CNC Router : Base
I will try and go through my build as well as possible in the order of construction,
although it was over a year ago that I built it now so If I miss anything that seems
important post a comment or send me an email with any questions or suggestions.
The very first thing you must consider when building a cnc machine is how big do you really need it, think of what you plan to use the machine for and the availability of materials and plan accordingly. I did not do this, my original motivation was to be able to construct custom cases for electronics (I still haven't made any yet) but I figured bigger was better and made a nice 1m x 1.5m base. Something this size is nice for 3D wood carving but remember if you are planning on building it on the cheap it will likely be fairly slow.
CNC Base With Dimensions
For my machines construction I chose 25x25mm aluminum SHS because I work part time in a hardware store and could get it quick and cheap, I also liked the idea that aluminium will not rust. Without a tig welder my options for construction were fairly limited so I decided to bolt the entire frame together. So I started to drill and tap a bunch of 1/4" holes, with 44 50mm right angled brackets and 176 1/4" gutter bolts I ended up with the construction I have modeled above. Unless you are some sort of wizard, to get any decent accuracy/consistency with something like this you have to make some jigs to help line up all the holes, I took a photo of all the jigs I used for this project because I have no idea how to describe them.. they are just jigs, scraps of screwed together mdf with holes in them.. One nice advantage of the soft aluminium SHS is that I could attach the tap to my drill for speedy tapping (I know tapping under power is bad form but it worked and saved me lots of time).
How to Make it Better
The main problem with this base design is the deflection under load, as I can not use fully supported rails with
the homebrew linear slides I had in mind the only way to overcome this would have been to have a shorter rail,
a different rail cross section (I used SHS) with a higher moment of inertia (I'll explain this shortly) or a material
with higher modulus of elasticity such as steel (E = 210Gpa). Of course in hindsight the best solution would have been
to use fully supported rails which I will discuss in the next section.
The easiest way to do some very basic and very rough analysis to get an idea of the deflection of the base is to model it as a simply supported beam as I have done in the image below. From the equation for maximum deflection below you can see that by increasing the polar moment of inertia (I) you can reduce the deflection, the polar moment of inertia is defined as an objects ability to resist torsion. For example a round solid rod has a low moment of inertia along its length where an I-beam has a high moment of inertia, just imagine trying to bend each of them, the rod will be much easier to bend. For the analysis I have used basic properties of aluminium and assumed that the load is 10kg distributed equally across each of the rails, which from memory is fairly close to the weight of the gantry section.
The equations get me a value of a 3mm deflection when the gantry is at the centre position, remember very rough (but actually pretty close from my observations) but the point is that if I had done this analysis before I built it I would have changed my design as 3mm is a big number if you are trying to achieve high accuracy. Fortunately most of the time the objects I have made are fairly small compared to the machine's size and the router blade actually follows the deflection, so far deflection has caused no notable errors.