What Printing Registration Tolerance Should We Expect on Packaging?

The Short Answer

On a molded, curved container, a realistic registration tolerance for multi-color printing is around plus or minus 0.2 to 0.3 millimetres between colors, with wider tolerances on parts with significant dimensional variation. That is tighter than many designs assume and looser than some brand guidelines demand. The tolerance depends on the process, the part geometry and how well the container is held during printing, so it should be agreed per project rather than taken from a general figure.

Where Registration Error Comes From

Every stage adds error. The container itself varies in diameter and roundness; the printing fixture locates the part with some clearance; the screen or plate is positioned with its own tolerance; and each successive color pass adds another positioning step. On a cylindrical body printed by rotating the part against a screen, the main variable is the part's diameter, because a smaller diameter rotates through a different arc for the same linear screen movement. That is why registration on molded parts is looser than on flat sheet.

Typical Values by Process

Pad printing on a flat or gently curved cap can hold around plus or minus 0.1 to 0.2 mm. Silk screen on a cylinder is typically plus or minus 0.2 to 0.3 mm. Hot stamping on a flat panel can approach plus or minus 0.1 mm but degrades on curved surfaces. Labels applied automatically usually hold plus or minus 0.3 to 0.5 mm depending on the applicator and the container's roundness. These are practical figures for well-controlled production; tighter values require special fixtures, slower running and higher cost.

Designing Within the Tolerance

The most effective response is design. Keep a clear gap of at least 0.5 mm between adjacent colors so that normal variation does not produce visual contact. Avoid thin outlines in a second color around a first-color shape, because misregistration on an outline is far more visible than on a solid block. Do not place small text in a second color over a curved area. Where two colors must align precisely, consider printing them as a single pass with a combined stencil, or converting one element to a label that is applied in a separate, more controllable operation.

Agreeing the Specification

Registration should be written into the specification with a value and a measurement method. The method matters: measuring the offset between two features on a magnified image is reproducible, while judging by eye is not. The specification should also state the sample size and the acceptance criterion, for example a maximum offset on any of ten sampled parts. Without these, a discussion about whether a batch is acceptable becomes a matter of opinion, which is expensive for both sides.

Checklist

• Agree a numeric tolerance with a measurement method.
• Keep at least 0.5 mm clearance between adjacent colors.
• Avoid thin second-color outlines and small multi-color text.
• Confirm the tolerance is achievable on the actual part geometry.
• Approve a production sample, not a flat proof.
• State sample size and acceptance criteria in the specification.
• Check registration after several thousand impressions, not only at start-up.
• Record offsets per batch for trend monitoring.

When a Design Uses Many Colors

Each additional color adds a positioning step, so tolerance accumulates. A four-color design will not hold the same registration as a two-color one, and the difference is visible where elements are meant to touch. Where a brand insists on a complex multi-color graphic on a curved container, a shrink sleeve or a label produced on a flat press gives far better control, because the printing happens on a flat substrate and the shape is created afterwards. Converting the design rather than fighting the process is usually the cheaper decision.

A final practical point: registration should be checked against the approved sample at the end of a run, not only at the start. Plates wear, screens stretch and fixtures shift, and the drift across a long run can exceed the tolerance that was met on the first hundred pieces. Sampling the last pieces of a run, and comparing them with the first, reveals whether the process is stable or merely started well.

Conclusion

Registration tolerance is a property of the process and the part, and it is larger on molded, curved containers than on printed flat surfaces. Designing with clearance, avoiding fragile multi-color details and agreeing a numeric specification with a measurement method prevent most disputes. Where precision is essential, a label applied in a separate operation is often more reliable than adding a print pass.

References

• ISO 12647-1:2013, Graphic technology — Process control for the production of half-tone colour separations, proof and production prints.
• ISO 2409:2020, Paints and varnishes — Cross-cut test.
• ASTM D5264-98(2019), Standard Practice for Abrasion Resistance of Printed Materials.
• ISO 105-A02:1993, Textiles — Tests for colour fastness — Grey scale for assessing change in colour.
• ISO 2859-1:1999, Sampling procedures for inspection by attributes.
• ISO 9001:2015, Quality management systems — Requirements.