Cobalt bits and a drill press.
1 Like
I agree it might be better, except that I built an enclosure for my MPCNC and it would no longer fit if I would put the rack and the track on the outside… The track needs about 6.5cm of space next to the tube. I might be able to shave off one cm or so, but I only have 3.5cm in my enclosure.
I don’t think much dust will stick to the belt in the current design though. The only time the teeth are facing up is when it is touching the rack. In the second image of my first post you see that the teeth are facing downwards actually, so any dust or chips will fall off there. Besides, most of the dust blowing at the track coming from the workspace will be shielded by plastic. The plastic is wider than the belt for this reason, and also to keep the belt from running off the bearings.
2 Likes
Check out Youtube videos of guys making those Epoxy River tables and stuff like that.
They take a propane torch and run it over the top of the epoxy to get the air bubbles out.
No idea if it works, but I’ve seen many different people using this method.
That pops the surface bubbles, but doesn’t always help with the ones deeper down inside.
Warming the two parts of the epoxy in a hot water bath before mixing can help with bubbles. Vibrating the fixture after the pour can also help bring bubbles to the surface.
I finished and tested iteration two of the roller design. Here are the files if anyone is interested: rack_roller.zip (229.6 KB).
I included an OpenSCAD model of the whole assembly, which looks like this:
To build one roller, you need to print one instance of each STL file. (Except frame_tube_clamp.stl, you need two of these to mount one rack on a frame tube.) Print time is a little longer than the standard roller plus stepper mount, but still in the same ballpark.
Furthermore, you need the following hardware for each roller:
- 1x GT2 6mm 280mm closed loop belt
- 7x 608 bearings
- 2x M8x30 hex bolt
- 1x M8x40 hex bolt (or longer)
- 4x M8x50 hex bolt (or longer)
- 7x M8 nyloc nuts
- 4x M3x10 bolts (optionally with M3 washers) for mounting the stepper
- 5x M4x20 bolts plus M4 nuts (one for the tensioning, four for mounting the gantry tube clamp)
- 4x M4x20 bolts plus M4 nuts to mount the rack to the frame tube
I think assembly is pretty self explanatory. Putting the belt on can be a little tricky. It helps to insert the three bolts in the thin motor plate, then put the bearings on, and lay it all down with the bolt heads on the table. Put the belt on, and hold it such that there is a loop where the stepper pulley has to be. Then slide the stepper plus thick motor plate on and wiggle the belt around the pulley. Then use the M4 screw on the side to put some tension on the belt.
There are four holes (two on each side) on the roller for mounting other stuff (I use it for a cable chain). They are not used otherwise, so don’t worry if you don’t know what to do with these holes.
The distance from the frame tube to the gantry tube is the same as for a standard roller, so you can experiment with this by making only one rack and roller and leaving belts on all the other ones.
I considered putting it on thingyverse, but I don’t like that site a lot… besides it’s all still a bit experimental.
1 Like
I also made some measurements. I mounted the epoxy rack and a version 2 roller on one side and left the standard roller with a belt on the other side. I then hung two different loads on the end mill (placed in the center of the workspace) and measured the deflection of the rollers.
Long story short, the epoxy rack and roller seems to be roughly as stiff as a belt when the roller is about 20 cm (8") from the corner.
My measurements implied a stretch of about 0.200mm per kg per meter for the belt, which is a bit low. Earlier measurements by myself and Jamie put that number at around 0.320 to 0.350mm. I measured both sides under the exact same circumstances though, so I think my conclusion of 20 cm is valid.
When the gantry is in the center, the epoxy rack only deflects 0.025mm per kg of load on the end mill. (Note that this number may be up to 0.050mm if my measurement method is wrong and I underestimated roller movement by a factor of 2. Also note that the gantry in the center is not the worst case. Worst case estimate would be about 0.100mm per kg).
1 Like
I printed the roller STL… only to find that my 23.5mm EMT was too loose to tighten. What OD is the tubing you used for all your rails?
EDIT: I imported the mesh into Onshape and measured 25mm for the gantry tube. Sound right??
1 Like
Yes it’s for 25mm OD. Sorry I forgot to mention that! It’s not too hard to make the model parametric, but I’d have to check all places where it depends on the OD value. I did not only partly so far…
Btw, I’m also working on a kind of double belt setup, where the rack is another belt. Instead of glueing it though, it is held in place with plastic parts. Seems to work really well so far.
1 Like
Thank you, Peter. Parametric would be nice but please don’t make the effort just for me. My CAD skills are sufficient to scale and alter parts when/if needed… and I printed your parts primarily just to check some fits and clearances. I plan to pursue the printed rack and pinion setup I’ve messed with previously – that vaguely resembles a flipped version of the roller configuration in your first photos – but without any epoxy, glue, or belts. So, rather than a design to actually duplicate exactly, your thread has served primarily to spur interest again in r&p ideas I’d already entertained when building a different machine… applied this time to MPCNC. Please try to understand where I’m coming from… I mean no offense and I’m very appreciative of your efforts. Thank you.
2 Likes
No worries!
1 Like