In 2004 I decided to build a CNC router for my shop. In the past I`ve built other machines for the shop. Including a combination wide belt sander | planer and pin router . These machines were built with woodworking and metalworking tools including a south bend lathe, metal cut-off saw and a old drill press.

 Part of the reasons for this build were to construct a machine that was capable of machining wood and metal. The end result provided a machine that capable of milling wood, aluminum and steel. As with all my projects I visited my local surplus store. I managed to find a fabricated aluminum frame comprised of tig welded I beams. This assembly defined the overall size of the machine 36 X 80.

eBay as usual is the place to find used components at a faction of the price for new precision assemblies. I found ball screw assemblies with the bearing blocks and purchased three sets. Only the bearing blocks were used for the machine. The ball screws purchased were XPR inch precision rolled ball screws from Nook Industries. All were 5/8 diameter by .2 pitch. This pitch is very common in the DIY world as it provides an excellent machine resolution.

It should be mentioned that all ball screws have a critical speed. At a certain speed the ball screw starts to shake. At a given ball screw diameter the critical speed issues can be minimized by reducing the maximum ipm of the machine. In my case, a 67 inch screw length produced extreme shake at 250 ipm. Setting the maximum speed of the machine to 200 ipm resolved the problem.

Thomson Ball Bushing BearingI started to mount my Thomson 1 1/4" diameter linear motion rails to the bottom of my frame. This required 48 bolt holes to be drilled in the aluminum frame. I secured one rail and then set up a measuring system to check the other X axis rail for parallel. I mounted a dial indicator to an aluminum extrusion and linked it to two Thomson blocks. I adjusted the blocks as tightly as possible to eliminate any clearances. After about 5 hours I managed to position the two X rails within .002" of parallel. As a test, I mounted a 15" x 36" piece of MDF to the four sliding blocks and pushed the new assembly along the full length of the X axis to test for any binding. It seemed fine but I did push one side and the complete assembly twisted a bit. As a design note: it was recommended by Thomson Industries that the spacing of the sliding blocks should be set to match the width of the machine. To clarify, my machine is 36" wide, so the spacing of the blocks should be 36 inches apart. If I had followed these guidelines I would have reduced the usable travel of the machine considerably. So I compromised and set the spacing to 15 inches.

I then went to work designing the gantry. I decided on 4 inch structural tubing with a wall thickness of 3/16. I had thought about using aluminum extrusions. In my searches I never found a size close to 4 x 4. Aluminum would have been easier to work and considerably lighter then the steel. It was my belief that a larger tube would allow me to space the THK sliding elements farther apart, making a more rigid design. After welding of gantry ends were completed. I sent the gantry weldments off to a machinist to mill the top and bottom plates. This provided a reliable reference surface to mount the Y axis tube to gantry uprights. The top tube shown here was also sent to a machine shop which had a large Blanchard grinding machine. All four surfaces of the Y axis tube were ground. This cost $200 and provided a very flat surface to mount the THK linear rails.

 My next step was to mount, 4 THK linear rails and blocks. I planned to butt joint the rails end to end to achieve the needed length. It is standard practice to use full length rails for CNC construction. I managed to install the butt jointed rails without any problems. As a design note: it’s a good idea to find engineering information regarding the strength of materials. It’s very surprising to find structural steel elements will deflect under load more than you might expect. The 4 X 4 tube had no noticeable defection taking into consideration the amount of weight the Z axis applied to the Y axis tube.

The next step was designing the Z axis assembly. It was my intent to use a spindle designed for metal instead of the typical wood router. I used a spindle purchased from the Little Machine Shop. This spindle allowed me to use R8 collets for all my cutters. I needed to drive the spindle and decided on a Baldor 1 HP DC motor. This motor system is designed to be controlled by a electronic speed control. This allows infinite speed control of the spindle. Something I felt I needed if I was going to machine steel.

Since this spindle was belt driven, it required a sliding motor mount to allow tensioning of the pulleys. The picture below shows a THK wide body slide which is secured to a black painted aluminum sub-plate. This functions as a sliding motor mount. To manually take up the belt tension a surplus acme screw was used. I did have a design oversight with this spindle. I originally setup a large timing pulley on the motor and a smaller pulley on the spindle. This configuration worked fine for my woodworking but failed in my attempts with steel.

By woodworking standards the spindle speed is considered slow at 4500 rpm. In all my woodworking tests I found this speed worked just fine. Another added benefit was the very low volume of noise the spindle produced. 4500 rpm was the highest speed that I could run the spindle. Being that the spindle was designed for metalworking and considering the bearing type. It was not possible to gear the spindle to match the speeds of a wood router. The bearing type and amount of preload on the bearings would have severely overheated the spindle.

In my attempts with steel I found that the spindle had almost no torque when it was reduced in speed to 800 rpm. This was of course the result of the original pulley sizing. Taking a 1 HP motor geared for a maximum speed of 4500 and running at a lower input speed via the electronic speed control negated any horsepower to the spindle. The solution to this problem was to incorporate a two step pulley system. I purchased a pulley set from The Little Machine Shop and used the supplied double step pulley for the spindle. I then machined an aluminum double step pulley for the motor that allowed for better torque transmission at high and low speeds. These extra steps did resolve my problem.

Here are the results of the spindle modifications. The picture below (far left):YouTube video of machining shows the machine at work milling the new cast-iron work surface.A few years ago I added this cast iron table to the end the machine table. This addition to the machine was needed to provide a rigid base for machining metal.

The center picture shows the machine table complete with the machinist vise mounted to the newly installed cast-iron work surface.