Digital Fabrication

Digital fabrication is the direct production of physical objects from digital models by computer-controlled machines. Its defining property is the absence of manual transcription: no one measures a drawing, marks out a sheet, or interprets a dimension. The model is converted into machine instructions and executed. Everything else — which machine, which material, which process — is detail below that principle.

How it works

The chain runs model → toolpath → machine → object, and each link has its own failure modes.

Modelling produces geometry that is not only correct but manufacturable. Closed solids, consistent normals, tolerances compatible with the tool, and features larger than the tool can resolve are all requirements that a visually acceptable model may fail.

CAM and toolpath generation convert geometry into motion. For subtractive work this means selecting tools, stepovers, depths, and feed rates. For additive work it means slicing into layers, generating contours and infill, ordering paths, and placing start and stop points where a seam will be least visible or least structurally significant. A post-processor emits code in the controller's dialect.

Machining or deposition executes the code. Here the physical world intervenes: tool wear, machine deflection, thermal effects, and — in concrete — the changing rheology of a material that is stiffening while being processed.

Finishing and verification close the loop. Digitally fabricated parts still need deburring, curing, surface treatment, or assembly, and verification by measurement or scanning is what confirms that the object matches the model rather than assuming it.

Concrete fits into this picture through what research literature calls digital concrete: extrusion-based 3D printing, particle-bed processes where a binder is jetted into a bed of aggregate, shotcrete 3D printing, digitally controlled slipforming, and digitally fabricated formwork such as milled, printed, or knitted moulds. The last of these is frequently overlooked and is often the most practical route to a complex concrete element today, because the resulting element is ordinary cast concrete and can be engineered with ordinary rules.

Key parameters

The choice between these is usually decided by three questions: what resolution the surface needs, whether the element must be structural, and how many copies are required. Additive extrusion wins on one-off complex geometry; milled formwork wins when a genuinely smooth, structurally conventional element is needed; particle-bed processes win on fine detail at small scale.

Applications

In construction and related production, digital fabrication covers CNC-cut timber and panel systems, milled and printed formwork, additively manufactured concrete elements, laser- and plasma-cut steel, digitally bent and welded reinforcement, and CNC production of furniture, moulds, and fittings. It is also the mechanism behind mass customisation: because retooling cost approaches zero, a product family can vary from unit to unit without a cost penalty. That property changes what is worth designing: variants that would never justify a mould become ordinary output, and a catalogue can be a definition rather than a fixed list of items.

Advantages

The model-to-machine link removes an entire class of transcription errors. Output is repeatable and measurable, so quality can be controlled at the process rather than by inspection. Geometric complexity stops being a cost driver. Lead times shorten because no tooling has to be made for a one-off. Variation and customisation become economically viable at small volumes. And the whole process leaves a digital record, which makes traceability and qualification straightforward compared with craft production.

Limitations

Machines are expensive and only economical when utilised, which is a difficult condition for small operations and one-off projects. Model quality governs everything downstream, and a large share of construction model data is not fabrication-grade. The skills required span design, CAM, and shop-floor practice, and are scarce in that combination. Each process has hard physical limits — tool radius, bead width, overhang, envelope size — that constrain design and are easy to overlook until a part fails. Material behaviour, particularly for cementitious materials, is not fully captured by the digital model, so simulation does not replace physical trials. And in construction specifically, the regulatory and contractual framework assumes conventional methods, which slows adoption regardless of technical merit.

Concreef context

Concreef's fabrication chain is a parametric model that generates its own toolpath, feeding a Crane WASP concrete printer in a workshop in Sofia. That is a single-process setup within the additive extrusion family, not a full digital fabrication shop.

The work to date is process development rather than production: test walls of roughly one metre in height, printed to establish which parameter combinations are stable, together with material experiments on cements, sands, and admixtures available locally. There are no completed buildings, no delivered client projects, and no legal entity yet. The main practical lesson from that work is the one stated above — that the digital model does not capture material behaviour. Two prints from an identical file can differ because the mortar was mixed at a different temperature, and no amount of modelling resolves that. Physical trials and recorded parameters are part of the fabrication chain, not an optional check on it.