OpenCAL, initiated by researchers at UC Berkeley, makes computed axial lithography accessible through open hardware and software. It does not build layer by layer. Instead, angle-dependent patterns are projected into a rotating volume of photosensitive resin; their accumulated optical dose forms the three-dimensional geometry at once. Project documentation puts commercial component cost at about $1,600 and a typical exposure at one to two minutes.

What OpenCAL makes possible

The system is not intended to replace a finished industrial product. It is a research reference for volumetric manufacturing. Its GitHub repository supplies firmware, Raspberry Pi 5 control and assembly material, while the documentation covers hardware and calibration. VAMToolbox provides a separate reconstruction layer that produces the angular projections required to create the target dose distribution.

The accompanying manuscript describes a modular platform assembled from commercial parts. More than forty contributors participated, and the scientific lineage reaches back to earlier CAL work at Berkeley and Lawrence Livermore. That is a stronger foundation than an isolated hobby build, but it does not guarantee identical results in every laboratory.

What happens inside the resin

An individual projection need not reach the polymerisation threshold. Dose accumulated during rotation crosses the threshold only where the part should exist. The idea resembles computed tomography in reverse: rather than reconstructing a volume from images, it computes images from a desired volume.

Speed comes with coupled physics. Resin absorption and scattering, vessel geometry and refractive index, projector distortion, rotational synchronisation and threshold dose jointly determine the result. Changing resin, glassware or the optical path is therefore a recalibration, not a simple consumable swap.

Verified facts and limits

The repository, documentation, research manuscript and bill of materials are public and verifiable. The one-to-two-minute exposure and roughly $1,600 component figure apply to the project scope and may exclude laboratory infrastructure, safety controls, post-processing and metrology. Active development also means that commit, configuration and dependencies must be frozen for reproducible work.

CAL is not a geometry-without-limits process. Scattering, resin heating, over- and under-dose, vessel-wall optical error and background curing can blur features. Washing and post-curing can further change dimensions and material properties.

Reproduction checklist

A new build should first measure mechanical runout, rotational speed stability and vibration. A geometric artefact can then verify projector axis, magnification and distortion. Resin needs absorption, threshold-dose and exposure tests, followed by repeated dimensional measurements at several positions.

The minimum record includes bill of materials, firmware and software revisions, projection parameters, vessel, resin lot, temperature, rotation profile, washing, post-cure and measurement uncertainty. Without those fields, visually similar parts are not evidence of an equivalent process.

Editorial conclusion

OpenCAL matters because it provides a shared hardware and software starting point for a research field that was expensive and difficult to reproduce. It is neither a finished production machine nor a qualified recipe. Its full value will emerge when laboratories share calibration, failures, repetitions and raw measurements alongside striking minute-scale objects.

Sources and transparency

System architecture and cost data come from official OpenCAL documentation, its GitHub repository and the project manuscript. Short exposure time and easy assembly remain project-team claims; a public multi-laboratory round-robin has not yet been reported.