A GRIN rod lens array can meet its catalog specification and still produce banding, blur, or stray light when installed in a contact image sensor. Platen thickness, working distance, rod alignment, and illumination geometry all influence the signal that reaches the sensor.
TracePro optical design and analysis software supports gradient-index modeling and non-sequential system analysis. Modeling the rod lens array with the platen, illumination, and sensor helps engineers distinguish array behavior from losses and unwanted paths elsewhere in the assembly.
A gradient-index lens bends rays through a spatial change in refractive index inside the material. For a radial GRIN rod, use the supplier's index profile, wavelength dependence, diameter, length, and working-distance data. An equivalent homogeneous cylinder will not reproduce that internal propagation.
Keep two meanings of pitch separate. GRIN optical pitch describes the ray oscillation through the material. Array pitch describes the spacing between neighboring rods. Both affect the system, but they are different design quantities.
Validate a single-rod model against the supplier's conjugate and imaging data before building the array. Check the selected GRIN property definition and numerical accuracy settings. A result that changes materially when tracing accuracy is tightened needs further convergence work.
Place the rods at the specified spacing, row offset, and orientation. Include the surrounding matrix, end-face properties, platen glass, and sensor position. Contact image sensor arrays combine overlapping images from neighboring rods, so alignment and working distance influence their combined response.
Begin with a uniformly emitting document plane to isolate the array's transmission pattern. Then add the actual illumination and document reflectance to evaluate the full assembly. This progression makes it easier to identify whether a scan-line variation originates in the array or the light guide.
Use sensor-plane irradiance, in radiometric units, for the optical power distribution. Convert it to a predicted detector signal using spectral responsivity, exposure time, and pixel area. Photopic illuminance in lux weights light for human vision and is not a substitute for detector response.
Extract an irradiance profile along the scan line and examine periodic variation near the rod spacing. Define the ripple metric explicitly, including the analysis region and whether peak-to-valley variation is normalized by the mean or another reference.
Choose receiver sampling fine enough to resolve the features being evaluated. At 600 dpi, the nominal sample spacing is about 42 micrometers. Use sufficient rays per analysis region and repeat selected cases with higher sampling to distinguish physical ripple from Monte Carlo noise.
Replace the uniform object with a bright-to-dark edge to evaluate geometric edge spread. Compare the response at several array positions and wavelengths. Treat this as a geometric system result; diffraction, pixel integration, and electronic processing require additional consideration when predicting final image quality.
Move the object through the specified height range while retaining the real platen thickness and refractive index. At each position, compare edge width, ripple, and collected power against the acceptance criteria.
This defines the usable depth of field for the assembly. A platen change can move the effective optical conjugate even when the mechanical spacing appears unchanged. Check each illumination band because the GRIN profile and focus can vary with wavelength.
Evaluate rod tilt, array bow, and sensor position separately before combining their limiting cases. The results show which assembly dimensions consume the available focus margin.
Illuminate a limited object region and inspect power reaching the intended dark region of the sensor. Include rod-wall interactions, matrix absorption, and end-face reflections. Compare an ideal absorbing boundary with characterized production properties to understand the contribution of the surrounding material.
Normal overlap between rod images is part of array operation. Crosstalk assessment should isolate light arriving outside its intended image region, then use path information to identify the mechanism. A material change should address that mechanism rather than assume that lower reflectance or a different index always improves performance.
Trace the LED light guide, platen, document, array, and sensor together. Include a suitable reflectance model for the document, especially when glossy surfaces can direct light into the array. Compare the model with measured uniformity and edge response from a representative assembly.
Request a TracePro trial to evaluate your GRIN rod lens array in its intended contact image sensor geometry before selecting the final array and assembly tolerances.