ASA Annealing Process and Strength Gain: How Much, How Long, How to Avoid Ruining the Part

What ASA annealing actually does to a 3D printed part

The ASA annealing process strength gain conversation is full of numbers thrown around without context. People say “annealing makes parts 50% stronger” and skip past which strength, in which direction, and what trade-off they paid in dimensional accuracy. Annealing is real and worth doing for the right parts, but the gains are highly directional and the cost is real shrinkage. This is the honest version.

Annealing is a controlled heat treatment that lets the polymer chains in your printed part move slightly and re-organize themselves. In FDM-printed ASA, the original print is highly anisotropic — strong in the X/Y plane along the print lines, weak in the Z direction between layers. Heat applied below the glass-transition temperature gives the polymer just enough mobility for inter-layer chains to entangle without melting the part. The result is a part where the Z-direction layer adhesion approaches the X/Y in-plane strength.

The temperature window that defines ASA annealing process strength gain

ASA’s glass transition temperature (Tg) is around 100-105°C depending on the specific formulation and any UV stabilizers added. The melting onset is around 240°C, well above any practical oven. The annealing window for ASA is therefore between 90°C and 100°C — hot enough that the polymer becomes mobile, cold enough that the part does not slump under its own weight.

Going hotter does not give you more strength. Once you cross Tg, the part is rubbery and any unsupported feature deforms permanently. People who anneal at 110°C and report ruined parts learned this the hard way. Stay at 95°C ± 5°C as a safe target. Hold the temperature for 1-2 hours per centimeter of part thickness, with smaller parts (under 5 mm thick) needing only 60-90 minutes total.

Strength gains broken down by direction

Published studies and hobbyist tests on ASA annealing converge on a fairly consistent set of numbers. Layer adhesion (Z-direction tensile strength) gains 30-50% after a proper anneal — this is where the marketing claims come from. In-plane (X/Y) tensile strength gains 5-15%, much smaller because those chains were already well-fused during printing. Impact strength is roughly unchanged or slightly lower, because the part becomes more brittle as crystallinity increases.

So if your ASA part fails by delamination between layers — which is the dominant failure mode for printed parts under any sustained load — annealing helps a lot. If it fails by a sudden impact, annealing makes it slightly worse. If it fails by abrasion or fatigue, annealing makes no measurable difference. Match the post-process to the failure mode you are designing against.

Dimensional shrinkage is not optional

The other half of the trade is dimensional. ASA shrinks 1-3% during annealing as the polymer chains pack tighter. This is not a print-orientation effect — it is volumetric. A 100 mm cube becomes a 97-99 mm cube, with the largest shrinkage typically along the print direction (Z) and somewhat less in X/Y because in-plane shrinkage is partially constrained by the layer geometry.

For functional parts that need to fit something, you have two options. Either anneal a test print, measure the shrinkage in each dimension, and scale your CAD model up by the inverse percentage before slicing the production print — or design tolerances that accommodate 2% shrinkage and accept the slight slop. The first option produces a stronger part with the right dimensions; the second is faster but only works when the mating tolerance is loose enough to absorb the shrink.

Oven annealing — the practical method

A standard kitchen oven is the most common annealing tool. Convection ovens are better than radiant because the temperature uniformity is much better, but a regular oven works if you are willing to verify the actual temperature with an oven thermometer. Many home ovens read 25°C off their dial, and “low” on a gas oven is often above 130°C — verify before you trust the dial.

Place the part on a flat surface that will not deform — a piece of glass, ceramic tile, or aluminum sheet. Avoid plastic shelves; they will sag at 95°C and your part will sag with them. Set the oven to 95°C, wait 15 minutes for it to stabilize, place the part inside, and start a timer. After the hold time elapses, turn the oven off and let the part cool inside the oven for at least an hour. Pulling a hot part out into room-temperature air introduces thermal shock that can crack thinner sections.

If your oven is also used for food, run a clean cycle first or anneal in a covered glass dish. ASA does not off-gas heavily at 95°C but it is not nothing — and the smell lingers in oven insulation if you ignore it.

Salt bath annealing — better but specialized

The hobbyist community has popularized salt bath annealing as a higher-quality alternative. The method packs the part inside a sealed container surrounded by ordinary table salt, which is then heated in the oven. The salt provides uniform thermal contact on every surface and physically supports the part against gravity-driven deformation. Parts that would slump in air can hold their shape in salt.

The trade-off is mess and time. Salt has to fill every cavity of the print, you cannot use it on parts with internal voids that will be hard to clean, and pulling a 1 kg salt-packed casserole dish out of a 95°C oven takes care. For high-value functional parts where dimensional stability and full strength gain both matter, it is worth the hassle. For a bracket that just needs not to break, plain oven annealing is fine.

Cooling rate matters more than people think

Both methods rely on slow cooling — and “slow” here means more than an hour to drop from 95°C to room temperature. Fast cooling locks the polymer chains in a partially-organized state and you lose half the strength gain. The simplest practical rule is: turn the oven off when the timer ends and do not open the door for at least 90 minutes. The thermal mass of the oven keeps the cooling rate gentle.

For salt baths, the salt mass acts as a thermal capacitor and makes cooling automatically slow — another reason the method produces consistent results. Just leave the container in the oven overnight if you can.

Which ASA parts to anneal and which to skip

Anneal ASA parts that are: structural, load-bearing, exposed to outdoor temperature swings (already a good ASA use case), or thin enough that layer adhesion failure is the dominant risk. Skip annealing on parts that are decorative, have already-tight tolerances you cannot recompensate, contain mechanical inserts that will move during the anneal (threaded brass inserts especially), or are very thick (over 30 mm) where uniform heating becomes hard to guarantee in a home oven.

The other context for annealing is post-processing alongside fume management — if you have already invested in a setup for printing ASA properly with proper ventilation per our ASA fume and ventilation guide, you have the same airflow infrastructure to handle the mild VOC release during the anneal. Doing both well is a small additional cost.

Comparison with other ASA post-processing methods

Annealing is one of three common post-processes for ASA. Acetone vapor smoothing dissolves the surface and re-bonds it, which closes the layer lines and adds modest layer-adhesion strength gain — perhaps 10-20% in Z direction, far less than annealing. UV exposure cures any free radicals in the surface and improves long-term outdoor durability, but does not affect mechanical strength meaningfully. Annealing is the only treatment that produces large layer-adhesion gains, but it is also the only one with significant dimensional shrinkage.

If your part needs both surface smoothness and strength, do annealing first, then acetone smoothing. The acetone vapor will not affect the annealing-induced crystallinity — they operate on different scales (molecular vs surface). Doing it the other way around makes the smoothed surface re-shrink during annealing and lose its smoothness.

Final practical workflow

Print the part with strong layer-adhesion settings already (high nozzle temp, slow speed on outer perimeters, enclosed printer with chamber heat — see our all-metal hotend filament options guide for thermal context). Print a sacrificial test part of the same shape. Measure it with calipers in three orientations. Anneal it at 95°C for 90 minutes plus 90 minutes cooldown. Re-measure. Calculate the per-axis shrinkage. Scale your CAD model by the inverse, re-slice, print the production part, anneal identically. The production part comes out the right size with the layer adhesion gain baked in.

The first time through this workflow takes a long evening. The second time it takes an hour because you already know your printer’s shrinkage numbers and you trust them. By the third time, you anneal anything important without thinking about it.

Common annealing mistakes that ruin good ASA prints

The four mistakes that account for most ruined annealing attempts are predictable and easy to avoid once you have seen them once. The first is annealing too hot — using 110°C or 120°C “to get the gain faster,” which crosses ASA’s glass transition and produces irreversible slump on any unsupported feature. The second is rushing the cooldown — pulling the part out of a 95°C oven into a cold kitchen and watching the dimensional shift jump from a clean 2% to an uneven 4-5% with internal stress lines visible at the surface. The third is annealing parts with brass threaded inserts already installed — the brass conducts heat, expands, and the surrounding ASA reflows around it, which usually ends with the insert loose or visibly off-axis. Install threaded inserts after the anneal, not before. The fourth is treating annealing as a substitute for good print settings — a part with bad layer adhesion to begin with does not become great after annealing, just less bad. Print well first, then anneal to push it to the next level. The combination of “marginal print quality plus aggressive annealing” tends to produce parts that look fine on the outside and crumble at the layer lines when stressed because the underlying interlayer bonding was insufficient for the polymer chain mobility to bridge.