CNC Router/Mill for Wood/Aluminum/Steel - Journey from Craptacular to Great

Four years ago I was messing around with a stepper motor I had and was wondering if I could build a CNC machine from it. Started doing some research found this forum and realized not only was it possible, but fairly common and looked pretty easy. I didn't have a real goal in mind, just was playing around, so I didn't want to spend any real money on it. I ended up slapping together a handful of cheap stepper motors with some belts, and a frame made out of 2x4s. I had some old drawer slides laying around and I used those with an Arduino running GRBL to come up with the first version of my CNC.

All told I ended up using mostly spare parts laying around and spent less than $100 to get something that was kinda sorta functional! I was surprised at how well it work for how little I had in it. The biggest issue was really the dust getting in the drawer slides. Besides for that it was reasonably accurate and plenty repeatable enough for wood working. I made a bunch of little carvings and other random projects with it. I think it was a great introduction, and I learned a ton doing it. It was a great introduction to get some hands on experience without spending any real money.

https://www.cnczone.com/forums/attac...d=453254&stc=1

I used the for a year or so making minor improvements to it here and there before I decided that it was worth spending some money on and improving it. This went on for the next four years slowly improving and redesigning this and that until finally today I'm at a point where I can make real aluminum parts in a reasonable time frame holding ~0.005" tolerances.
https://www.cnczone.com/forums/attac...d=453260&stc=1

This is a manifold I finished up the other day:
https://www.cnczone.com/forums/attac...d=453256&stc=1

I thought I'd like to share my journey and show a bunch of the iterations. Might help someone else avoid some of the pitfalls I made, or skip some of the intermediate steps. I'm also planning a big upgrade in the future, which I'm just starting designing now. Here's a teaser:
https://www.cnczone.com/forums/attac...d=453258&stc=1

I love it...

Instead of falling into analysis paralysis, you decided to just build it and then tried it out to see what it could do. Sounds like you had fun and learned a lot in the process. That is how many new innovations are born.

Keep on keeping on.

Sent from my SM-G973U using Tapatalk

I hope some of the daydreamers who have yet to make the leap into machine building see this thread and take note.Its good to see ample triangulation playing it's part.

Quote:

---

Originally Posted by routalot

I hope some of the daydreamers who have yet to make the leap into machine building see this thread and take note.Its good to see ample triangulation playing it's part.

---

I think thats the biggest takeaway from this stage. Rather than spending endless time stuck in design iterations, or lamenting the high cost of precision rexroth linear guides, just go build something and see how it goes.

That was the 1.0 build, I quickly made a few changes to it. The biggest ones were to add some skateboard bearings that preloaded the X-axis and eliminated alot of the freeplay in the drawer slides and changed the orientation of the drawer slides to be able to take advantage of their in plane rigidity. I see a CNC as combined of 4 major components. The Spindle, the Linear motion, the frame and the controller.

Spindle: I used a DW611 1.25HP router. This is a pretty common choice for small CNC routers, so there is alot of good info out there on it. It seems to hold up pretty well to continual use, particularly if you don't run it at its full speed. The bearings seem pretty good and at least on mine there is very little runout. It does run at a pretty high speed, which is good for small bits, but it poses some challenges for cutting aluminum as the torque drops off at lower RPMs. I'm also learning now that the 1.25HP estimate is a bit generous, but that will come later as this 1.0 design certainly can't take advantage of even 1/10th of that power.

Linear Motion: I used drawer slides for the bearings here, obviously not the ideal choice, but they are dirt cheap. I found that they are fairly rigid in the in plane direction, but are quite floppy out of plane. You could make up for this using two pairs rotated 90 degrees, however as long as they are oriented in the right direction its not too bad. For driving it, cheap amazon steppers work great, and put out way more power than is needed. Toothed belts work well, but can be very difficult to tighten enough to eliminate backlash.

Frame: This is the area where there is the most freedom and design choices. That can make it one of the most challenging portions to design. I think sometimes folks get stuck in analysis paralysis when trying to design a frame. I slapped a bunch of scrap wood together for this, and when compared to the rigidity of the drawer slides it was plenty sufficient. Following some general guidelines, like larger cross sections are better than stiffer smaller ones, more mass is more better and lots of cross bracing and its very easy to make a frame that is going to be sufficient for your needs.

Controller: I used an Arduino Uno running GRBL fed by BCNC on an old laptop. I'm shocked at how well this worked. I spent a long time looking at very expensive controllers, or cheap software that required expensive interface boards. It did take quite a bit of fiddling and learning how to set everything up, but its not too bad. until you start getting into closed loop or absolute encoders I don't really see any reason to go beyond this.

I did make a video on this iteration a few years back:
https://www.youtube.com/watch?v=k32B6ufuBwY

I used that setup for a year or two making little improvements here and there. As long as cutting forces were kept low, it worked fairly well, the biggest issue was that saw dust would get in the slides and jam them. I came up with many iterations of guards to prevent that, but was never really successful. Eventually I decided that I needed a larger work envelope and that I needed to improve the linear motion. I rebuilt the frame, once again using wood, but I made it somewhat larger and a little more substantial.

I still wasn't willing to spend real money on this project, so I didn't want to shell out for some commercial linear bearings, even cheap ones. I came up with three different solutions for the X, Y and Z axis. For the Y axis, I went with a fairly conventional solution, of using extruded aluminum rails with ball bearings riding on it. In this case, I used the V-slot rails from OpenBuilds. I wanted to use a bearing setup so that there wouldn't be any chance of it rocking from one bearing to another and causing some backlash issues. This didn't turn out to well, but that's a topic for the next post!
https://www.cnczone.com/forums/attac...d=453558&stc=1

For the Z Axis I had a large aluminum plate that I was using and thought I could make up a good linear bearing by mounting bearings around it. I took some wood, and cut slots in it and drilled some holes to mount the bearings in. I then sandwiched this around the aluminum plate. I used a screw to be able to adjust how the bearings clamped up on the edges of the plate, and the screws clamping the two plates together ensured that it was tight front to back. This gave me a nice rigid solution that was very inexpensive. The large size of the bearing plates did limit the Z range somewhat, but that wasn't much of an issue. A thicker aluminum plate would have worked better as this 1/4" plate did flex some out of plane under load.

https://www.cnczone.com/forums/attac...d=453564&stc=1
https://www.cnczone.com/forums/attac...d=453562&stc=1

For the X-Axis I needed something ~4' long so I decided to take making my own linear bearings. I got a piece of 2" steel electrical conduit and created some bearing holders to ride on it. This worked out very well. I used the CNC to machine blocks out of wood that I could mount some roller-skate bearings in. The conduit is quite straight and reasonably consistent in diameter. I did sand it lightly to smooth out a couple of rough spots. The bearing carriers had two features of note. The first was that they had a stiff spring to close them, which allowed them to flex slightly and absorb any changes in diameter. The springs were stiff enough that the small cutting forces would not deflect them substantially. The second feature was to install some foam brushes on the top of the carriages to sweep off the saw dust that would accumulate. These actually worked amazingly well and I never had any issues with sawdust jamming them. This solution worked quite well. The large diameter steel tubes are quite rigid allowing for much higher cutting forces and are consistent enough to make accurate parts. I would recommend it to someone looking for some linear bearings on the cheap.
https://www.cnczone.com/forums/attac...d=453560&stc=1

At this point I've got a machine that is working quite well. Its rigid enough now to easily machine wood without having to worry much about babying the feed rates. The limiting factor at this point is the drive system, which is still using 5mm toothed belts. I built a couple of clamps to better tighten the belts, but they just aren't stiff enough. The Y-axis also doesn't work as well as I would like. However even with those limitations, I am able to make a few aluminum parts, albeit very slowly.

https://www.cnczone.com/forums/attac...d=453566&stc=1