The corner brace design I found on Thingiverse was too weak for 25mm extrusion. After assembling the frame the extrusion could flex the plastic touching it.

I took the design into Sketchup and adding more support. I’ve added the STL to a new GitHub repository I have started using for the pieces I redesign.

The new brace design adds support on the bottom of the extrusion to allow for 3 additional mounting points. This decreases the amount of flex present in the frame.

Side shot (The bottom holes needed to be reamed a bit because I sized them too small. The other pieces are sized correctly):

A shot beside the old design

After ~14 days worth of printing with my Rostock Max v2, I’ve come across issues with being able to print the entire length of the heated bed. Delta printers use triangles and math to figure out how to move the three carriages in a coordinated way to move the effector platform around the bed. If the triangles get messed up, the math doesn’t work correctly.

This issue is well documented in a SeeMeCNC thread where multiple users are having the same issue.
tl;dr Moving the effector platform outside of the triangle formed by the three towers results in a loss of precision while trying to print in this area. The issue has been known for some time (the original thread was posted in April 2014) but no silver bullets have been discovered. Lots of different areas of the printer can affect it’s printing precision. Incorrectly placed towers, uneven arm length, etc can add up and cause inaccuracies in the finished print.

The issue of precision outside of the triangle is visualized below from an image in that thread:

The main reason I purchased a Rostock Max v2 to begin with was to have a 280mm circular build platform the Onyx Heated Bed offers. With part of the heated bed outside of the triangle, printing to this area has been near impossible. Through countless hours of calibration the nozzle raises ~1mm above the bed in some places while driving into the bed in others. There are multiple attempts to get the printer working reliably to the outside perimeter of the bed, but none have been too successful. For me, the easiest way to resolve this issue seems to be increasing the size of the triangle by increasing the distance between the towers. This can’t be done with a static melamine board that the Rostock Max uses as its frame though. I decided to look for alternative delta printer plans that could give me this flexibility.

I found the Kossel Mini promising. While the original plans are for a much smaller printer (and only 15mm extruded aluminum parts), scaling it up to encompass the entire printed bed the Onyx has to offer would be simple. I found a set of STLs on Thingiverse that was already altered to work with larger extrusion.

I’ve decided to use 25mm aluminum to be able to re-use the towers from the Rostock Max v2. I have managed to get my hands on 24 feet of 25mm for free from a family member 🙂

Using some calculations it looks like I will also be able to re-use the effector arms as well. The optimal length of the arms should only be ~20mm larger than they currently are.

Here’s what I’ve got so far, the printed parts for the base and the top, and the extrusion cut to size. The extrusions will be 23.5″ each. I originally planned for 24″, but my hack saw skills left the ends not square. After switching to a miter saw and squaring up the ends I was left with 23.5″.

Here is one of the base parts being printed:

And the base laid out ready for nuts and bolts.

The extrusion I am using has slightly larger spacing on each side. I need to find a nut/bolt solution to this or resort to ordering some from Misumi.