Origin Two DLP 3D Printer: How the Process Works

• 1
  What the Origin Two DLP 3D printer actually does
• 2
  Why layer-by-layer curing sets the tolerance ceiling
• 3
  Resin families and what they are good for
• 4
  Geometry that suits the process, and geometry that does not
• 5
  Where the process pays off, and where it does not
• 6
  Origin Two DLP 3D printer vs CNC machining

DLP stands for digital light processing. A projector throws a full 2D image of the layer onto a resin vat, and every pixel in that image cures at the same time. The build plate lifts, fresh resin flows under the part, and the next image fires. One layer is one flash, not one nozzle pass.

That single-flash behavior is why DLP holds build speed as parts get wider. A laser-based process has to trace every contour, so a plate packed with small parts takes far longer. On the Origin Two DLP 3D printer, a full plate and a single part take roughly the same time per layer.

The second half of the system is post-processing. Uncured resin stays tacky and dimensionally unstable, so the part goes into a cure unit that floods it with UV at controlled temperature. Only after that step does the material reach its final strength and stiffness.

So the machine is really two stages: a projector that defines the geometry, and a cure station that defines the properties. Skip or shorten the second stage and your measured numbers will not repeat from batch to batch.

Every layer is cured against the previous one, so the bond between layers is chemical, not mechanical. That is good for strength in the build direction and bad for surface finish on sloped faces. Stair-stepping is a function of layer thickness, not of the projector resolution alone.

At 50 μm layers, a 45° face shows visible steps under raking light. Drop to 25 μm and the steps shrink, but build time roughly doubles and the resin has less room to flow. Most production parts land between 50 μm and 100 μm for that reason.

XY resolution is a separate number. The projector pixel pitch sets the smallest feature you can hold, and it is fixed across the plate. A small part and a large part get the same pixel size, which is the opposite of how a laser spot behaves.

In-plane dimensions repeat well, often within a few tens of microns on a well-calibrated machine. The Z direction is where variation creeps in, because it depends on the peel force and the resin viscosity at each layer.

DLP resins are formulated for a target property, not for a metal substitution. Rigid grades mimic ABS or PC stiffness. Tough grades take impact. High-temperature grades hold shape in an oven or under a hood. Clear grades are used for light guides and fluid paths.

Each family has its own cure window. Under-cure leaves the part soft and dimensionally drifting. Over-cure makes it brittle and can yellow a clear grade. The cure unit schedule matters as much as the print schedule, and both belong in your process sheet.

Moisture is the quiet problem. Most DLP resins absorb water from the air, and a wet part will shift after printing. Parts that sit in a humid room for a week can move more than the print tolerance you were chasing.

For anything that sees sustained load, heat above 80 °C, or repeated sterilization, we treat the resin route as a prototype path and quote a machined equivalent in aluminum, stainless, or PEEK.

DLP likes small, detailed, thin-walled parts in one orientation. Lattice structures, internal channels, and fine text come out clean because there is no tool to reach into the feature. Those are the cases where the process beats milling outright.

It dislikes thick solid sections. A 15 mm block of solid resin cures unevenly and can warp as it cools and shrinks. If you need a thick body, design it as a shell with ribs and let the cure unit do its job.

Drain paths matter. Trapped resin inside a blind cavity will cure into a hard plug that is hard to remove. Add a drain hole at the lowest point and plan the build orientation so liquid can escape.

Threads are another boundary. Printed threads under M6 are unreliable in service. Design a pilot hole and cut the thread with a tap, or use a heat-set insert. This is where a printed part and a machined part often get combined.

The economics are simple. No tooling, no setup charge, and the cost per part is nearly flat from one piece to a few hundred. That is the window where injection molding cannot compete because the mold is not paid off yet.

Once you pass a few thousand pieces in a stable design, molding wins on unit cost. DLP does not. The resin price per kilogram does not fall with volume the way pellet prices do, and the cure step is labor.

The other hidden cost is inspection. A printed part with internal channels cannot be checked with a caliper. You need CT or sectioning, and that changes the quality plan. Machined parts have the same problem, but you can often measure the feature directly.

We quote both routes side by side when a design is near the crossover. Sometimes the answer is a printed prototype for the first 50 units and a machined or cast part after that. That is a normal split, not a compromise.