X-axis Part I
5 Bears Home Homebrew CNC bench mill

As seen from the Y-axis, some very basic components go into each way - a ballscrew/ballnut, appropriate bearing support for both driven end and support end, and a set of recirculating linear blocks and rails for the motion. Of course, the whole point of a mill is to secure your work for cutting, and this requires a t-slotted or drilled and tapped table of appropriate flatness and rigidity. The original ebay mill (which I rejected, but did pirate the general design) used an extruded aluminum T-slotted table which couldn't have been any better than +/-.010" flat, and of questionable rigidity. A true cast-iron T-slotted table would be the way to go. I considered simply swiping one from another small bench mill, but rejected this as costly and wasteful. Another source for a small, cast iron table would be any one of a number of X-Y compound tables targeted at drilling and light milling. These have nothing more than a pair of X and Y ACME screws and handles to go with the table and base, and are used in a drill press or larger mill.

Palmgren makes a number of these, and the quality is certainly there, but they are pricey. A number of importers also offer Chinese-made versions. One importer on eBay is Pitstool Wholesale. These guys offer a large assortment of imported tooling at rock-bottom prices and some of it is quite nice. I took a chance ($79) on one of their tables, and scored big. You can too because it is a new, off-the-shelf item, if you choose to do so.

This picture is from their web site, and shows their smaller 5" X 12" table. Unless you are going this small, I'd recommend their 6" X 18" table. First I must say that everything below the table is quite frankly junk. The leadcrews, handles, and base are crude and suitable only for coarse drill-press use. But the table is excellent! Go figure. Nice, reasonably close-grained cast iron, it machines beautifully, no hard spots, and best of all the table is accurate, as far as I can measure with my tiny granite surface plate. A quality straight-edge layed across the surface at various angles shows a good black-out fit.

The table is secured to its original base via normal dovetails machined into both saddle and table. On my example, the shipper had dropped the box, and the casting which mounted the handle and leadscrew to the table was broken. It was a good thing that I wanted only the table. Everything below the table was stripped. I didn't even save the leadscrews, they were that bad. Into the trash they went.

The table itself though was so nice that I did not hesitate to buy another for CNC mill Mark 2. I also stripped that one and stored it well packed with anticorrosives for the future mill. Oh yes, the T-slots are standard Bridgeport size. And the T-slots were milled, not cast into the table, meaning the bottoms of the T-slots are smooth.

The basic layout of the end of the X-axis table. The top portion of the picture shows the iron table (upside down), and the ballscrew fixed-support bearing block, which is an NSK product I bought on eBay. It took a bit of fiddling and fussing to come up with a component layout which is compact, simple, and importantly, adjustable.

The bottom picture is the actual aluminum table endplate, generated by simply drawing the plate around the top portion of the picture, and transfering the necessary drillings and borings from the table and the bearing housing.

The endplate is secured to the table with 4 heavy SHCS of 10mm dia. The bearing housing is likewise connected with SHCS, in this case 4 ea. 1/4" X 20, located through the body of the bearing housing. The other 4 holes in the endplate are tapped 10-24 for the motor mount. With each axis, I used the same hole pattern and spacing, so that my shop-designed NEMA23 motor mounts can be interchanged.

Earlier, I had modified the bottom of the iron table to accept 2 THK SHS15 rails by milling two accurate flats onto the dovetail, and drilling/tapping for the rail-securing SHCS. There is just enough room to have a 100mm rail to rail spacing with 15mm rails. You will see the rail-mounting cuts, as well as ballscrew clearance cuts, in the next pictures.

Only a bit of work remained on the table itself. The two holes here, already drilled and tappped 10mm, are "native" to the table and secured the crude handwheel and ACME screw. I made use of them in my design, but had to add two more 10mm holes outboard for rigidity.

Before the drilling, shown here, I had carefully milled the end of the table to a true surface perpendicular to the table top. It must be 90 degrees, or when everything is tightened, a bending moment will be generated in the ballscrew; not good.

A tall drill press is needed to fit 18" of table, + drill bit, beneath the spindle. In a desperate pinch, these holes could probably be drilled by hand, but I wouldn't recommend doing this. The table is clamped to a big angle iron, and both outboard holes were drilled and tapped 10mm. The iron table drills and taps with ease.

I recommend drilling the holes a bit smaller than normal, unless your setup for the drilling is very rigid. Any time there is a bit of flex, the holes tend to go oversized. Be sure to get all the iron swarf from the table nooks and crevices, as it is abrasive and bad for precision components like ballscrews and linear blocks.
With all machining on the table complete, we can lay out the critical components, and begin to assemble them into a super-accurate X-axis assembly.

To the left is the ballscrew (5mm pitch) with NSK bearing block, ballnut and machined nut carrier, and at the far end, the simple-support end. This consists of a 28mmOD, 8mm ID radial bearing, and a machined aluminum bearing support flange.

On the table itself, you can see the machined surfaces for the THK rails (the machined rail areas are on top of the mail dovetails). Along with those, there are holes drilled/tapped for future swarf guards, and in the middle of the table, cuts were made to mount the ballscrew, as well as clear the ball nut in its travels.

It is absolutely critical that the surfaces machined into the bottom of the table be parallel to each other, and to the working surface of the table. The best way to do this is do all the work in one setup, with the table clamped inverted directly to your B'port table.

Cut both of the rail mounting flats in one setup without changing the locked quill depth. Likewise, mill any ballscrew support sections in the same setup, but at the appropriate depth. If the rails' longitudinal path diverges from the ballscrew, binding and inaccuracy will result. Take care!

An alternate view of the ballscrew driven end. It is important to understand that the ballscrew simple support end is connected directly to the table, while the driven end (sometimes called the fixed end) is connected to the end plate, which in turn is connected to the table. This means that the simple end is firmly fixed, while the drive end of the ballscrew may be adjusted up and down, relative to the table, via the inherent slop of the securing bolts. This is necessary so that the ballscrew may ultimately be installed perfectly parallel to to working surface of the table. If this is unclear, hopefully the following installation pics will help.
The first components to be mounted are the SHS15 rails and blocks. Just like the Y-axis alignment, great care is taken to ensure an accurate and parallel set of rails is installed. When tightening the cap screws for the rails, do it gradually, alternating screws, and ultimately torquing them as evenly as possible.

These blocks come tapped 4 ea 4mm. Note the convex grease nipple of brass, and the fact that the master rail is snugged firmly against the milled rim in the iron table.

Do you guys have benches as cluttered as this? ;-)

After the rails are secured, and the blocks installed, the simple support end of the ballscrew is secured. This is a machined, split bearing block which allows the radial bearing to be adjusted longitudinally, then securely clamped. The block mates to the table via 2 ea. 1/4" button head CS.

With the bearing set for a sliding fit, the block is secured to the table.

Moving towards the drive end of the ballscrew, we come to the ball nut. It is secured to its carrier with 4 ea. 8-32 SHCS, using either lock washers or wavy washers. The nut is not tightened yet to the carrier. Everything is loose enough to allow for adjustment, as there is a specific order in which everything will be tightened. All we want to do at this point is loosely assemble everything.

Note the necessary clearance cuts milled along the table to clear the ball nut and its carrier.

For photographic/web purposes, the left hand rail is not yet installed. This view shows the driven end. In Part II, the fun begins as we start to tighten everything in the correct sequence for best accuracy.