The shell is most of the print
Ask someone how much plastic is in a print and they will reach for the infill percentage, as though a 15 percent part were 15 percent of a solid one. It never is. A printed part is a solid skin wrapped around a sparse interior, and on small parts the skin is the overwhelming majority of the material.
Take a 50 mm cube. Its volume is 125 cm3 and its surface area is 150 cm2. Two perimeters at 0.42 mm give a 0.84 mm wall; four top and four bottom layers at 0.2 mm give a 0.8 mm skin. Averaged over a surface that is two thirds vertical, the shell envelope is about 0.83 mm thick, which is 12.4 cm3 of plastic before a single line of infill is laid. The remaining 112.6 cm3 at 15 percent infill adds 16.9. Total: about 29 cm3, or 36 grams in PLA. The shell is 42 percent of the plastic in a part most people would describe as mostly infill.
Shrink the part and the ratio gets worse, because surface area falls with the square of the size while volume falls with the cube. A 25 mm cube has one eighth the volume and one quarter the surface area, so the shell goes from 42 percent of the material to well over 60. This is why raising infill from 15 to 25 percent on small parts barely moves the scale, and why dropping wall count from three to two does.
Grams, cubic centimetres and metres
Three units get used interchangeably and each has a fixed conversion. Volume to weight is density: PLA is nominally 1.24 g/cm3, PETG 1.27, ABS 1.04, ASA 1.07, TPU around 1.21 and nylon PA6 around 1.14. Volume to length runs through the filament cross-section. A 1.75 mm filament is a circle of radius 0.875 mm, so its cross-section is pi times 0.875 squared, which is 2.405 mm2. One metre of it is 2405 mm3, or 2.405 cm3.
The check worth memorising: a 1 kg spool of PLA at 1.24 g/cm3 holds 806 cm3 of plastic, which at 2.405 cm3 per metre is about 335 metres of 1.75 mm filament. If a calculator ever tells you a kilogram of PLA is 200 metres or 500 metres, something in it is wrong. In 2.85 mm the same kilogram is only about 126 metres, because the cross-section is 2.65 times larger.
| Material | Density (g/cm3) | cm3 in 1 kg | Metres of 1.75 mm per kg |
|---|---|---|---|
| PLA | 1.24 | 806 | 335 |
| PETG | 1.27 | 787 | 327 |
| ABS | 1.04 | 962 | 400 |
| ASA | 1.07 | 935 | 389 |
| TPU 95A | 1.21 | 826 | 344 |
| Nylon PA6 | 1.14 | 877 | 365 |
Surface area is the input people skip and shouldn't
Volume is easy to find and surface area is one menu item away in the same tools, but people leave it blank and take the cube estimate. For a chunky bracket that is close enough. For anything with holes, slots, ribs, lettering or a lattice, the cube estimate can be low by a factor of two or more, and since the shell dominates the total, so is the weight.
The vertical share field handles the other half of the geometry problem. Vertical faces receive the perimeter walls, so their thickness is wall count times line width. Flat faces receive the top and bottom solid layers, so their thickness is layer count times layer height. Those two are usually different numbers, and which one dominates depends on the part. A flat plate is mostly horizontal surface and its weight is driven by the solid layer count; a tall column is mostly vertical and its weight is driven by perimeters. Set the share to match the shape you are printing rather than leaving it at the cube default.
Where this parts company from a slicer
A slicer generates real toolpaths and this generates an envelope. The differences are systematic and mostly in one direction. Slicers add gap fill between perimeters that do not quite meet, they add a brim or skirt, they add support material and its interface, and they prime the nozzle before the first layer. Infill patterns rarely hit their nominal density exactly, and features like ironing or extra solid infill near surfaces add material this model knows nothing about.
Against that, this model double-counts a little where the top skin meets the perimeters at a corner. On simple parts the two errors partly cancel and the answer lands within about 10 percent of a slice. On a support-heavy print of a complicated organic model, it will be well under, and no adjustment to the fields will fix that, because support volume depends on overhang geometry rather than on anything entered here. Once you have a number you trust, carry it into the print time estimator and the filament cost calculator.
Questions people ask
What infill percentage should I use?
For most functional parts, somewhere between 15 and 25 percent, with the strength coming from wall count rather than infill. Decorative pieces and prototypes work fine at 10 or below. Parts that carry a real load want three or four perimeters at 20 percent rather than two perimeters at 50, because in a wrapped-shell structure the outer material carries the bending load and the interior mostly stops the walls buckling inwards. Above 50 percent the return on extra plastic and time is small; above 80 percent you are close enough to solid that you may as well print solid and stop pretending. The one exception is parts that will be tapped, drilled or take a threaded insert, where the local density around the feature matters more than the average.
Why is my slicer heavier than this calculator says?
The usual causes, roughly in order: support material, which this does not model at all; a surface area that was estimated as a cube and is really much larger; a brim or raft; and gap fill, which slicers add wherever two perimeters cannot quite close a thin region. If the difference is more than about 20 percent on a part with no supports, the surface area is the first thing to check. Enter the true figure from your CAD package or slicer rather than leaving the field blank.
Does infill pattern change the weight?
Slightly, and less than the pattern names suggest. At the same nominal density, the common patterns land within a few percent of each other, because the slicer adjusts line spacing to hit the target density regardless of pattern. What genuinely differs is whether a pattern is two-dimensional or three: grid, lines and triangles repeat the same shape on every layer, while gyroid, cubic and honeycomb vary through Z. That affects stiffness direction, print time and how the pattern interacts with the top skin, not the total plastic. Choose a pattern for strength direction and print speed, not for weight.
How do I find the volume of my model?
Almost every CAD package reports it under a mass or physical properties command. Mesh tools show it for STL files, and a slicer will show volume or the equivalent filament figure once the model is loaded. If you already have the slicer number in grams, you do not need this page at all for that model; use it for the case where you are deciding settings before you slice, or estimating a model you have not downloaded yet from published dimensions. Make sure the volume is in cubic centimetres. A model reported as 125000 is almost certainly cubic millimetres, which is 125 cm3.
Can I use this for a hollow vase-mode print?
Not directly, and the fix is straightforward. Vase mode prints a single continuous perimeter with no infill, no top and, usually, a few solid bottom layers. Set the wall count to 1, the infill to 0, and the top and bottom layers to just the bottom count doubled, since the field takes a combined figure. What you get is the surface area times one line width, which is exactly what a vase-mode print is. The result is usually surprisingly small: a 200 mm tall vase can be under 60 grams.