.

When readable marks fall outside the permitted area, the reader may still decode them. Keep mark-location acceptance separate from identifier decoding if placement itself is an inspection requirement.

Checklist before ordering or releasing the station

• Documents: Record the controlling drawing, marking requirement, and revisions.
• Variants: Separate diameter, notch or flat, marked face, and approved alternate positions.
• Coordinates: Confirm datum construction, viewed face, axis directions, and image mapping.
• Coverage: Include symbol boundaries, required clear space, placement tolerance, and fixture movement.
• Sampling: Calculate pixels per module and validate with actual mark contrast and focus.
• Hardware: Confirm working distance, illumination compatibility, interface, and image access.
• Acceptance: Distinguish valid identification, permitted location, and marking quality.
• Recovery: Save the validated recipe and associated reference images before later adjustments.

The reader configuration backup procedure provides a useful handover structure for settings and reference captures.

For the next step, send the approved location dimensions, representative wafer images, calculated search field, and required interface through application support. Those inputs make a hardware evaluation meaningful without assuming a universal wafer code position.