Other Blogs
PLA shrinks about 0.3 to 0.4% in typical FDM printing, and across brands and setups you'll usually see something in the 0.2 to 0.5% range. That's the smallest shrinkage of the common filaments, which is a big part of why PLA is the material people reach for first. For most parts you can print at 100% scale and never think about it.
Before we go further, one thing worth clearing up. The shrinkage a maker cares about is the dimensional error on the actual printed part, measured with calipers after it cools. That is not the same number a resin datasheet quotes as "mould shrinkage," and most FDM filament datasheets (Prusament, Polymaker) don't even publish a shrinkage figure at all. So the honest way to think about this is empirical: print something, measure it, and compensate if the gap matters.
Every thermoplastic contracts as it cools from print temperature back to the temperature of your room. The size of that contraction tracks two things: how high the glass-transition temperature (Tg) sits, and how far the plastic has to fall from the nozzle down to ambient.
PLA has a low Tg, roughly 55 to 65°C, and it prints effectively amorphous (it doesn't crystallize into a tighter-packed solid on the way down). Both of those work in your favor. The print-to-ambient temperature drop is small, so there's simply less contraction to happen. That's the whole story, and it's why PLA lands at the bottom of the ranking below.
A couple of details that apply to PLA the same as any FDM material:
Here's where the four common filaments land, sorted from most shrinkage to least. The driver is glass-transition temperature and the print-to-ambient drop, not crystallinity, since all four print effectively amorphous.
| Material | Typical shrinkage | Range | Compensation scale |
|---|---|---|---|
| ABS | 0.6 to 0.8% | 0.4 to 0.8% | ×1.006 to 1.008 |
| ASA | 0.4 to 0.7% | 0.4 to 0.7% | ×1.004 to 1.008 |
| PETG | 0.4 to 0.6% | 0.2 to 0.6% | ×1.003 to 1.008 |
| PLA | 0.3 to 0.4% | 0.2 to 0.5% | ×1.002 to 1.005 |
Roughly, ABS and ASA sit together at the top, PETG in the middle, and PLA at the bottom. If you print several materials, this ordering is the intuition to keep.
One myth worth killing: you'll see blog posts claim PLA shrinks 2 to 3%. That number is conflating thermal shrinkage with annealing, where you deliberately heat-treat a printed PLA part and it does change size as it crystallizes. Straight off the bed, PLA does not shrink anywhere near that much.
If you do need the compensation, the workflow is the same for every material:
The scale formula is:
scale% = 100 / (100 − shrink%)
So for a measured 0.3% shrinkage, you scale the model to 100 / (100 − 0.3) = 100.3%, or ×1.003. For PLA that lands in the ×1.002 to 1.005 window from the table.
You don't have to do this by hand. Our material shrinkage calculator takes either a shrinkage percentage or your measured-vs-expected dimensions and gives you both the scale factor and the slicer compensation value. Its PLA default is 0.3%, which matches the typical figure here. It also handles the XY hole-compensation idea, since inner contours shrink a little differently from outer walls, and that's exactly where PLA parts fail a press fit even when the outside measures fine.
A practical rule: for small decorative PLA prints, skip compensation entirely. For a long functional part, a threaded interface, or anything that has to mate with hardware, calibrate once and save the value in your slicer's filament profile.
If you print more than one material, the shrinkage numbers shift and so does the workflow. Here are the siblings to this guide:
And if you're setting up a printer from scratch, our roundup of free 3D printing tools and calculators covers flow rate, E-steps, and the other things worth dialing in alongside shrinkage.
Compensation is just scaling, and scaling is something you can hand off. Describe the part you want in plain English, tell GrandpaCAD the material and the fit you need, and it can model the geometry pre-scaled to compensate, so what comes off the bed is the size you actually asked for.