Just made my first cut with my self designed self build CNC router.
Mostly uses openbuild parts. But also laser cut 6mm wood. A few 3d printed parts. It can cut up to at least 120x120cm. But only if i put an extra extension desk in front of it.
Spend less then 500 euro so far, of which a large chunk is the router itself. But the wooden parts and 3d printed parts where free. As well as the motors and control electronics (discarded Ultimaker 2 parts, working at Ultimaker has it’s advantages)
It has a bit of flex in the Y direction, the Z stage tilts a bit due to the wood flexing. So I’ll most likely replace those plates with alu. Prototyping with wood is so much cheaper 
The left Y guide rail is bolted to the wall. The right side has no guide rail and just rides along on the table, but still has a belt for accurate positioning.
This is how it looks in CAD, with 250mm X/Y axis instead of the 1500mm ones that I used in the build. (Makes it easier to get an overview of the machine with shorter axis)
missing/deleted image from Google+
+1 for the super weird setup: one side with a sturdy wall-mounted rail, one with a sloppy but functional belt clamped on a drawer handle 
at least it does not get in your way for now. Lay a ruler on the bench and the other side on the back of a chair and you get a super wide cnc
No offence but in this setup I don’t see anything accurate. Sorry I don’t understand this kind of setups.
@George_Novtekov thanks for that “insightful” comment. How do you define “anything accurate”, 1mm accuracy? 0.1mm? 0.01mm? Also, which part do you think causes these inaccuracies?
I think the circle has about 0.5mm inaccuracy, but that’s without any tuning. As well as a few fixable design flaws. I haven’t measured it yet. As I don’t care too much, I can see it’s not a perfect circle. But that’s also down to the amount of speed an acceleration I threw at it.
My goals are pretty clear for me. And I know this doesn’t fit everyone. My goals are wood, 2.5D, plate materials, mostly cases/panels to fit projects in. 0.5mm accuracy would be good enough for that. Not want to spend too much money. Don’t need to cut any metals (if any project calls for that, I have a machine at the office that can do that)
As I have access to 3D printers, I don’t have the need to create complex precise shapes, as I can print those a lot better.
But most important, this machine only takes up a tiny amount of space when not in use. Which is important for my cramped garage.
@Jeremie_Francois Yes, that’s the point, I can extend it quite easily without getting in the way when not in use. You cannot see it on the photo, but the garage is so full, that there is only 1 spot to stand in front of the machine. Still cleaning up.
Idea came when I saw this machine: http://printrbot.com/project/crawlbot/
However, the problem with that, is that you X alignment depends on clamping the sides of the table. And then you can all kinds of complex spring loaded wheels to adjust for inaccuracies. Which upped the part count quite quickly.
Which is why I came up with the hybrid between that and a more traditional machine.
Cura for CNC confirmed!
(just kidding)
looks like a fun build man. You’re gonna want double stacks on the X/Y axis at least, but I’d suggest Nema 23.
How’s the noise when cutting? My MPCNC screamed like a banshee because it wasn’t very rigid
Well, I do have a Cura for milling prototype, but it’s a huge mess. Made it before I got some more experience with milling on our shopbot with v-carve.
But first, I’m playing with making my own firmware.
I could swap for NEMA23 quite easy if they give me issues. But these are quite beefy NEMA17s, at 44Ncm (Ultimaker 2 feeder motors). And I’m not even running them at full power yet. Best thing, they where free 
In my experience, our shotbot would scream like a banshee if the chip-rate was way too low. Lowering spindle speed or increasing feedrate helped there.
But I’ve only done this one cut, so I’m not sure on the noise levels yet. Didn’t have issues with the noise with protective gear on.
Oh, and what are “double stacks”?
They sometimes refer to Nema motor sizes by stacks. A “typical” motor of a given standard (14,17,23, etc) will be considered as 1 stack motor, IE a single winding stack around the rotor. For a double stack they may reuse the stack twice in a longer body.
Ah, ok, these are already longer then usual motors. 47mm vs our normal 38mm. And in that size, these are the strongest we could find (guess what, there is quite some difference between manufacturers)
As I have access to NEMA17 motors of 4 different lengths, I find the “stack” naming confusing If I would go by the smallest one as 1 stack, these would be quad-stack.
I’m a bit more scared for the Z motor. As that has the spindle attached to it directly, so no way to beef that one up.
