A hobbyist 3D printer that positions to 1 micron and deposits 4-micron resin droplets

The µRepRap project wants micron-scale fabrication to be as accessible as RepRap made desktop 3D printing, using common parts and a UV LED under the bed.

In brief

The current PIKA V0.02 build repeats positions to about 1 µm, with reliable accuracy of 10 µm or better inside an 8 mm cube. A probe tip draws UV resin up by capillary action and deposits droplets as small as 4 µm.

The earlier MAUS C build reached about 5 µm over 4 mm, but its motion wasn't linear.

What the wiki describes

The RepRapMicron wiki sets a long-term goal of a replicating machine at micron scale, built from accessible materials, with the same intent that RepRap had for desktop printing.

A three-axis positioner moves a probe tip that serves as a Z touch sensor, an engraving tool and a deposition head. Resin sits in a smeared reservoir; the probe draws it up by capillary action and deposits tiny droplets, and a UV LED in the print bed cures them. An optical-fibre probe is the alternative: it carries UV light down to the work, and has made 9 µm cured lines. Control is G-code over serial; the page considers Arduino GRBL adequate. Two USB microscopes help align probe height and XY, and a 100x to 200x bench microscope inspects the output.

Build Positioning Range
PIKA V0.02 ~1 µm repeatable, 10 µm or better accuracy ~8 mm cube
MAUS C V0.04 ~5 µm at best, non-linear motion 4 mm square

Steppers are 400-step NEMA17; 200-step motors come close to 1 µm. The wiki gives no cost figure.

The precision has cheap precedent; the replication goal does not

A commenter who also explained what the USB microscopes are for doubted the RepRap framing:

"But even if you could get a positioning system that small, I'm pretty sure you can't scale resin printing down three orders of magnitude smaller than this printer."

DannyBee, an ex-Google open-source lawyer by their own bio, pointed out that micron-level position feedback is already cheap:

"you can get etched glass scale encoders that will give you 1um accuracy for basically nothing these days (IE <50 bucks)."

tliltocatl, who laid out where piezo stacks and flexures get expensive, explained why the usual route to finer motion does not fit a build others must copy:

"Piezos are sure the way to go for a fab-in-a-shed youtuber, but not yet for those who want to make it reproducible by others."

So the 1 µm figure is plausible with hobby-priced parts, but the project's stated aim is reproducibility. Until the wiki publishes a cost and a second builder confirms the 8 mm cube, treat the hobbyist-accessible claim as open.

Get breakdowns like this every day

The top Hacker News stories, unpacked — by email, Telegram, or RSS.

This is the last story in the October 19, 2026 issue.

Go to the issue archive