The Base; Y Axis Mount
5 Bears Home Homebrew CNC bench mill

It is important, with any machine tool, to build plumb. Right angles need to be 90 degrees. Parallel surfaces need to be parallel. As much as possible, one should keep in mind that the ability to adjust fixed members, aligning them true during final assembly, is very desirable. This is where the Bosch structural members shine. Given that many parts of the mill are connected to other portions via T-nuts and slots, it is a simple matter to make minor adjustments prior to torquing down for the final time.

Recirculating linear ways and rails, and ball screws in particular, do not take kindly to misalignment. It is very common to torque down a pair of rails that you are sure are parallel and true, only to find a bit of stiffness due to a minute divergence. Misalignment with ball screws will cause adverse wear, and increase torque requirements. Therefore, take care to construct a base which is both rigid and true, with a column at the correct 90 degrees, the Y-axis rail mounts parallel and in the same plane, etc.

It is critical that you produce certain portions of the mill with accuracy, specifically anywhere you are going to mount either a linear rail and block assembly, or a ballscrew. For example, the XY saddle "sandwich" which is the heart of the mill requires great care. Ground iron plate would be best, but I have no grinder, so I chose aluminum. When machining plate, the stock should be clamped to the vertical mill table for face milling rather than being gripped in a vise... even the best vises have some "lift", and with plate, you will likely produce a surface not parallel to its other side. A table carriage not coplanar will transmit (and magnify) its errors into the work table, meaning the table itself will be canted. Ultimately shims may be used in certain circumstances, but try to avoid that!


Front and Back Plates, face milled true, and bored/drilled/tapped.
The first items produced were these front and back plates. The front plate is a 12" X 5" piece of 1/2" thick aluminum extruded flat, while the rear is 12" X 6" X 0.75". 1" thickness plate for the rear would be better, as the rear plate provides most of the support for the Z-axis column.

I found to my chagrin that the aluminum flat stock, as extruded, is neither flat nor uniform in thickness. To make the base of the mill true, I needed to face mill the plates to uniform thickness. This was done with a sweet ebay purchase, a Lovejoy face mill of 2" diameter. Any face mill will do, but between 1.5" and 2.5" is ideal. The idea is to plane down the extruded flat so that it is within .002" (.001" is better; .0005" is possible with a good Vmill and face cutter). As purchased, the extruded flat is no better than +/- .005" in thickness. Variations in flatness will warp the critical Y-axis "nest" which we will make from 2 pieces of Bosch aluminum.

While mounted for machining, the pates were drilled for 4 ea 12mm bolts which will hold the Bosch 4590H members. The back plate (rear in photo) is also drilled for 6 ea 5/16" SHCS, which will secure the vertical column. The front plate was drilled and bored for the ball screw bearing housing, and tapped for the servo motor mount. All drawings were done with QuickCAD, a cheap but powerful 2D CAD program.

Two pieces of mild steel T-stock form the junction between the back plate and the vertical column. Six 5/16" cap screws do the honors. Note that the column will be adjustable vertically as much as required.

Note also the 4 holes drilled outside the T-nuts - these accept the 12mm cap screws for the Bosch 4590H supports.

Temporarily put together to show the idea. The four small holes at the corner of the back plate (the front plate has the same) are 2 ea. 1/4" x 20 tappings. These can accept blocks for leveling feet, or stiff extensions parallel to the X-axis to provide additional stability.

After I discovered that extruded aluminum flats aren't really very true, I began to use a type of aluminum plate called mic6 jig and fixture plate available from Metal Express. These guys are great, give them a try. The mic6 plate is reasonably priced, and it is accurately cut. Best of all, from what I have measured, the thickness is uniform to within .001" or better over a large area.

Here, a 3/8" mic6 plate forms the lower base plate for the mill. Eight 1/4 X 20 SHCS secure the plate to the 4590H extrusions in their T-slots. The 4590 extrusions are secured between the front and back plates, and form the platform for the Y-axis rails. Thus, the extrusions are best milled dead parallel, in place, for the Y-rail mounting. The first step, not shown, is to mill one side of each extrusion, taking off perhaps 0.005" - make it a skim cut, just enough to true. Then, the two extrusions are temporarily assembled with the front plate, back plate, and the mic6 bottom plate to form a box.

Seen here is the bottom plate being secured. A 90mm spacer is placed between the 4590's to simulate the column width. The assembly is clamped with a wood-working clamp. When everything is true, the back plate alone is torqued, and the bottom plate also firmly fixed. Then, the back plate is removed.

What we now see is the base of the mill mounted on my larger vertical mill for a truing cut, minus both front and rear end plates. The large mic6 base plate is clamped to my big mill; it in turn holds the two 4590H extrusions via the eight 1/4-20 SHCS. Plenty rigid to mill these extrusions with no chatter or flex, with the two longitudinal members perfectly parallel to each other.

All that remains now is to skim cut the upper surface. The upper surface is where the Y-axis rails will be mounted, and it is critical that the upper surface be parallel to the bottom plate, and that the two extrusions be milled to exactly the same height. If not, the work table will be tilted... not good.

The surface of the Bosch 4590 extrusions are anodized. The abrasive nature of the anodized surface will dull HSS cutters, but a good carbide mill cutter has no problems. Be sure to protect your ways from the swarf and anodize dust, which is a type of abrasive aluminum oxide! Think of grinding wheel dust. Very similar, quite fine, and bad for your mill.

I used a vacuum cleaner to evacuate the dust on the fly. Once the skin of the extrusion is removed, the abrasive swarf is no longer a hazard to your mill.

The face mill here is another Lovejoy tool...

...and it leaves behind an excellent finish!

Face mills are great for this type of process because they can plane a relatively large surface in one pass. Even a 3/4" end mill would require three passes, and unless the setup is perfect, the subsequent passes would produce a surface less than planar. Of course, the caveat is that your mill head must be trammed to perfection. Otherwise, the large diameter of the face mill will exaggerate any tram errors, and produce a rippled surface.

Probably 0.008" was removed on the upper side of the two supports to true the Y-axis mounting plane. Once finished, it is best to leave the two 4590 support members assembled to the base plate, keeping this Y-axis rail plane in truth. I did mark the plate and the two support members with stamps so it can be reassembled in the same orientation in which it was milled, if necessary.

Back on the bench, you can get a better idea of the base structure. The front and rear plates are temporarily attached. The right-hand Y-axis THK rail (HSR25) has been set on the mounting plane, while the left side is still on the work bench. Note the Y-axis ballscrew to the right of the base assembly.