MESHBY KINETIC CARVING
Modeling · Extrusion · Slicing · Help
Return to Kinetic Carving MembershipJoin
Hardware · MotionbeginnerFDM~5 min read

Cartesian, CoreXY and Delta Printers

Short answer

A Cartesian printer gives each axis its own motor; on the common bed-slinger the bed itself moves front to back, so that axis carries the most mass. A CoreXY keeps both XY motors on the frame and moves a light head, which is why it suits fast printing. A delta moves the nozzle with three towers and arms, and its print quality depends on accurate geometry calibration.

What the motion system decides

Every FDM printer has to put the nozzle at an X, Y and Z position relative to the bed. The motion system is the arrangement of motors, belts, rails and screws that does it. It decides which parts move, how heavy they are, and therefore how hard the printer can accelerate before vibration shows up on the print. It also decides which calibration matters most. Firmware hides the math: Marlin and Klipper both let you set speed and acceleration limits in ordinary X, Y and Z terms on all three layouts below.

Cartesian and the bed-slinger

Marlin describes a Cartesian printer as one that moves the nozzle along three straight axes at right angles, with each axis driven by its own motor. Klipper's kinematics notes say the same thing from the firmware side: movement on each axis maps directly to movement in Cartesian space, which makes it the simplest case to control.

The most common Cartesian layout is the bed-slinger. The head moves left and right on a gantry, the gantry moves up and down for Z, and the bed slides front to back for Y. Some Cartesian machines keep the bed still and move the whole gantry for Y instead; the firmware treats both the same way.

The practical effect: on a bed-slinger the Y motor carries the bed and the print. Marlin points out three consequences:

  • Y usually needs lower acceleration than X, because the bed is the heaviest moving part.
  • A tall, thin print on a fast-moving bed can wobble.
  • Heavier moving parts ring more, so ghosting near sharp corners tends to show first on the Y axis, and its ringing frequency is usually lower than X's.

The upside is that settings mean what they say: steps per millimeter, speed and acceleration are set per axis, and a problem on one axis is usually that axis's motor, belt or pulley.

CoreXY

On a CoreXY the bed moves only in Z. The head moves in X and Y on a light gantry pulled by two belts, and both motors are fixed to the frame. Neither motor owns an axis. Marlin explains that when the two motors turn the same way the head moves in X, and when they turn opposite ways it moves in Y. Klipper's notes describe the same thing as one motor following the sum of X and Y and the other following their difference.

On a bed-slinger the bed carries the print in Y. On a CoreXY two motors fixed to the frame move the head in X and Y together, and the bed moves only in Z. Belt paths are not drawn.

The practical effect is low moving mass. Marlin says this is why CoreXY is popular for fast printers, and that a light, stiff CoreXY frame can often run much higher acceleration than a bed-slinger. At those speeds ringing, not motor power, becomes the limit, which is where input shaping comes in.

What commonly goes wrong:

  • Motor directions. Marlin calls getting the motor directions right the step that most often goes wrong. If X and Y are swapped, swap the motor cables or the driver assignments; if one axis runs backward, flip that axis's direction setting.
  • Steps per millimeter. Both motors contribute to every move, so Marlin advises keeping X and Y steps per millimeter the same.
  • Sensorless homing. Because both motors move for either axis, stall sensitivity has to be sensible on both, or homing stops early or crashes.

Delta

A delta has three vertical towers around a round bed that never moves. Each tower carriage drives a pair of arms that hold the nozzle platform (the effector). Marlin notes that no single motor moves the nozzle in X, Y or Z: even a straight move needs all three motors at changing speeds. Klipper computes each tower's position from arm length, tower offset and height using Pythagoras's theorem, and it lowers maximum speed and acceleration near the edge of the build area, where an arm approaches horizontal.

The practical effect is light moving parts and a tall, round build volume, at the cost of calibration. Marlin is blunt about it: if the geometry values are even slightly wrong, a flat move comes out domed or dished and circles print as ovals. It says calibration matters more on a delta than on any other printer type, and that using a bed mesh to hide geometry errors makes the middle look fine while the edges get worse. On a Marlin delta, run the geometry calibration (G33) before any bed leveling, and repeat it until the results stop changing.

Comparing the three

LayoutWhat movesUsual limitCheck first when prints go wrong
Bed-slinger CartesianHead in X, bed in Y, gantry in ZBed mass in YY belt and acceleration, tall prints wobbling
CoreXYLight head in X and Y, bed in ZRinging at high accelerationMotor direction, belt condition, input shaping
DeltaThree tower carriages and armsGeometry accuracyGeometry calibration before mesh leveling

Acceleration, ringing and input shaping

Whatever the layout, firmware limits how quickly the head changes speed. Klipper uses constant acceleration and enforces it at the print head, which indirectly limits the torque each motor needs; it says limiting speed changes reduces the risk of disrupting the print. Heavier moving parts store more energy at each corner and shake more when they stop, which shows up as ghosting. Both Klipper and Marlin offer input shaping to reduce these vibration artifacts. Klipper notes it can also allow faster printing while keeping quality. See Ghosting and Ringing for the symptom and Print Speed and Acceleration for the slicer side.

How to check your machine

  1. Find out which layout you have: watch which parts move during an X move and a Y move.
  2. Before the first home on a new or rebuilt machine, make short positive moves on each axis and confirm direction. Marlin advises verifying homing direction and endstop behavior before homing.
  3. On a bed-slinger, keep Y acceleration more conservative than X, as Marlin suggests, and watch tall prints for wobble.
  4. On a CoreXY, if diagonal moves look wrong or one axis is reversed, check motor direction and driver assignment first; Marlin says direction is the step that most often goes wrong.
  5. On a delta, calibrate geometry first, then level. Domed or dished first layers point to geometry, not to the bed.
  6. If corners show echo lines, reduce acceleration as a test before changing anything else. If the echoes shrink, the cause is vibration, and input shaping or a stiffer frame is the fix. If prints shift sideways, go to Layer Shifting.
Sources

Last reviewed 2026-11-05