PLEASE NOTE: My gantry corner brackets have ERRONEOUS hole dimensions at the moment and won't work. I will try to fix this soon. This part alone may change a hundred times before I'm happy. Also, the github repo for the OpenSCAD parts models is now live.

GOALS:

The aim is to design and build a basic CoreXY 3D printer using cheap and ubiquitous 20mm x 20mm x 3mm aluminium tubing. Currently very much a WIP. The initial draft will have a print volume of roughly 200mm x 200mm x 200mm, but the aim is to make the system scalable to any reasonable size, using any reasonable square tube profile.

The primary goal is to get CoreXY happening inside US$200/AU$300 at 2024 prices. Upgrades will come later. I also hope it might be something that Ender 3 owners might be able to print all the parts for before cannibalising their ender for most of the mechanical parts.

The linear rails can be either 8mm stainless steel dowel for both x and y rails or stainless steel tube for the x rails. I have a stock 16 bit Ender 3 control board floating about so the initial prototype will use that to prove the concept, however, I reckon anything from BigTreeTech's SKR range will be very satisfactory and Klipper or Marlin (I know Marlin quite well, probably Marlin to start with.)

Parts are printing as I write this but, not all the parts have been designed. The critical cylindrical bearing housing/trolleys have yet to be designed, spacers for the motor and bed mount aren't even shown yet. However, aluminium is soon to be sourced, some parts have been harvested from my scrapped printers bin, but currently this is a complete unknown.

DESIGN:

The corner brackets and gantry brackets have specific orientation requirements.

The frame corners are specifically designed to optimise the layers for equal strength in all three axes, hence those orient naturally on the face of the corner chamfer. This makes the layers orthogonal to the vectors of each tube as it slides onto the bracket.

The gantry corners are printed upside down to minimise support material. This is also why the corner guides (AKA "the fins") are triangular, to obviate the need for support. There is a 0.1mm clearance from the walls of the frame's verticals to allow room for less friction without adding too much play.

The y-rail "trucks" are yet to be designed, but are basically just something to hold a linear bearing to two linear rods (the x rails), allowing them to move along the y axis. The print bracket, similarly, has not yet been designed, but is another mount for linear bearings and a hotend mount that will slide along the X axis.

The z-axis will be leadscrew driven. 4 motors are shown, but the design could be forked for three screw tramming, with a single motor in the center, at the rear. Probably just an extra set of nema17 motor holes in the floor plate and a bracket on the front facing side. Marling has triple z tramming built in, it just needs enabling in the geometry.

Some of this may change as even the design progresses, never mind at the build stage. Remember, THIS IS A WORK IN PROGRESS!!!

Not only a WIP but, because it has evolved from an intellectual exercise in how to corner join 3 tubes in 3 directions with even and maximum layer strength, I can't promise I'll finish even an initial design. I have a few too many projects in progress right now and, being retired, I am only working on them for pleasure. Don't rush me ;-)

05 SEPTEMBER 2024

Might have simplified and cheapened the spec a little more. I'm thinking the gantry and printer truck could run on F603-MC3 flanged bearings (3mm bore, 9mm OD, 0,7mm flangeheight on one edge) on the actual 20x20x3 (or 202x20x1.2) tubing, with the mounts designed so that the flanges are the stability. This eliminates the linear rods but has the same effect. 3 or 4 point levelling still works in theory, provided the static bed doesn't need more than 1mm tramming height range.

Alternatively, use a pair of F603, unflanged bearings (axially perpendicular) in each corner of the lifting frame and flanged bearings for the gantry (Y) and truck (X) - that way the frame is the linear rail. This way it could be possible to assemble the whole frame and moving tracks from a 6125mm length of 20x20 aluminium square tube.

This should bring down the cost significantly as 1612 linear bearings and the stainless cylindrical rail are likely more expensive than the requisite than the necessary F603 and F603-MC3 bearings alone.

Another cost saving is using 5mm leadscrew for the Z axis lifting, although, with the proposed vertical bearing arrangement, even a timing belt lift might be possible for 3 or 4 axes.

Think, design, rethink, redesign, repeat from "rethink."

ELECTRONICS:

Because of using a 3 motor or 4 motor Z tramming and lifting system, a simple, 5 motor board (E, X, Y, Z1, Z2) won't be sufficient. For a basic, single hotend machine, this printer will need at least six motor drivers, So, I recommend the Big Tree Teck SKR V1.4 control board with the EXP-MOT motor driver expander module. These are a fair slab of of the AU$300 budget at AU$120+ for the two boards, but they'll give you individual control over the usual extruder, X and Y, as well as giving you 3x Z for tramming and a second extruder, or 4x Z and a single extruder. Personally, I'll take 3 way tramming of the gantry and an extra extruder, anyday. A standard SKR board and splitter cables will give you 2 parallel pairs of Y motors, allowing for manual tramming on the X and electronic tramming on the Y, with a 5 motor board on its own. The idea behind the AU$300 budget target is for those with an Ender 3 or similar, they can cannibalise, really.

DESIGN CONTRIBS:

I've added my github repo for the OpenSCAD code. I'll drop a link to that when I've sorted that. If my design makes sense to you and you can improve on anything, you'll be able to open a pull request on that part. This is a 100% open source, free software and designs project. It is, however, commercially restricted. No making money out of printing parts until the design is proven and tested. Again, I MAY change that when it gets to a "golden master."

BRANDING:

The name, "Kyoob", is provisional. I haven't been able to find anybody in the 3D space that has used this for a brand but, if you can prove a registered brand in the 3D space, let me know gently and I'll take the name back to the drawing board. Yes, there is an architectural firm, again, if 3D is in the architectural namespace, let me know gently, I'll change things.