I have no personal experience of building and using a metal-frame CNC router. However, I do have a "working" MDF machine in my garage, built to a slightly modified version of the open-source JGRO design (Google will give you plenty of references). I have indeed used it to cut some useful items. After 6 months or so, I have decided that I need to replace it with a decent metal-framed machine. Does that tell you something?



I've learnt a lot:
- I have a box of electronics that works well and is sized for a bigger and more powerful machine, complete with steppers also reusable.
- I've had a lot of practice with CAD and toolpath-generation software. Free does not necessarily equal best...
- A 2KW water-cooled spindle provides much more cutting power than the machine structure could possibly cope with. And its weight is more than MDF can cope with, as well. See below.
- A home-made anti-backlash delrin nut assembly works fine. I can't say the same about M10 stainless rod used as a leadscrew. Not sure what the theoretical critical speed is, but in practice it works out at about 800mm/min for the X axis.
- A "rapid" feed that is slower than most people's cutting speed isn't too important when the max cutting speed is (depending on cutter, etc) around 200-300mm/min. At this point, I hear the experts laughing, holding their sides, and saying, "I told you so..."
- You can get an enormous amount of satisfaction out of building something that works. I've cut a presentation plaque in hardwood, engraved some Christmas presents very acceptably, and even sold a few small items. I just had to do it very, very slowly, and with a lot of shimming of spoil boards to get something to approximate a level bed. I even improved the original design by replacing some MDF components with pieces made on my 3D printer (also home-built) but this just shows up deficiencies elsewhere.
- MDF is not the most stable constructional material. Actually, it's not stable at all. Or even suitable as a constructional material. OK, I should have varnished or painted it, but I doubt that it would have made much difference. Just before Christmas the bed had a sag in the middle of around 3mm in 300mm, and with the recent high humidity it's probably worse than that and still going. My son has built a very similar machine and stiffened the bed with a couple of lengths of 1" square steel tube which has helped, but the Z axis plate is now a few degrees off vertical due to MDF sagging under the weight of the spindle.

There are people who have built a JGRO design more or less as per the plans and used it for some fairly serious work. Well, not in my local micro-climate! On the other hand, if I reuse all the steppers, control electronics, spindle motor and VFD, etc, I shall be throwing away less than £100 of materials, mainly MDF and steel pipes used as linear bearings. Even the skate bearings for the three axes altogether probably cost less than a tenner off Ebay. I always had the thought in mind that this would teach me what I, personally, could do with a machine like this and that the first machine would have to be considered disposable.

I am currently putting together an outline design for a slightly larger machine, to be steel framed with decent bearing rails and ballscrews, and this will be presented to the community for comment. In particular, I have some idea of what is possible with a somewhat crude and flexible machine, and I don't want to go the other way and grossly over-build. But if you want a sensible size machine, don't think about wood. Ply might be better than MDF, but for a machine for commercial purposes in any size? Don't go there!