Holes and fits: designing a calibration coupon

What a coupon measures, and what it does not

A fit coupon is a small part drawn to answer one question: how much clearance you need, on your printer and your material, for a hole to accept a given shaft. According to Prusa's documentation, there is no single universal tolerance value. Model size, horizontal or vertical orientation, the geometry of the interlocking parts, calibration, settings, and material properties all play a role. The coupon exists to isolate those factors one at a time, not to produce a number that works everywhere.

That makes it different from general dimensional accuracy troubleshooting, which hunts for the cause of a part that came out wrong, and from a simple scaling correction applied to a whole model. Here you design a deliberately redundant object, measure it under a fixed protocol, and record the result together with its conditions. Prusa's documentation states that an Original Prusa is accurate to at least 0.2 mm, while reminding you that different materials can warp or shrink during printing. That figure describes that manufacturer's machines in their own context. It does not replace measuring your own coupon, and it is not a certificate for your printer.

Choosing the geometry: holes, shafts, and orientations

Start with a simple plate and draw a row of vertical through-holes whose nominal diameters step up around the size you care about. The following dimensions are illustrative: a nominal 6 mm shaft and nominal holes of 6.0, 6.1, 6.2, 6.3, and 6.4 mm. Nominal diametral clearance is the CAD hole diameter minus the CAD shaft diameter; nominal radial clearance per side is half of that. Prusa also gives at least 0.3 mm for movable parts, without defining a radial or diametral convention in that passage. Do not treat that reference as the definition of the diametral clearance chosen for our example.

Add a second row of horizontal holes, cut into a vertical wall of the same plate, because Prusa lists orientation among the factors that change the result. Include one or two printed shafts, one standing and one lying down, to measure before any fit test. In your CAD software, check how many faces each circle actually has: Prusa's documentation warns that a curve smoothed on screen by shaders can export and print as a polygon. Also inspect mesh integrity and the slicer result; successful slicing alone does not verify every dimension.

Labeling nominal dimensions without ambiguity

A coupon is only as useful as its labels. Give every hole a short identifier, embossed or engraved next to it, and keep a separate table that maps that identifier to the CAD nominal, the orientation, and the face you measure. Always separate three ideas. The nominal dimension is the number typed into CAD. The measured dimension is what your instrument reads on the printed part. The clearance is the difference between a hole and a shaft, and it must be tagged as diametral or radial to avoid misunderstandings.

Very fine engraving may not print at all: Prusa's documentation notes that in its MK3S reference with a 0.4 mm nozzle, the PrusaSlicer extrusion width is 0.45 mm, and that walls thinner than one perimeter cannot be printed. Size your markers accordingly, or replace letters with a count of dots or notches. Record the file version too. Prusa recommends a parametric approach and saving versions along the way before merging bodies or cutting cavities that must fit something else. Every coupon should point back to one exact version of the CAD project, otherwise you will not know which nominal was actually printed.

The measurement plan: instrument, points, and repeats

The measurement plan should say what you record and where, so you can repeat the procedure. Choose an instrument suited to the hole or shaft size and accessibility. Record its resolution, the height actually accessible, and how it is applied; displayed resolution does not certify measurement accuracy. For a hole or shaft, you can record two perpendicular directions at a defined height, such as mid-thickness, then write each reading down without rounding toward the nominal dimension. This describes checks to perform; no coupon measurements are reported here.

Add a qualitative fit check with terms set in advance: does not enter, enters with force, enters and turns, floats. These notes describe an observation, not a statistic. For comparisons between coupons, keep orientation, feature height, material, nozzle, profile, and post-processing constant. If you use a caliper, record the instrument, its resolution, and the same measurement points; the Prusa documentation cited here does not prescribe a particular metrology instrument or certify your measurement’s precision.

First layer and elephant foot: what the slicer can correct

The first layer can affect dimensions at the bed. Prusa’s documentation says it is squished against the heated bed and usually comes out a bit wider than intended, which can interfere with tight fits. The Elephant foot compensation page places the setting under Print Settings > Advanced and says it requires Advanced or Expert mode. It says values around 0.2 mm usually work for the default 0.4 mm nozzle, official Prusa profiles turn it on by default, and PrusaSlicer 2.2 and newer detect thin first-layer lines to avoid over-shrinking them.

The documentation names two effects: the brim connects less firmly to the model, and the preview may look distorted where thin features meet larger bodies, while real squish fills part of the gap. For the coupon, record dimensions separately at the bed and at the specified height. A difference may be a reason to inspect the compensation setting, but it does not by itself establish the cause of a dimensional error. The setting affects the first layer, not every dimensional issue; see the elephant foot page.

Reading the sheet and iterating on the design

After measuring and checking the fit, distinguish nominal CAD dimensions from measured diameters. Measured diametral clearance for a pair is the measured hole diameter minus the measured shaft diameter; the difference between their CAD dimensions is the nominal clearance drawn. If a pair gives the fit you want, record both clearances and the coupon conditions. You can use the nominal clearance as a starting point for a comparable part, then check that representative geometry before treating it as validated.

If results differ by orientation, record them separately instead of averaging: Prusa lists orientation among the factors affecting fits. Materials can warp or shrink differently; Prusa’s guide also says PC, ASA, or PETG may reproduce small details less well than PLA. Treat a measured clearance as specific to the recorded conditions and check again before applying it to another material, profile, or printer pairing.

Sources and limits

• Prusa — Modeling with 3D printing in mind
• PrusaSlicer — Elephant foot compensation, page 2.9 (Legacy)

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