How to Measure a Part for 3D Printing Before You Reproduce It

Start with what the part has to do

Before you grab your calipers, figure out what the part actually does. A knob, spacer, clip, or bracket rarely needs every surface copied to a tenth of a millimeter. What matters are the functional dimensions: the ones that touch another part, take a screw, locate a shaft, or carry a load. Everything else can be simplified, rounded off, or beefed up.

A Prusa blog post about modeling a replacement part in OpenSCAD is a good example. The plastic knob under an office chair had snapped, and nobody knew what the original looked like. So the author didn't try to build a one-to-one replica. The goal was a part that did the same job, printed easily, and held up. The design started from how the part is loaded: it works in compression under the screw head, and a metal washer on the chair side helps spread the force.

Do the same on paper. A rough sketch, nowhere near a real drawing, is plenty to record each dimension, what you measured it from, and how. That sketch becomes your reference once you open your CAD program.

Caliper technique that gives you numbers you can trust

A set of digital calipers handles most of the job. Use the outside jaws for diameters and thickness, the inside jaws for bores, and the depth rod for counterbores and steps. Close the jaws, zero them, then take each measurement several times in different spots. A worn, warped, or poorly molded part won't read the same everywhere.

When you can't reach a dimension directly, measure around it. In the Prusa example, other hardware blocked the gap between the underside of the chair and the bottom of the screw head. The author measured two reachable dimensions and subtracted one from the other, which gave a fairly accurate result. The post also records a range instead of a single number: the minimum distance with the screw tightened and the maximum with it almost backed out. The printed part has to land between the two. Too thin and it won't hold the screw head. Too thick and you can't tighten the screw.

Finally, know your tool's limits. The same post notes that the calipers were plastic, so there was some play. Steel calipers are more consistent, but no reading is perfect. Question every number.

Holes, screws, and wall thickness

With holes and standard hardware, identify the standard before you trust the reading. In the Prusa post, the screw measured 7.7 mm, but it's an M8, so the nominal diameter is 8 mm. Part of the difference comes from the threads, which always run a little under nominal. Wear and play in the calipers account for the rest. Model the standard value, not the raw reading.

For the hex head, the author measured the circumscribed circle, the one that passes through the corners, instead of measuring across the flats. It read 14.7 mm and was rounded to 15 mm, because OpenSCAD's cylinder function needs that diameter to generate a hexagon. Take each measurement in the form your software expects.

For wall and section thickness, separate hard limits from comfort. The section around the threads had to stay between 4 and 17 mm, and the author picked 10 mm. Around the head, the part needed at least 5 mm, but the author chose 30 mm for a better grip, with an overall diameter of 50 mm. Our guide to holes and fits in CAD covers the next steps.

Why no clearance value works on every printer

The size you measure on the original isn't always the size you should draw. If two parts need to slide, snap together, or rotate, you need clearance. That clearance depends on your printer, nozzle, filament, speed, and print orientation. It can even differ from one hole or pin to the next. A vertical hole and a horizontal hole rarely come out the same.

That's why no clearance value works on every machine. Be skeptical of one-size-fits-all numbers from forums or videos, because they describe someone else's setup. The reliable approach is a test coupon with a series of stepped clearances. Print it with the same filament and slicer profile you'll use for the real part. Then note which step gives the fit you want: a press fit, a sliding fit, or free rotation.

The assembly clearance calculator turns those results into a design dimension. Keep the results with your slicer profile, and print the coupon again whenever you change filament or nozzle.

Mesh measurements vs. real-world measurements

Once you've exported the model, check it in your slicer. According to the Prusa Knowledge Base, the PrusaSlicer measurement tool sits in the left toolbar. It detects vertices, edges, circles, and planes, and it shows distances and angles right on the model. It measures between features of a single object, not between two separate objects.

Be clear about what that number means. The slicer reads the digital mesh. It confirms that the file is the right size and was exported in the right units. It tells you nothing about the part that comes off the build plate, with its shrinkage, elephant's foot, or holes that print slightly small. A dimension that looks perfect on screen can still produce a part that's out of spec.

The documentation also lists some limitations. On objects with overlapping parts, the tool misses edges created by the intersections but detects planes hidden inside other parts. It also treats negative volumes as parts. You can scale an object uniformly by clicking a measured distance and setting the length you need. That quickly fixes a file exported in the wrong units, but it doesn't replace fixing one specific hole.

Print a test piece before the full part

Before committing to the full print, print just the critical zones: a thin slice that includes the hole, the screw-head pocket, or the mating feature. It prints quickly, uses little filament, and answers the only question that matters at this stage: does it fit?

Measure that test piece with the same calipers and compare three numbers: the original part, the CAD model, and the print. If every outside dimension is off in the same direction, the cause is probably the printer, so check the dimensional accuracy troubleshooting guide. If only certain holes are tight, fix the model at those spots instead.

Then test-fit the part without forcing it. Check how it tightens, how it moves, and whether the holes line up. Change one dimension at a time, label the revision, and reprint. Short loops like this keep you from wasting long prints. Finally, a printed replacement comes with no guarantee if a failure could hurt someone. For that kind of part, use the manufacturer's original.

Sources and limits

• Prusa — Practical problem solving in OpenSCAD
• PrusaSlicer — Measurement tool

We organize manufacturer and community documentation; we have not measured these products ourselves. The manual for your exact model takes precedence over general guidance.