From RepRapWiki
Jump to: navigation, search

Drive trains, also called mechanisms or mechanical systems, are ways to transform mechanical motion from the stepper motors or DC motors into a precise movement of the desired part of the RepRap.

The mechanical systems are the part of the RepRap machine that move either the print head, the build platform or both. Generally you have a motor which rotates and that rotation is either used directly as rotational movement, or it is converted into linear movement. The various forms of movement are discussed below.

Some of the articles in this category focus on one specific part of a drive train -- an improved gear, pulley, or rack-and-pinion.


rotary to linear motion conversion

Most drive trains convert the rotary motion of the motor into linear motion. There are 2 popular techniques used by most RepRaps and RepStraps so far:

Other promising techniques currently under development that may turn out better than both include:

overall layout

Other articles in this category describe the overall layout of a complete drive train, of which there are currently 4 major categories being developed by RepRap developers:

  • serial Cartesian drive train (the first ones to successfully RepRap)
  • serial polar drive train
  • parallel drive train (the Rostock is apparently the first working RepRap in a non-Cartesian drive train)
  • parallel SCARA drive train
  • partially parallel Cartesian drive train

(This is far from exhausting the theoretical space of all possible drive trains -- for example, there has been a brief discussion[1] of a parallel Cartesian drive train.[2], and Rhombot is an early prototype of that completely parallel Cartesian drive train)

One advantage of the parallel drive trains (including the partially parallel Cartesian drive train) is that all of the motors can be mounted in a fixed location. Serial drive trains generally require at least one motor to be mounted to a moving part, which means that some *other* motor (and the mechanism attached to it) has to be strong enough to move that motor around and keep it constrained to the desired position.


Some drive train mechanisms use smooth rods or some other kind of rails to properly constrain the motion of one part relative to the base (or to another part).

Several people are trying to substitute some alternative mechanism that requires fewer or no rails. Why is that? Current speculation is:

  • "Eliminating rails is a big step towards a fully printable reprap." -- Rhombot
  • "Sliding rails are somehow (?) mechanically inferior to a few rotating joints" -- ?
  • "A sliding rail costs more (?) than a few rotating joints of equivalent precision" -- ?
  • Some other reason?


Many machine designs show some part sliding along 2 parallel smooth bars. The part slides along one axis, and the way the part is attached to the bars is intended to constrain the part in the other 5 dimensions -- the other 2 axis of motion, and all 3 axis of rotation.

A super-common problem occurs when a designer's human sense of symmetry makes the designer try to attach that part to both bars in the *same* way. Since it's impossible (?) to divide 5 axis of constraints evenly between the 2 smooth bars, and since it's practically impossible to make the 2 smooth bars *perfectly* parallel, we end up with over-constrained motion, leading to binding. (search the RepRap forums for "binding"). (search the RepRap wiki for "binding")

  • Some researchers allow the part to slide in 1 dimension, using one rod to rigidly constrain the moving part in 4 dimensions, then use the other rod to rigidly constrain the moving part in the 1 remaining dimension. See the "Mendel's improvements over Darwin" video on the Mendel page.
  • Some designers attach the same bearings to both rods, then rigidly attach one of those bearings to the moving part, but add a flexure (perhaps a living hinge) between the moving part and the bearing on the other rod. See illustration 3-12 on p. 25 and the "consider the common problem of aligning shafts that support a linear motion carriage" discussion on p. 24 and illustration 3-19 on p. 39 and the discussion on p. 38 of FUNdaMENTALS of Design (Alexander Slocum 2008).
  • The "5 axis of constraints" are apparently only required for serial Cartesian drive trains. The parallel Cartesian drive trains (such as the capstan and bowstring designs) and the serial non-Cartesian drive trains (polar) and the parallel, non-Cartesian drive trains (such as the Rostock) all seem to bypass the binding problem.

Cartesian Robot (3-axis machine)

Most of the machines described on the RepRap wiki are classified acording to their mechanical arrangement. The categories are linked to on the mechanical arrangement list.

Have all possible ways to connect X, Y, and Z in series already been explored?:

  • ... is there some other series arrangement I'm missing?
  • Is there any significant advantage or disadvantage of one series arrangement over another?
  • Which (if any) parallel or partially-parallel drive train is better than the best series drive train?

The Cartesian Robot is a very standard style of positioning system that uses the familiar X/Y/Z coordinate system. Basically, you have 3 perpendicular linear axes: X, Y, and Z. A stepper motor controls each axis and can move it back and forth. One axis will generally be mounted on another axis, while the other axis is standalone. The build platform will be mounted on one axis, and the print head will be on the other. Common configurations are:

  • Print head on X/Y assembly, build platform on Z axis
  • Print head on Z axis, build platform on X/Y
  • Build platform on X axis, print head on Y/Z

The RepRap project has developed many different Cartesian Robot machines:

  • Darwin's 3-Axis System - print head on X/Y axis, build platform on Z. Made from RP parts and steel rod. Belt driven.
  • McWire Cartesian Bot - McWire Cartesian Bot. Print head on Z, build platform on X/Y. Threaded rod driven.
  • Mendel -- Print head on X/Z axes, build platform on Y. Belt driven X and Y, screw driven Z.
  • There are a range of H-Bot machines that have been developed. This is a drive system with the print ehead on an XY assembly but with both motors (substantially) stationary and the head is translated in two axis by the use of more or less complicated timing belt (or wire or line) arrangement that requires coordinated motion of the motors to achieve the required movement of the head. It has appeal as the two axis have similar inertia as neither one has the extra motor mass.

Delta robot

Rostock (see also Category:Delta)

parallel SCARA robot

The Armstrong A1 by ttsalo is apparently the first RepRap in a parallel SCARA arrangement -- a 3D-printable parallel SCARA 3D printer. (see,156261 ). It apparently uses the "Marlin Modified for parallel SCARA printers" firmware. Congratulations!

See category: scara for other RepRaps that remind us of SCARA robots: RepRap Morgan ... Wally, etc.

More details on parallel SCARA and other SCARA arrangements:

Polar Coordinate Robot

see Polar

This style of printer consists of a turntable, an X axis, and a Z axis. The table rotates a build platform, the X axis moves a print head from the center of the table to the outside of the table, and a Z axis moves the print head up or down vertically.

The RepRap project does not currently support a particular polar robot, but we are excited to see several promising prototypes. See Polar for details.


This category has the following 2 subcategories, out of 2 total.


  • Delta(6 C, 46 P)