I know there are plans to build your own, but I’m not feeling like building it (this time).
Most of the 4th axis units on Amazon use Nema 23 and cost about $300 and up, but this one uses Nema 17 and cost about $160 including chuck and tailstock. Seems like a good deal to me.
I am not sure exactly what I’ll be doing yet but I am thinking maybe puzzle pieces like small versions of this. Maybe start with a stick and carve the perimeter of several different pieces and then separate them on a band saw.
It’s possible that the puzzle pieces might require “true” 4 axis (X, Y, Z, A). I suspect the puzzle pieces will have undercuts that are not reachable with the usual 2 + rotary (X, Z, A). That will keep it interesting.
Seems like we could have a “mostly printed” 4th axis, but the software / tooling is probably going to be the hard part. I’ll be interested in what you figure out there.
Bought the same one a while back. Put it on my old 3018. I didn’t set up as fourth axis at first but set it as y axis lined up with x axis under spindle. Works well, but haven’t used it much in a while. Goal was to make chess pieces…
If a pre-processing step unwraps an object to a 3D surface then Estlcam could process the topographic map while being oblivious to the fact that it’s a rotary axis and Y is actually A. I think that’s a fairly common way to do rotary 3-axis (X, Z, A).
Another option is basically multi-sided milling where the object is rotated into a discrete number of rotary orientations and in each orientation milling proceeds as a 3D surface. Estlcam can do each of the orientations with a rotated model, but it’s not streamlined to do it easily.
These parts could be made on a small lathe. But I have always wanted this game. I only played it once and it was really fun. The top (making any top would be fun) and the pins could be made on a machine like this:
I was using grblgru. If i remember correctly can do up to 5 axis. Just finished debugging a 7 axis welding robot cell at work, but actually enjoyed playing with my little one, not sure if it qualifies as a lathe or mill… lol
I looked through the manual and it’s not clear whether it can take a 3D model and generate a 3-axis or 4-axis rotary-axis tool path from it. The videos on the GrblGru channel (or at least the previews) show symmetrical chess pieces like pawns but no examples of a knight.
As a starting point I would consider it if it does CAM from unwrapping a solid into X, Z, A, and better yet if it does more axes.
…
After some digging I found a post that says that GrblGru does not do that kind of 4-axis work.
JamieK, I have been down this research road the last year or so. The test 3d models I have are for a Lewis Chess set.
The free programs that are available are rather limited, or unable, in their ability to “unwrap” a 3d model/ There were several attempts at a program to do this that was open source, but all of those have not been updated in many years, and are less than ideal.
In the low cost level for a program to do the unwrap, CamBam has a “Unwrap” plugin that can take the 3d STL model and unwrap it to a flat STL model. Unfortunately CamBam doesn’t have a Post Processor for Marlin, but does have one for GRBL for tool paths. It can also do full rotary axis machining. I have been researching a PostP for CamBam, but have some other things in the mix of life/work/hobby, so that has been slow. The PostP format for CamBam is a bit different.
Example base file:
Thanks for the suggestions. Looks like Snapmaker Luban might also have this ability and it’s open source.
In all cases I’ve seen so far, it looks like “4 axis” is really 3-axis with X, Z, and rotary axis, but that’s okay. I had a feeling that true 4-axis was going to be rare anyway.
I’m sad there wasn’t something more via software yet.
In my mind, without software to support, all I kept reverting to, is the rotor tool, being possibly better than the manual rotation that Dennis Van Hoof did for some of his projects.
I was watching and hoping for more. If nothing else, I was hoping for a way for current methods to use the rotary tool 1/4 when needed, to match what he did manually, so as to allow the roughing bit to do all four sides, before requiring the operator to do anything.
I know that wasn’t the goal of this thread, but that would have been one small step in the right direction (and least in my mind).
Grblgru can easily wrap SVG files around curved forms unlike a whole lot of expensive programs out there. It’s also free. In this video I created a series of overlapping circles in another program. I imported the SVG into Grblgru then ran the GCode it created. I wanted to carve the graphic 0.5mm deep so any depth errors would be obvious. I’m also using a fifth axis I created. Grblgru kindly added the real math necessary to keep the cutter perpendicular or normal to the surface. This is using true 4th and 5th axes acting with the linear axes using G93. I had to restart the cut since I was out about 1mm from my X starting point.
Control card is a Mega 2560. Grblgru is updating his program to run with 32 bit cards too. I’m testing an STM32 black pill for him. I’m also trying an ESP32 running some version of ESP Grbl by Bart Dring I suspect. It doesn’t have backlash though unlike GrblHal for the STM32.
He hasn’t quite gotten to cutting 3D models like the chess pieces mentioned above but I think it is in the future plans. He’s also splitting the program into a CNC version and a lathe version.
Cheers
Was just checking it out and then went to his videos. This satisfies some things around here. The table, and the 4th axis. Both seem very simple. The table, the cnc would need to be protected, but wow! Shop Work: Making a DIY CNC Lathe - YouTube
.
