Digitizing real objects with a contact 3D scanner

I couldn't attend the live onboarding session, so I went through it independently using the interactive walkthrough and Roland scanning-tips guide, then did the hands-on work in the lab. The Modela MDX-20 is normally a small milling machine; swapping the spindle for a probe head (two screws) turns it into a contact 3D scanner. The bed and coordinate system stay the same.

Instead of a laser, it uses a piezoelectric probe that senses sideways bending as it sweeps across a surface, so it "feels" edges approaching rather than stabbing straight down. Because the head only travels up and down, overhangs and undercuts cannot be captured — the scanner sees the object the way rain would fall on it.

The trade-off that matters: pitch vs. time

The scan pitch (spacing between sample points) decides everything. A coarse pitch is fast but blocky; a fine pitch is detailed but slow, and can grow to many hours. The working method is to run a coarse draft first, then rescan only the regions that deserve detail at a finer pitch.

Exploring the machine with Ahmed Ibrahem and Fatma Samir, we looked for something to scan and — as it turned out — picked badly: a soft silicone part. We didn't yet know how much the material would fight the probe. We fixed it to the bed with double-sided tape and used it to learn the Dr.PICZA interface end to end.

1. Mounted the silicone part flat on the bed with double-sided tape.
2. Opened Dr.PICZA and defined the area of interest, entering start and end points for X and Y in inches.
3. Set the scan parameters (pitch) and the highest Z point in the design.
4. Ran a draft scan to see the result.

Even as a failure, this run was the best possible UI tutorial: it's where I actually learned how to set coordinates, bound the scan area, define Z, and read the parameter panel.

When my turn came, I chose a micro:bit V2 — rigid, flat, and full of interesting surface detail. I scanned the front face, which carries the two buttons, the LED matrix, and the crocodile-clip holes along the bottom edge.

1. Mounted the board flat, front face up, on the bed.
2. Measured and entered the board dimensions and selected the area of interest.
3. Set the maximum Z so the probe cleared the tallest components.
4. Ran a draft scan at ~1 mm pitch to capture the overall shape quickly.

Dr.PICZA displays pitch in inches. The draft screenshots show roughly 0.038 in / 0.028 in — about 1 mm.

The draft captured the board's outline but blurred the parts that matter. Rather than rescan the whole board at a fine pitch (slow), I selected just the high-value regions — the LED matrix, the two buttons, and the crocodile-clip holes — and rescanned those at a finer pitch, dropping X/Y from 1 mm to 0.5 mm. Dr.PICZA layers the new fine mesh over the existing draft.

1. Material choice comes first. A contact probe needs a rigid object; the soft silicone deformed under the stylus and made the scan unusable.
2. Fixturing matters. Double-sided tape holds the part flat and still — any movement corrupts the mesh.
3. Set Z before scanning. Knowing the tallest point lets you set a safe Z upper limit so the probe clears components.
4. Draft coarse, then rescan. A ~1 mm pass maps the shape in minutes; a targeted 0.5 mm rescan adds detail only where needed.
5. Pitch is a time budget. Fine pitch across a whole object costs hours — spend it selectively.
6. No overhangs. The head only moves vertically, so undercuts won't be captured — orient the object accordingly.