Annealing 3D Prints: Measure Dimensional Change by Axis

Measure annealing change by feature and axis; do not apply one universal shrink percentage. The result depends on the exact polymer, print history, geometry, support during heating, thermal cycle and cooling. A strength improvement can arrive with a fit problem.

Use the material maker's supported process. Annealing is not a way to turn an arbitrary printed object into a heat-rated or otherwise qualified component.

What published testing supports

A 2023 Materials study varied layer height and annealing time and temperature for PLA, PETG and carbon-filled PETG. It reported material- and axis-dependent dimensional behavior, with the tested filled PETG showing smaller changes than the other tested materials. Those findings do not provide a universal correction for every spool or geometry.

Prusa's annealing investigation also examines the tradeoff between thermal treatment and dimensional stability. Read the specimen and process alongside any claimed benefit, rather than copying a temperature as a general recipe.

Define the acceptance features first

  • Outside lengths, width and height.
  • Holes, slots, mating faces and fastener locations.
  • Flatness, warping and assembly clearance.
  • The actual service property the thermal treatment is intended to improve.

Measure before treatment and after the same defined cooling or conditioning interval. Use suitable equipment and keep individual sample records. A single exterior dimension can hide a distorted bore or a bowed surface.

Worked example: compensation is an inverse ratio

Suppose an illustrative 100.00 mm feature becomes 98.50 mm after a controlled process. The observed change is -1.50%. A first screening scale factor would be 100.00 / 98.50 = 1.01523, about a 1.523% increase, rather than simply adding 1.50%.

If another 100.00 mm axis instead becomes 101.20 mm, its inverse factor is about 0.98814. One uniform scale cannot correct those opposite changes. These are hypothetical numbers demonstrating the calculation, not study data or a validated compensation setting.

Do not compensate unstable behavior

A consistent dimensional bias may be investigated with a new model and another proof run. A large spread, unpredictable warping or changing holes is a process-stability problem. Scaling the CAD file does not make that variability disappear.

Use representative geometry and multiple specimens. A small solid coupon and a thin enclosure can respond differently to the same thermal cycle. Filled materials also need their exact manufacturer instructions rather than inheriting an unfilled material's process.

Keep a complete thermal record

  1. Record grade, lot, moisture state and print profile.
  2. Record the supported thermal cycle and actual equipment limitations.
  3. Document support or restraint and the cooling procedure.
  4. Measure the same marked features before and after.
  5. Repeat the relevant service or assembly check.

Use appropriate dedicated equipment and the maker's handling instructions. Do not invent a household heating recipe from an academic specimen temperature. Report dimensional results separately from mechanical or thermal claims so a visually stable part is not mistaken for a qualified one.

Evidence note: this guide combines the cited public documentation with GoodPrints3D analysis. Any numerical example labeled illustrative uses assumed inputs; it is not a GoodPrints3D equipment test. Sources reviewed November 19, 2026.

Sources and references

Related GoodPrints3D guides and tools