For FDM printing, tune the fit of a threaded pair with a calibration series that separately applies a reduction to the external thread and an enlargement to the internal thread. Do not use the single term “clearance” interchangeably for a radial model adjustment, a surface offset along the normal, and the actual distance between the thread flanks: these are different geometric quantities.
Scope and limitations
This method applies to threads printed by thermoplastic extrusion. Calibration specimens must be produced using the same printer, nozzle, material, slicer profile, and orientation as the final part.
The calibration determines whether the parts can be assembled, how smoothly they rotate, and how much play they have. It does not verify the load-bearing capacity of the connection or its resistance to creep, wear, temperature, vibration, or cyclic loading. Load-bearing and safety-critical assemblies require a separate structural analysis and testing of the finished parts under their intended operating conditions.
Smooth engagement does not mean that the thread can withstand its operating load. Do not use fit-calibration results as evidence of the connection’s strength or service life.
Quantities that must be distinguished
Define two independent parameters for the external and internal parts:
- Δe — radial reduction of the external thread;
- Δi — radial enlargement of the internal thread.
If the external thread is reduced radially by Δe, its diameters decrease by 2 × Δe. If the internal thread is enlarged radially by Δi, its diameters increase by 2 × Δi.
The sum Δe + Δi describes the total radial compensation of the pair. It is not automatically equal to the shortest distance between the inclined thread flanks. The actual normal clearance depends on the profile angle, helical geometry, axial position of the parts, and the method used to modify the model.
Radial compensation must also not be confused with a surface offset along the normal. An offset operation applied to a helical surface may change crests, roots, and transitions differently from a radial translation of the original profile.
A verifiable method for modifying the geometry
The most predictable method is to modify the original thread profile before generating the helical surface. The axial section should simultaneously show:
- the nominal external-thread profile;
- the modified external-thread profile;
- the nominal internal-thread profile;
- the modified internal-thread profile;
- the part axis and the Δe and Δi dimensions.
External thread
Move the entire generating profile of the external thread radially toward the axis by Δe without changing its axial coordinates. Then generate the helical surface using the original pitch, handedness, and number of starts. This translation preserves the shape of the profile in the axial section: the flank angle, axial distances, and pitch remain unchanged.
Internal thread
Move the entire generating profile of the internal thread radially away from the axis by Δi. Use the same helical-generation parameters as in the nominal model.
What to verify after generating the thread
- The pitch and axial positions of the turns remain unchanged.
- All radial dimensions of the external profile have decreased by Δe.
- All radial dimensions of the internal profile have increased by Δi.
- The crests and roots have not disappeared or turned into thin, unprintable features.
- The aligned parts contain no intersecting solids.
- Free space remains between the mating surfaces in several sections along the thread length.
If the CAD system can generate threads only with a built-in command and does not allow control over the original profile, create separate parameters for the nominal external- and internal-thread diameters. Check the axial section after every change. Do not rely only on a parameter name or on the visual absence of an intersection.
Do not use global part scaling to tune a thread. It changes the pitch, engagement length, wall thickness, chamfers, and all other mating dimensions.
Record the calibration conditions
Before creating the series, record:
- the printer model and nozzle diameter;
- the material and the specific spool;
- the layer height and line width;
- the orientation of the thread axis;
- the external-wall speed;
- the nozzle temperature and cooling settings;
- the wall-printing order;
- the number of perimeters;
- the first-layer and horizontal-size compensation settings;
- whether supports are used;
- the intended fit and measurable acceptance criteria.
Do not transfer the result to a different nozzle, layer height, material, orientation, or slicer profile without verification.
Verify that the profile is printable
Compensation cannot correct a thread whose features cannot be reproduced by the selected process. Before calibration, inspect the profile layer by layer in the slicer.
- Pitch. The flanks should be formed by several consecutive layers and extrusion lines rather than isolated, discontinuous fragments.
- Profile height. The crests and roots must not disappear after slicing.
- Feature thickness. Walls around the internal thread must print without gaps and must not deform during hand tightening.
- Thread entry. Both parts need a chamfer or another controlled lead-in that prevents cross-threading during the first turn.
- Orientation. Horizontal and vertical threads have different limitations related to overhangs and layer structure, so they require separate tests.
How to choose the test-series range
There is no universal range in millimeters. Select it based on the printer’s repeatable dimensional error, line width, layer height, and the dimensions of the thread profile itself.
First, print one nominal pair without compensation. If the profile is reproduced but the connection cannot be assembled, define the series increment s. It should be large enough for adjacent variants to differ after printing, but not so large that a single increment significantly damages a profile crest or root.
For the first series, compensation can be distributed equally between the two parts:
| Variant | External thread Δe | Internal thread Δi | Sum Δe + Δi |
|---|---|---|---|
| A | 0 | 0 | 0 |
| B | 0.5s | 0.5s | s |
| C | s | s | 2s |
| D | 1.5s | 1.5s | 3s |
| E | 2s | 2s | 4s |
Mark both parameters on every pair, for example Δe=…; Δi=…. A single-number marking is unacceptable when it does not show how the compensation is distributed.
When only one part can be modified, use a separate series. For example, if the nut cannot be changed, set Δi=0 for every variant and vary only Δe. The result of this series must not be treated as equivalent to a series with equal compensation applied to both parts.
Eliminate first-layer effects
First-layer bulging can increase the external diameter of a screw or reduce the entrance of an internal thread. The connection may then jam at the end even though the remaining turns have sufficient clearance.
Add entry chamfers and check the first layer with a separate flat calibration specimen. Do not increase compensation across the entire thread until first-layer bulging, misaligned engagement, and edge defects have been ruled out.
If the pair engages over its full length after removing only the protruding entry edge, correct the first layer or lead-in rather than the entire profile.
Print the series with the production profile
Print all variants simultaneously or under comparable conditions. Do not change the orientation, layer height, number of perimeters, wall speed, temperature, cooling, or dimensional compensation between specimens.
After printing, remove supports, strings, and incidental protrusions without affecting the working thread surfaces. Do not sand or repeatedly run in the thread before the initial evaluation, because this treatment changes the fit.
Define measurable acceptance criteria
Terms such as “tight” and “loose” are insufficient for a reproducible selection. Define the project-specific parameters before testing and measure every variant in the same way.
| Parameter | How to test it | Project criterion |
|---|---|---|
| Start of engagement | Assemble without tools while keeping the axes aligned | No more than the specified number of failed attempts |
| Rotational torque | Use a torque-measuring tool or the same measurement fixture | Permitted torque range |
| Axial play | Measure with an indicator under a specified reversing axial force | Maximum permitted displacement |
| Radial play | Measure with an indicator under a specified transverse force | Maximum permitted displacement |
| Wear | Evaluate after a specified number of complete cycles | Permitted change in torque and play |
| Damage | Perform a visual inspection at the same magnification | No cracks, shearing, or crushing beyond the specified limit |
Numerical limits must be defined by the project. They cannot be declared universal for all diameters, materials, and applications.
How to interpret the results
| Observation | Likely cause | Action |
|---|---|---|
| The thread does not begin to engage | Defective lead-in, first-layer bulging, misalignment, or insufficient compensation | Check the ends and chamfers first, then increase Δe or Δi |
| Binding occurs once per revolution | Seam or localized wall defect | Compare the binding location with the seam position |
| Resistance increases along the full length | Insufficient total compensation or a geometric profile error | Inspect the section and move to the next variant |
| Low torque but excessive play | Excessive compensation or incomplete thread turns | Reduce Δe + Δi and inspect the layer-by-layer representation |
| The thread turns crush or delaminate | Insufficient strength, unsuitable orientation, or excessive load | Change the design, material, or orientation; fit adjustment alone will not solve the problem |
Refine the selected range
After the coarse series, identify adjacent variants between which the result changes: one fails the criteria and the next passes. Create intermediate values while keeping the same distribution between Δe and Δi.
Do not select a pair that assembled successfully only once. Select a variant that repeatedly passes the defined checks. Print several identical sets and evaluate the variation in torque, play, and damage. The number of repetitions is determined by the project requirements and the permitted process variability.
Slicer compensation settings
Horizontal contour compensation and first-layer compensation can affect threads, but their behavior depends on the software and version. These settings may simultaneously change holes, external dimensions, slots, and thin walls.
First, tune the printer’s systematic dimensional error using separate specimens. Then repeat the threaded calibration series. Do not treat hole compensation as equivalent to a radial profile translation without checking the layer-by-layer representation and measuring the printed parts.
What to record for future use
Printer: [модель]
Nozzle: [диаметр]
Material: [тип и катушка]
Layer height: [значение]
Line width: [значение]
Thread: [номинальный размер, шаг, длина]
Orientation: [описание]
External compensation Δe: [значение]
Internal compensation Δi: [значение]
First-layer settings: [значения]
Torque and play criteria: [значения проекта]
Number of test cycles: [значение]
Repeat-test results: [измерения]
Profile date and version: [данные]Final checklist
- The method is applied to FDM printing with the production slicer profile.
- Δe and Δi are defined separately.
- Radial compensation is not described as the actual normal clearance.
- The profile is modified in the axial section before helical generation.
- The pitch and axial coordinates of the turns are preserved.
- Every specimen is marked with its Δe and Δi values.
- The series range is selected according to the printing resolution and profile dimensions.
- The first layer and thread entry are checked separately.
- The fit is evaluated by torque, play, cycling, and damage.
- The selected variant is confirmed with repeated specimens.
- A load-bearing connection is supported by separate analysis and testing.
- Calibration is repeated after changing the material, orientation, or slicer profile.