FDM print time is set by the path the nozzle has to travel and how fast it can travel it, not by the part’s volume. Walls print slowly and use solid plastic, infill prints fast and sparsely, height sets the number of layers, and short moves never reach full speed. Two parts with the same volume can differ in print time by a factor of three or more, which is why a price based on volume alone is wrong in both directions.
Key points
- Mesh volume, extruded volume and print time are three different numbers.
- Height sets the layer count, and every layer carries a fixed overhead.
- Perimeters (walls) are the slowest part of most prints; the outer wall is slowest of all.
- Infill is fast and sparse, so a chunky part can print faster than a thin-walled one of the same volume.
- Short moves spend most of their time accelerating, so fine detail costs more time than its volume suggests.
- Calibrate any estimate against real prints on your own machines.
Why is volume a poor predictor of FDM print time?
Volume tells you how much space the part encloses. It does not tell you how that space is built. An FDM printer lays down three kinds of extrusion, at very different speeds and densities:
| Feature | Density | Typical relative speed | Why |
|---|---|---|---|
| Outer wall (perimeter) | Solid | Slowest | It is the visible surface, so it is printed slowly for quality |
| Inner walls | Solid | Medium | Strength, printed faster than the outer wall |
| Top and bottom skins | Solid | Medium | Closes the part, often printed at wall-like speeds |
| Infill | Sparse, often 10-30 % | Fastest | Hidden inside, printed fast and thin |
| Supports | Sparse | Fast | Removed afterwards |
A chunky part is mostly infill: it encloses a lot of volume but extrudes little of it, quickly. A thin-walled part is almost all wall: nearly every cubic centimetre is solid plastic laid at wall speed. Same volume, completely different amount of work.
What actually sets FDM print time?
Five things, roughly in order of how much they move the number.
1. Height and layer count. The number of layers is height divided by layer height. Each layer has a fixed cost: the Z move, retraction, a wipe, sometimes a minimum layer time for cooling. On a tall thin part the minimum layer time can dominate, because the printer waits for each small layer to cool.
2. Perimeter length per layer. Every layer traces every outline, inside and out, once per wall. A part with a long outline, many holes or many separate islands spends most of each layer on walls.
3. Solid area. Top and bottom skins, and any region too thin to hold infill, are printed solid. Parts with lots of horizontal faces, steps or ribs carry more skin.
4. Infill volume and pattern. The inside of the part at your infill density, adjusted for the pattern. Patterns differ in how much plastic they lay at the same nominal percentage, so a pattern choice is also a time and material choice.
5. Acceleration and travel. The head does not jump to its set speed; it accelerates. On short segments it never gets there. Travel moves between islands add time without adding plastic.
How much does acceleration matter?
A simple way to model it is to treat each segment’s effective speed as the set speed multiplied by
1 - e^(-L/La), where L is the segment length and La is an acceleration length for your printer. With an
example acceleration length of 5 mm:
| Segment length | Share of set speed reached on average |
|---|---|
| 1 mm | About 18 % |
| 2 mm | About 33 % |
| 5 mm | About 63 % |
| 10 mm | About 86 % |
| 50 mm | Practically 100 % |
The consequence: a part full of small text, fine lattices or tiny holes prints at a fraction of the speed you configured, while a long straight wall prints at full speed. A volume-based price undercharges the first and overcharges the second. The real acceleration length depends on your firmware’s acceleration and jerk settings, which is why it is worth fitting from your own prints.
A worked comparison: two parts of the same volume
Take two parts of almost exactly the same mesh volume, 64 cm³, and print them with the same settings. Part A is a solid 40 mm cube. Part B is an open-topped box, 120 × 100 × 60 mm, with 1.6 mm walls and a 2 mm floor.
Worked example: the same volume, very different print times (simplified, example settings)
Settings: 0.45 mm line width, 0.2 mm layers, 3 walls, 1 mm top and bottom skins, 15 % infill. Outer wall 40 mm/s, inner walls and skins 80 mm/s, infill 120 mm/s, 2 s per layer change. Travel and acceleration ignored. PLA at 1.24 g/cm³.
| Part A: 40 mm cube | Part B: open box | |
|---|---|---|
| Mesh volume | 64.0 cm³ | 64.2 cm³ |
| Height and layers | 40 mm, 200 layers | 60 mm, 300 layers |
| Plastic extruded | 8.3 cm³ walls + 2.8 cm³ skins + 7.9 cm³ infill = 19.1 cm³ | 64.2 cm³, all solid |
| Filament mass | 19.1 × 1.24 = 23.6 g | 64.2 × 1.24 = 79.7 g |
| Outer wall path | 160 mm × 200 = 32 m | 440 mm × 300 = 132 m |
| Outer wall time | 32 m at 40 mm/s = 13.3 min | 132 m at 40 mm/s = 55.0 min |
| Other walls, skins, infill | 31.8 min | 121.2 min |
| Layer changes | 200 × 2 s = 6.7 min | 300 × 2 s = 10.0 min |
| Estimated print time | 51.8 min | 186.2 min (3.1 h) |
Same volume, about 3.6 times the print time and 3.4 times the plastic. A quote of volume × a rate would charge them the same. In practice the gap narrows a little because the cube’s short 40 mm sides lose more time to acceleration than the box’s long walls, but the box is still far slower. We made the same point, from the pricing side, in how to price FDM prints.
How do supports and orientation change the estimate?
Support structures add their own extrusion, usually sparse and fast, plus interface layers that print slowly, plus labour to remove them. Orientation changes all five drivers at once: laying a tall part flat cuts layers sharply but may lengthen each one and add supports. For pricing, the useful habit is to estimate time in the orientation you will actually print, not the one the file arrived in. Our design for 3D printing guide covers the geometry choices that make a part quicker to print.
How accurate is a slicer’s time estimate?
A slicer computes the full toolpath, so it knows the path length exactly. Where it can drift is the motion model: acceleration, jerk or junction deviation, minimum layer time and any flow limit the firmware imposes. A slicer set up with your printer’s real firmware values is usually close; one using a generic profile can be off by a wide margin. Either way, check it: print a few representative parts, log the real times, and keep a correction factor per printer and material.
For quoting, slicing every upload is slow and needs a configured slicer per machine. A geometry-based estimate, built from per-layer cross-sections (perimeter length, solid area, inside area) and your speeds, gets close enough to price on, as long as it is calibrated against your real prints.
What this means for pricing
Machine time is usually the largest line in an FDM price, so the time estimate is the price. If you estimate from volume you will win thin-walled, detailed parts you lose money on and lose chunky parts you could have made well. Feed the estimate into the full cost model in how to calculate 3D printing cost; if you run many printers, the hourly rate that estimate is multiplied by is covered in print farm pricing.
Quotometry estimates FDM time from a per-layer profile of the part and the speeds, acceleration and overheads you set for each printer, and its calibration fits those coefficients to jobs you have run. How that works for FDM is set out on the FDM process page.
Questions people ask
Can I estimate FDM print time from the part's volume?
Only roughly, and only for families of similar parts. Volume does not tell you how much of the part is walls, which print slowly, versus sparse infill, which prints fast and uses little plastic. It also says nothing about height, which sets the layer count, or about short moves where the head never reaches full speed. Two parts with the same volume can differ in time by a factor of three or more.
Why does my slicer's time estimate differ from the real print time?
Slicers model acceleration, jerk and cooling pauses with simplified firmware settings, and those settings may not match what your printer actually runs. Firmware also caps speed by volumetric flow on some materials. Compare a few finished prints with the estimate and apply a correction factor per printer and material until the two agree.
Does a lower layer height double the print time?
Close to it. Halving the layer height doubles the number of layers, and each layer still has to trace every perimeter and fill every area. Per-layer overheads such as layer change and retraction also double. Higher flow per line partly offsets this on large flat areas, but for most parts 0.1 mm layers take nearly twice as long as 0.2 mm.
What is the fastest way to cut FDM print time on a part?
Reduce height first, by choosing a flatter orientation if the part allows it, because height sets the layer count. Then reduce wall count and infill density if strength allows, use a larger nozzle or thicker layers for non-cosmetic parts, and avoid tall thin features that force cooling pauses. Each changes the time far more than shaving volume.
Quotometry prices every part from your own machines and rates, on your website, in seconds.