Print time
How long does a 3D print take? Five parts, three layer heights
Print time is extruded volume divided by flow, plus an allowance for travel and layer changes. That is the whole model — it is what the print-time calculator runs, and it is accurate to about ±20%. Below it is applied to five part classes, then to your own model.
The five parts
| Part | Extruded volume | 0.2 mm | 0.12 mm | 0.28 mm | PLA mass |
|---|---|---|---|---|---|
| Keychain tag a 40 × 25 × 4 mm name tag | 1.5 cm³ | 5 min | 9 min | 4 min | 2 g |
| Phone stand a folding desk stand, 2 walls | 25.0 cm³ | 1.5 h | 2.5 h | 1.1 h | 31 g |
| Vase 180 mm tall, single-wall spiral | 35.0 cm³ | 2.1 h | 3.5 h | 1.5 h | 43 g |
| Helmet cosplay shell, 12% infill, supports | 180.0 cm³ | 10.6 h | 17.7 h | 7.6 h | 223 g |
| Large figurine 250 mm tall display piece, 15% infill | 60.0 cm³ | 3.5 h | 5.9 h | 2.5 h | 74 g |
| Basis: volume ÷ (layer height × line width × speed) × (1 + 15% overhead). Hours are estimates with a ±20% band, not slicer output. | |||||
Volumes are the plastic the slicer lays down, not the bounding box: a 120 × 80 × 60 mm enclosure has a 576 cm³ box and about 170 cm³ of extruded plastic at 3 walls and 20% infill. Mixing those up is the usual reason a hand estimate is hours out — that conversion has its own page.
The estimate, in one line
hours = volume_mm³ ÷ (layer height × line width × speed) ÷ 3600 × (1 + overhead). At 0.2 mm layers, 0.45 mm wide and 60 mm/s the machine deposits 19.44 cm³ an hour, and the 15% overhead for travel and layer changes brings the effective rate to 16.90 cm³/h. A 25 cm³ phone stand therefore takes 1.5 h.
That is faster than a real print: a slicer inserts accelerations, retractions, travel moves and cooling pauses, which is where the ±20% comes from. Use it to decide whether a job fits in tonight, not to promise a customer a 15:00 delivery.
Your own model
Put your slicer's volume in, set the profile's layer height, and read hours and grams off the panel. Material changes the mass, not the time.
Layer height is the lever with the widest range
Hours scale inversely with layer height, so going from 0.28 mm to 0.12 mm on the same model is a 2.3× change with no other settings touched: the helmet goes from 7.6 h to 17.7 h, and the vase from 1.5 h to 3.5 h. Printing a functional bracket at 0.12 mm buys nothing you can see and costs 133% more machine time.
Nozzle width is the other multiplier: a 0.6 mm nozzle deposits 1.33× the plastic at the same speed, so the phone stand drops to about 1.1 h — at the cost of surface detail.
What this model cannot see
- Supports and infill — already inside the volume if it came from your slicer; if you estimated it yourself, add them first, since supports can be 10–20% of a helmet.
- Acceleration and firmware limits. A machine that never reaches 60 mm/s on short segments prints slower than this, and at high speed the extruder, not the motion system, becomes the limit.
Time is the product you are selling
A printer has 168 hours a week and you have fewer in attention. At 20 productive hours a week you can deliver 2 helmets or 14 phone stands, and the machine earns only while it runs — which is why choosing a printer for paid work is a question about hours. Money and time are read together: cost per gram, the full bill, failure rates.
Affiliate disclosure: this page may contain affiliate links.
Frequently asked questions
How long does a typical 3D print take?
A 25 cm³ phone stand is about 1.5 h at 0.2 mm layers, a 35 cm³ single-wall vase about 2.1 h, and a 180 cm³ cosplay helmet about 10.6 h. Small tags finish in minutes; anything over 150 cm³ is an overnight job.
Does a lower layer height really double the time?
It scales with the inverse of layer height, so 0.2 mm to 0.1 mm is a 2x change, and 0.28 mm to 0.12 mm is 2.3x. Wall thickness, infill and speed stay the same; only the number of layers changes.
Why is my print taking longer than the calculator says?
Usually supports and infill were not in the volume you typed in, or the printer never reaches the assumed speed on short moves. The model carries a ±20% tolerance for exactly this; a slicer estimate of the same model is the better number.