Guide
Calibration Prints and the Benchy
There is an entire genre of calibration models and most beginners print too many of them. Here is which four actually tell you something, what each one is diagnosing, and when not to bother.
By Scooter M. · Published · Last reviewed
The short answer
Print a Benchy and a calibration cube. That is enough for a fortnight. Modern machines calibrate their own flow and resonance, so most of the calibration genre is solving problems your printer already handled. Print calibration models to diagnose a specific problem, not as a ritual before you are allowed to make something.
Why there are so many of these
Calibration prints came from an era when printers arrived uncalibrated. You tuned extrusion, retraction, temperature and resonance by hand, and each of those tuning steps had a purpose-built test model. That work was real and the models were essential.
Most current beginner machines do a large part of it themselves. Bambu Lab publishes automatic flow-dynamics and vibration calibration on the A1 series; the Creality Ender-3 V3 line publishes automatic bed leveling and Z-offset. That does not make calibration models useless — it changes what they are for. They are diagnostics now, not a setup ritual.
The four worth printing
3DBenchy — the general diagnostic
A small tugboat with a curved hull, an overhang under the bow, a bridge over the cabin, a thin chimney and text on the base. Every common failure shows up somewhere on it in a recognizable place.
What to look at: the hull for smooth curved surfaces and consistent extrusion; the bow overhang for drooping; the cabin roof for how well it bridges; the chimney for whether small cross-sections print cleanly; the base text for fine detail.
About an hour, under 15 grams, no supports. Print one early so you have a known-good reference to compare against later.
The calibration cube — dimensional accuracy
A 20 mm cube with X, Y and Z on its faces. Print it, measure each dimension. If they are all close to 20 mm, your machine is dimensionally honest and you can stop worrying about that whole category of problem.
A cube that is consistently over-size suggests over-extrusion; one that is under suggests the opposite. On a machine with automatic flow calibration this is rarely wrong, which is exactly why it is worth twenty minutes to confirm.
A temperature tower — when layers separate
A tall model printed in bands, each at a different nozzle temperature, usually stepping down from the top of the filament’s published range to the bottom. When it is done you look at which band is strongest and cleanest.
Worth printing when you have a specific problem: layers separating, poor bonding, or stringing that will not go away. Not worth printing as a routine, because the manufacturer’s published range is a good starting point — Overture publishes 190–220 °C for its PLA and that is where you should be.
A retraction test — when you have stringing
Two or more towers with a gap between them, so the nozzle has to travel across empty space repeatedly. Strings between the towers show how much plastic is oozing during travel.
Print it when stringing is your actual problem, and remember the first thing to rule out is not retraction settings but wet filament — see storage and drying. PETG strings far more readily than PLA, which is covered on the PLA against PETG page.
The ones you can skip
Calibration models that were essential on older machines and are usually redundant on a current one.
| Model | What it tests | Why you can skip it |
|---|---|---|
| Flow / extrusion multiplier test | How much plastic comes out | Modern machines calibrate flow automatically |
| Input shaping / resonance test | Vibration compensation | Automatic on current machines with the feature |
| Overhang test | Maximum printable angle | Useful once, out of curiosity. Not diagnostic |
| Tolerance / clearance test | How much gap moving parts need | Worth it only when you start designing your own parts |
| All-in-one calibration models | Everything at once | They test so many things that a failure tells you nothing specific |
That last row is the important one. A model that tests eight properties simultaneously produces a result you cannot interpret, because you cannot tell which property caused which flaw. A calibration print is only useful if a bad result points at exactly one cause.
What to do with a bad Benchy
Read it as a set of separate signals rather than a verdict.
- Rough or pitted surfaces, popping while printing: wet filament.
- The bow overhang drooping badly: cooling, or printing too fast.
- The cabin roof sagging: bridging, which is again mostly cooling.
- Strings between the chimney and the hull: retraction, or wet filament first.
- The base not stuck down or lifting at a corner: a bed adhesion problem — work through the ordered list.
- Ripples down the side of the hull: vibration, usually a wobbly surface rather than the printer. Where to put your printer covers it.
Questions people actually ask
+−What is a 3DBenchy?
A small tugboat model designed as a general printer test. It packs a curved hull, an overhang, a bridge, a thin chimney and fine text into about an hour and under 15 grams, so every common failure appears in a recognizable place.
+−Do I need to calibrate a modern 3D printer?
Much less than you used to. Current beginner machines publish automatic bed leveling, Z-offset, flow and resonance calibration. Print a Benchy and a calibration cube to confirm all is well, then go and make something.
+−What should a calibration cube measure?
20 mm in each axis, or very close to it. Consistently over-size suggests over-extrusion; consistently under suggests the opposite. On a machine with automatic flow calibration it is usually right, which is why it is a twenty-minute confirmation rather than a tuning exercise.
+−Should I print a temperature tower?
Only when you have a specific problem — layers separating, poor bonding, or persistent stringing. The manufacturer’s published temperature range is a good default, and printing towers as a ritual is a way to spend a week not making anything.