Hole, Bore, and Hex Pocket Calibration Models

Every printer and machine prints or cuts holes and pockets a little off - a 5 mm hole rarely comes out 5 mm, and the error usually isn't the same for a round hole as it is for a flat-to-flat dimension like a nut pocket.

These are the calibration models for FitCheck: Hole Sizes & Fits, an app I develop and keep improving. You print (or machine) one part, measure it, and the app corrects the clearance and pocket sizes for how your machine actually operates - so you can bake the right offset into a design, a slicer, or your CAM.

The Parts

Two parts. Pick the one for your machine.

• FDM/SLA Part

  • Three 5.00 mm holes - the only measurement that is required

  • One 22.00 mm bore

  • Two 8.00 mm hex pockets

• CNC/lathe Part

  • Three 8.00 mm holes

  • No hex pocket and no big bore, on purpose - an endmill's error doesn't care about contour or diameter.

Every feature has its size engraved on the part, and every hole and pocket has 0.5 mm chamfers top and bottom. That leaves a clean parallel band in the middle to measure, and strips the two zones that read false on calipers: the elephant's-foot squish at the bottom and the bulge at the rim.

Calibration

How to make it

FDM

• Slice with the same profile, nozzle and material you use for real parts - the offset only transfers to your work if this prints like your work does

• Flat on the bed, holes vertical, no supports, 0.2 mm layers. Prints clean on 0.4 and 0.5 mm nozzles.

• Leave slicer XY size compensation OFF (your slicer may call it hole horizontal expansion)

  • The whole point is to measure what your machine does uncompensated. Measure with it on and you're measuring your slicer's patch, and the app will then correct on top of it.

CNC/lathe

• Cut the 8.00 holes with the same tooling and process you use for real parts (drill + ream, or bore/interpolate)

How to measure

Use the caliper's inside jaws and aim for the middle of the bore - between the 2 chamfers. Light, consistent pressure - leaning on the jaws deforms the wall and a slight tilt reads high, so lean on the averaging rather than any single reading.

Let it cool to room temperature first, on the plate. An ABS 22 bore measured while warm off a 60 °C bed reads about 0.07 mm larger than it will at room temperature - bigger than the error you're trying to isolate, and it impacts the largest bore where it matters most. PLA moves less, but the habit costs nothing.

FDM

• Round holes (required)

  • Measure each of the three 5.00 holes twice, rotating the caliper 90° between readings. Average all six

• Hex pockets (optional)

  • Each hex has three flat-to-flat pairs. Measure all three on both pockets and average the six

  • Leave it blank and the app corrects hex pockets with the round-hole number instead

  • If your flats measure dead-on 8.00, enter 8.00 - that records "flats print true," which is different from leaving it blank

• Large bore (optional)

  • Measure the 22.00 bore three times, rotating roughly 60° between readings, and average

  • If your three readings spread more than 0.05 mm, don't use them - reprint. That much spread is a defect on that bore, like a seam blob or a lifted corner, not a machine characteristic.

CNC

• Measure each of the three 8.00 holes twice, rotating the caliper 90° between readings

• Average all six

Take every number from the same part. Mixing a 5 mm reading from an old print with a 22 mm reading from a new one folds the difference between the two prints into the correction - a spool change alone can move flow enough to matter.

What to do with the numbers

Enter them into FitCheck (free app for iOS and Android - more info here), and the clearance and pocket sizes it shows are corrected for your machine. Calibration is free, and saved machines stick around.

No app?

The numbers are yours to use. measured - commanded on the 5 mm holes is your printer's hole-size error, usable directly as slicer XY hole compensation. The 22 mm number tells you how much less of it a big bore needs, and the hex number tells you how much of it not to apply to straight-walled pockets.

Prefer to just look a number up?

The whole reference is free on the web, no app needed:

• Clearance hole sizes - metric and imperial

• Heat-set insert hole sizes

• Bearing pocket sizes

• Tap drill sizes - metric and imperial

• Printed holes that fit - design patterns that don't need calibration

Why these measurements

I moved to two separate measurements because on my printers, flat-to-flat accuracy and hole accuracy differed - so a correction found by averaging the round holes actually made flat-to-flat parts worse, like nut capture pockets.

The reason is that the two features share a common tuning component and round holes add a second. Flow sets your wall width, and that shifts round holes and flats alike. Round holes then pick up an extra curvature component that flat walls never see. So a round-derived number is always too big for a pocket, and applying it over-opens the pocket. How big either component is depends on your machine and your material.

The 22 mm bore is there for a related reason: hole error isn't constant across sizes. On my printers, a correction from the 5 mm holes alone over-corrected larger bores, so measuring a small and a large hole lets the app scale the correction with diameter.

Re-measure when your process changes. These numbers describe one machine with one nozzle, one material and one profile. Change any of them - including turning slicer compensation on or off - and print the part again. Naming your machines accordingly helps: "MK4 · 0.4 · PLA" beats "MK4".

Want to see what your machine does across the whole range instead of at two sizes? The hole error test plate has six holes from 5 to 22 mm.

Happy printing!

Like what I do and want to support me? Buy me a coffee!