What are the differences, and pros & cons, between 3D printers with varying layouts for moving head vs. moving build plate?

Example layouts would include:

  • X Head; YZ Bed;
  • XY Head; Z Bed;
  • XYZ Head;
  • etc.

In particular, what are their respective strengths, weaknesses, specializations, maintenance considerations, etc.?

  • $\begingroup$ I believe this question falls into the "too vague" category. I believe this answer falls into the not-specific-enough-answer category. $\endgroup$
    – fred_dot_u
    Commented Jun 17, 2016 at 1:06
  • 2
    $\begingroup$ I tried an answer underneath, because even though I agree that it is 'too vague', I think it is an answer that is of many people's concern in exactly that way when they begin. I would leave it as a portal to more specific questions which I would try to link in my answer. $\endgroup$
    – kamuro
    Commented Jun 17, 2016 at 8:29
  • 1
    $\begingroup$ The page reprap.org/wiki/Category:Mechanical_arrangement has links to "Cartesian-X-head", "Cartesian-XY-head‎", "Cartesian-XYZ-head", "Cartesian-XZ-head", "Cartesian-Z-head‎", etc. which give a list of (some) printers that use each arrangement. $\endgroup$
    – David Cary
    Commented Jun 17, 2016 at 13:04
  • $\begingroup$ I'm glad this question has stayed up, it provides a good overview on the topic. $\endgroup$ Commented Mar 16, 2018 at 0:51

2 Answers 2


Without going into too much detail, since this is a very exhaustive topic, I'll write some pro's of each down from the top of my head:

Cartesian XZ hotend, Y bed (eg. Prusa Mendel):

  • easy to build (relatively)
  • easy to maintain
  • easy to modify
  • understandable kinematics
  • with the right frame, no x-y-z orthogonality (90 degree angles) needs to be adjusted
  • affordable
  • bad for timelapse recordings
  • print quality will theoretically always be inferior at the same speeds and accelerations to kinematics that have less mass to move (heavy printbeds will lead to ghosting)
  • z-wobble is only existent in this approach
  • big build-plates are no option for this design (last feasible size might be 20x30 cm)

Cartesian XY hotend, Z bed (core-XY, sparkcube, Ultimaker, Makerbot)

  • less mass to be moved -> faster print speeds possible
  • almost no size limitaions
  • construction is easy to enclose in most models due to the cubic frame
  • looks almost always professional
  • enclosure can be hard to modify due to constraints in space

XYZ hotend (Delta bots)

  • master of circles
  • less mass to be moved -> faster print speeds possible
  • impressive to watch
  • more load on the processing unit due to more complicated kinematics (32 bit needed for fast print speeds and responsive control with display)
  • kinematics not easily understandable
  • error-cause search can be very complicated
  • more accurate in the center than on the outer limits due to the kinematic approach

The list is for sure not complete, and as a major disclaimer: print quality will always, with every approach, depend more on the setup and calibration of the printer than on the model. There are people around that produce great prints from an acrylic frame cartesian printer and lots of people that produce mediocre results with expensive printers in fancy designs.

I will add some links to the list items when I find the time, for now you have to believe me. I am highly appreciating corrections and additions!


I think this is simple, breaking it into simple parts... It is much easier to calibrate a machine where each axis does one thing. For instance, the Ultimaker; one of the few XY-on-the-same-axis printers... It has such a crazy complicated pulley system. If one thing gets out of whack, everything will start binding. Thankfully this doesn't happen as they did a good job, but it comes at a price. All the double pulleys and double rods - that costs money.

However if you break it into smaller bits then less will go wrong.

When I upgraded my MendelMax 1.5 to use a PBC linear solid slide for the X axis, it instantly solved all my issues with the rods binding (two rods into one big slide).

To quote my coworker,

It's like balancing a stool. Is it easier to balance a 4 legged stool or a 3?

Following from that, there is probably a good hour's worth of mechanical engineering information that can be said in regards to the linear motion itself. I would rather calibrate one axis at at time, and not worry about weird behavior caused by a XYZ head.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .