Tolerances for 3D Printed Moving Parts: A 2026 Fit Guide

Most moving joints in a printed assembly want somewhere between 0.2 mm and 0.5 mm of gap on FDM, and 0.1 mm is often enough in resin. There is no single right number, because the gap that works on a well-tuned printer in PLA will bind on a different one in ABS. The honest way to pick tolerances for 3D printed moving parts is to start from a clearance that matches your function, then prove it with a small printed coupon and a caliper.

The reason generic dimensional tolerances disappoint people is that a tolerance is written per feature, while a printed part moves as a whole. A hole and a shaft on the same part can each sit inside their own tolerance and still jam, because layer lines, shrinkage and a hole that always comes out undersized all push the same direction.

What Tolerances Are Right for 3D Printed Moving Parts?

Tolerances for 3D printed moving parts depend on printer accuracy, material shrinkage, feature size, what the joint has to do, and how much freedom of movement you want. Functional clearances beat fixed target dimensions, and you should confirm them with calibrated measurements rather than trusting the slicer preview.

Five things move that number, and it is worth knowing which of them you can control:

• Printer class. A well-maintained bed-slinger and a high-speed CoreXY machine with pressure advance and input shaping will not hold the same hole size from one Tuesday to the next.
• Material. PLA and PETG barely move. ABS, ASA and nylon shrink enough to close a tight fit on their own.
• Feature size and orientation. Vertical holes print smaller than horizontal ones. Small holes drift more, because a fixed error is a larger percentage of a 3 mm bore.
• What the joint does. A hinge that swings freely needs more room than a shaft carrying a light load through a bearing.
• Environment. A car dashboard in summer is a different design problem than a mechanism sitting on a desk.

If you take one habit from this guide, make it printing a coupon before the real part. Twenty minutes of test print saves a twenty-hour mistake.

Tolerance vs. Clearance vs. Fit: What Is the Difference?

These three words get used interchangeably, and they are not the same. Getting them straight makes every later number easier to reason about.

Tolerance is the permitted amount of variation on a single dimension. A 5.00 mm shaft specified as 5.00 plus or minus 0.05 mm has a bilateral tolerance; the 0.10 mm band between 4.95 and 5.05 is the tolerance, not the shaft.

Clearance is the actual gap that results when a hole and a shaft are assembled, measured as the difference between them. Hole minus shaft equals clearance. Positive is a clearance fit, zero is line-to-line, negative is an interference fit.

Fit is the condition of the assembled pair, classified three ways:

• Clearance fit — the hole is always larger. Free movement, no force needed.
• Transition fit — sometimes clear, sometimes interfering. The worst option for a prototype, because it is unpredictable.
• Interference fit — the shaft is always larger. Pressed, hammered or thermally assembled. Real, but hard on plastic.

A datum is the feature everything else is measured from, usually a flat face or a shaft axis. It matters in printing because a warped flat surface shifts every hole relative to it, so the two mating parts stop sharing a reference.

A hole-and-shaft example in real numbers

You model a 5 mm shaft and a 5.2 mm hole, a nominal 0.2 mm clearance. On a typical FDM print the hole comes out at 4.95 mm after the STL polygon approximation and the shrink that comes with a vertical bore. Real clearance is 4.95 minus 5.00, which is negative. The joint is an interference fit, and it will not turn.

That single conversion from CAD to plastic is why designers who copy numbers from a spec sheet keep getting stuck parts. Print the pair, measure it, and only then decide what to change.

Which Clearance Do Common Moving Parts Need?

These are starting points for a well-calibrated FDM machine, not engineering guarantees. Treat the left column as design intent and the right column as the number to test first.

Joint	Design intent	FDM starting clearance	Resin starting clearance
Shaft in a bore (light load)	Free rotation, no rattle	0.3 – 0.4 mm	0.10 – 0.15 mm
Rotating ring or collar	Even rotation with visible gap	0.4 – 0.5 mm	0.15 – 0.25 mm
Sliding piece in a guide rail	Slide by hand, no bind	0.4 – 0.6 mm	0.15 – 0.25 mm
Print-in-place hinge	Repeated flexing, stays intact	0.4 – 0.6 mm	0.20 – 0.30 mm
Snap fit (cantilever)	Audible click, holds when pulled	0.3 – 0.4 mm at the root	0.10 – 0.20 mm
Lever arm on a pin	Rocking motion, no side play	0.2 – 0.3 mm	0.05 – 0.10 mm
Bearing bore (press fit)	Grip by interference, square shoulder	-0.05 to -0.15 mm	-0.02 to -0.05 mm
Dovetail or tab joint	Slide together, adjustable	0.3 – 0.5 mm	0.15 – 0.20 mm

Notice how the numbers climb as the required motion gets looser. A press fit is the only entry with a negative value, and it is also the one most often ruined by shrinkage, because a bearing bore depends on plastic gripping the outer race rather than on a gap.

These numbers are the consensus in print-in-place tolerance discussions, and they line up with what people report after months of iterating on FDM mechanisms: 0.2 mm for tight fits, 0.3 to 0.4 mm for loose and sliding joints, and resin users landing happily in the 0.05 to 0.1 mm band.

How Do Materials and Printing Methods Change the Fit?

The process decides how repeatable your hole sizes are. The material then decides how far everything moves after it cools.

Process	Typical dimensional accuracy	What it does to fits	Finish
FDM	Around plus or minus 0.2 mm	Holes print undersized; layer seams add burrs; elephants foot affects the first layers	Rough, benefits from deburring
Resin (SLA/MSLA)	Around plus or minus 0.05 mm	Holes shrink slightly and need a wash and cure cycle; distortion around heavy supports	Smooth, sometimes glassy and brittle
SLS / MJF	Around plus or minus 0.15 mm	No support contact, so no support marks; powder gives a soft internal surface that wears in	Grainy, self-supporting
Powder metal (DMLS/SLM)	Around plus or minus 0.05 mm	Shrinkage is significant and must be compensated in the slicer; holes print close to nominal	Very fine, needs support removal
PolyJet	Around plus or minus 0.1 mm	Accuracy depends heavily on which polymer is used; glossy surfaces are slippery for moving contact	Excellent detail
TPU (FDM)	Around plus or minus 0.3 mm	Flexible and squishy, so clearances need more room and shrink noticeably on the bed	Rubbery, no detail

Material shrinkage is the second half of the problem. Scale your model before printing and a tight fit becomes correct on paper but loose in the hand, while press fits get worse in the opposite direction.

Material	Shrinkage behaviour	Practical response
PLA	Very stable, minimal shrink	Design as modelled
PETG	Small, predictable shrink	Design as modelled, add a little extra gap for long spans
ABS / ASA	Notably larger, temperature dependent	Scale up around 1 to 2 percent, or enlarge holes specifically
Nylon (PA)	High moisture absorption, post-print growth	Measure the dry spool, expect change with humidity
TPU	Soft, compresses under load	Design for a squeeze fit, not a precision one
Resin	Small, cure-related	Fully cure before measuring or testing fit

One warning about scaling: print a coupon before you apply a global scale factor to a 200-hour model. Scaling everything changes every fit in the assembly at once, including the ones that already worked.

How to Set Tolerances for 3D Printed Moving Parts by Function

Rather than picking a number and hoping, work backwards from what the joint has to survive. Start with the question: how much slop can you live with? A display model rotating next to a camera slider needs precision. A shop door hinge modelled in plastic needs almost none.

Then adjust the base clearance for the conditions:

• Load. More force on the joint means the material flexes more under use. Add clearance so a flexing part does not grip a stationary one.
• Speed. A fast-spinning shaft wobbles more visibly, so bias toward the upper end of the range. A slow pivot tolerates a tighter fit.
• Temperature. Warm plastic is softer and creeps. A joint that works on a cool bench can bind on a warm shelf, so leave more room if the mechanism will live in a car or an enclosure.
• Stiffness of the material. Filled and fibre-reinforced filaments print stiff but shrink more. A stiffer shaft in a softer hole needs slightly more clearance, not less.
• How many cycles. Print-in-place hinges wear at the contact surface. Start 0.1 mm wider than feels right if the hinge has to survive thousands of swings.

What changes the number from one joint to the next

On a single mechanism the joints are rarely the same. The pivot pin carrying the main arm can sit at 0.25 mm because load is high and a little play is invisible. The same pin carrying a thin linkage arm gets 0.4 mm so the mechanism still articulates after a few hundred operations. The small decorative gear on the same plate can run at 0.2 mm and nobody would ever notice.

This is the part most reference material skips: tolerances are not a property of the part, they are a property of the joint and its job.

How to Account for Shrinkage and Print Orientation

Orientation quietly ruins more fits than calibration does. A hole printed lying flat on the bed keeps a circular cross-section within one layer. A hole printed vertically gets the polygon approximation of that circle built layer by layer, and the corners that the filament cannot reach round off inward, so the finished bore is smaller than nominal.

Several effects stack up in the same direction:

• Elephant foot. The first few layers bulge outward from the pressure of the nozzle, which eats clearance at the bottom of a vertical hole and can make a part feel tight on the bed but loose once removed.
• Layer seams and stringing. A seam crossing a bore leaves a small internal burr. Two such burrs on opposite sides of a hole can close a 0.2 mm fit completely.
• Rounding. Small radii, slot ends and any feature under about 1 mm are where a printer deviates most from the model.
• Warp. A part that lifts off the bed changes shape enough to shift hole positions relative to each other.

The horizontal expansion setting that fixes most holes

Every mainstream slicer exposes the same idea under a different name: Cura calls it Horizontal Expansion, PrusaSlicer and Orca call it XY Size Compensation, Bambu Studio uses XY Size Compensation as well, and Simplify3D calls it Horizontal Expansion. A negative value widens holes and outer walls; a positive value narrows them.

Set it to minus 0.05 or minus 0.1 mm, print a coupon with a set of holes, and measure. A shift of minus 0.1 mm fixes a surprising share of undersized-hole complaints without touching the model at all.

The measurement loop itself takes a few minutes. Print a coupon at the same orientation and with the same layer height and material as the real part, measure the holes with a caliper, note the deviation, and apply the opposite compensation in your slicer or in the CAD dimension. Repeat once. Do this before every long print, not after.

How to Test-Fit and Tune an Assembly

Print the joint, not the assembly. The fastest path to correct tolerances for 3D printed moving parts runs through a small coupon that contains the critical interface at full size.

1. Build the coupon. Include the real hole, the real shaft, a range of clearances if you are searching (0.2, 0.3, 0.4 and 0.5 mm) and any curved or post-processed feature.
2. Print it in the real orientation. A coupon on the bed tells you about bed-printed holes only. Test the vertical orientation separately if the real part uses it.
3. Inspect before measuring. Look for stringing, seam blobs, delamination and lifted corners. A part with a visible defect will mislead you about dimensional error.
4. Measure with the right tool. Calipers for outside dimensions, small bore gauges or pass pins for holes. Take three readings and average them.
5. Assemble without force. If it needs a tool to go together, stop. Never press a printed joint together to make it fit.
6. Run the motion test. A rotating joint should spin through a full revolution with a light drag. A sliding piece should move by gravity on a slight tilt.
7. Adjust one variable and reprint the coupon. Change the slicer compensation or the model dimension, not both at once, or you will not know which one fixed it.

Define pass and fail before you start. A rotating joint passes if it turns through 360 degrees with no dead spot and no side load; fails if it stops anywhere in the rotation. A sliding piece passes if it travels its full stroke on a 15 degree tilt; fails if it creeps under its own weight on a flat surface. A snap fit passes if it holds with a firm pull and releases with a thumbnail; fails if it goes on without resistance or will not come off without heat.

Record the result. A note that says “0.3 mm hole gave 0.22 mm actual at 0.2 mm layer height in PETG” is worth more than any table on the internet, because it is true on your machine.

What Measurements and Tools Give the Best Results?

Each tool answers a different question. Knowing which one reaches for a specific measurement saves a lot of guessing.

Tool	Good at	Limitation
Digital caliper	Outside dimensions, wall thickness, shaft diameters, quick checks on holes with an internal jaw	Resolution and accuracy vary; cheap units drift and cannot resolve 0.05 mm reliably
Outside micrometer	Shaft and pin diameters, where you need 0.01 mm resolution	Cannot measure a hole; contact pressure can push into soft plastic
Bore gauge	Inside diameters directly, including small holes	Expensive, and reading the dial on a soft printed surface needs a light touch
Feeler gauge	Gap widths and slot clearances where the gap is open on one side	Only works on accessible gaps, and the blade thickness itself is the limitation
Pass / fail pin gauges	Binary confirmation that a hole is above or below your limit, quickly and repeatedly	Tells you pass or fail, never how far off you are

Measuring the actual printed part matters far more than trusting the slicer or the CAD file. The slicer reports what it was told, not what came out, and a CAD dimension is a request rather than a result.

Two habits help. Take measurements after the part has cooled fully and, for resin, after a complete cure, because both move slightly as they settle. And measure at several heights and around several positions, since a hole that is square in the CAD can come out slightly oval or tapered.

How to Fix a Too-Tight or Too-Loose Printed Part

Match the symptom to the fix rather than reaching for the same global scale every time. Most problems have one cause and one clean answer.

Symptom	Likely cause	What to change
Hole smaller than modelled	Polygon rounding, thermal shrink, undersized extrusion	Set horizontal expansion to minus 0.05 or minus 0.1 mm
Part binds on the build plate only	Elephant foot at the base	Add a small chamfer, lift the part in the slicer, or increase first-layer cooling
Rotating joint grinds rather than turns	Seam burr inside the bore	Move the seam, or deburr with a fine file or scraper
Holes vary between prints	Flow, temperature or tension inconsistency	Run a flow test and a temperature tower, then re-check the first layer
Snap fit will not release	Root radius too tight, or material too stiff	Increase the root radius and add 0.1 mm clearance, or try a softer filament
Press fit loose after cooling	Material shrank during cooling	Model a deeper bore, add ribs, or switch to a low-shrink material
Layer split along the joint	Not enough layer adhesion for the load	Raise the layer height slightly, increase part wall count, or reorient the part
Everything is loose but consistent	Over-corrected slicer compensation	Reduce the negative expansion value, or scale the affected features only

Two warnings. First, never assemble a tight printed joint with force. Plastic that has been forced into a hole will fail later at a spot with no visible damage, and printed parts carry almost no safety factor. Second, resist reaching for a global scale. If one interface is wrong, change that one interface.

And when reaming or drilling is the right answer, treat it as a deliberate design step rather than a rescue. Drilling out a tight hole to a target diameter works well, and reaming with a proper reamer gives a genuinely accurate bore, but a twist drill will wander and leave a burr. A pilot hole, a light cut and a final pass get you a cleaner result than fighting the drill.

Frequently Asked Questions

Conclusion

Start with a functional clearance rather than a number from a spec sheet, print a coupon that contains the critical interface at full size, measure it with a caliper or a bore gauge, and change one variable before you commit to the full assembly. Whether you are tuning tolerances for 3D printed moving parts on a printer you built yourself or tuning them for 2026 production runs, that loop is what separates a mechanism that moves from one that prints solid.