Z-axis ballscrew backplate

5 Bears Home Homebrew CNC bench mill

With the NSK 16mm ballscrew ready to go, with the correct bearings at both the fixed end and the simple end, I needed to mount this assembly with accuracy to the column. As I mentioned in the Z-axis planning stage, I wanted to be able to dismount the entire ballscrew from the column still aligned and attached to a mounting plate. Reference the picture to the left. This is very similar to the top view of the mill column (looking down from the top) as first shown in the planning stage. I have added the shaded region, which is the subject of this installment, and will be referred to as the Z-axis ballscrew backplate.

The distance from the column to the spindle mount plate is limited. The top edge of the ballnut must remain at or below the plane defined by the top surfaces of the linear blocks, which is where the spindle will reside. To do this, and still retain some rigidity in this plate, I started with 1/2" aluminum stock, with a channel milled into the surface to clear the ballnut during its travels up and down the column.

The ballscrew backplate, then, will have six holes for t-nut attachment to the column. The length will be slightly grater than the ballscrew. The fixed end bearing housing, and the simple end bearing, will in turn be bolted securely to this plate. Standard rules apply - the plate must be milled flat for accurate attachment to the column face, which will place the long axis of the ballscrew parallel to the column. This will prevent binding, and generally make for an accurate setup.One final bonus is that the back-plate will supplement the two thicker side plates in helping the column extrusion maintain rigidity.


To start, the plate is cut to length. The width of the stock is 3.5" or 88.9mm, and this allows for a nearly perfect fit against the 90mm wide face of the column, as shown. Being slightly loose will allow the backplate to be adjusted slightly for correct alignment. There is a 0.5mm gap between the side plates and the ballscrew backplate.
Only a rough drawing of the plate was created, and shown here, not to any scale. At either end of the plate, two "platforms" were milled to drop the mounting surfaces of the ballscrew hardware to a lower plane. This allows the ballnut assembly to remain below the level of the linear bearing blocks. Doing so forces the ballnut lower rim into collision with the plate, so an axial channel was milled to clear the nut. Finally, the bottom surface of the plate, which mates with the extruded column, was milled as shown, with three separate "pads". Only the pads contact the column. Doing this makes it easier to mate the backplate to the column with accuracy. The T-nut bolts, three sets of two, go through the backplate and mate with the column at the pad locations.
Creating a largish plate with truly parallel and accurate surfaces can be tricky. The first step is to mill away the two clearance areas on the back side of the plate, labeled as "A" in the drawing. This was done in the mill vise, as these cuts are not critical and need not be anything more than stock removal. The cut was made to a depth of 0.020" in two locations.
I enlisted the help of Karl, my son, in drilling the holes for the T-nut bolts. He is really good for a 12-yr old, and did both the drilling and the counterboring for 6 ea. 1/4 x 20 SHCS.
  The last two holes for the hold-down bolts are counterbored. Note the pattern of the bolt-holes matches the locations of the mating pads on the backside of the plate.
Remember, the mating pad surfaces are still as extruded... We have only milled the clearance cuts at position A on the drawing, to a depth of 0.020". Now, all three mating sections are milled so as to be in the exact same plane, and remove any inaccracies in the surfaces. I honestly thought that extruded aluminum flats were plenty accurate for this type of construction, but they aren't, and if you need something flat, you either face mill it, or buy the mic6 jig/fixture plate instead. If you produce your own aluminum monstrosity like this, trust me when I say the cost of the mic6 plate will be recouped in time saved in the shop.
  With the backplate to column mating surface complete, we can mount the plate for the critical milling operations in the most accurate manner... flat against the mill table. Four clamps were used to secure the plate, pads down against the table, to allow the milling of the upper surface, which is where the ballscrew will reside.

The first operation is to mill away the shelves labeled as "B". Onto these shelves will be mounted the ballscrew's angular contact bearing block (fixed end), and the simple end ball bearing block. A larger face mill is used here to create an accurate, planar surface.

It is critical that the shelves be of identical thickness. If one or the other is high, the ballnut will travel a sloping path, and the whole assembly will bind as the spindle moves up and down.

The first shelf has been milled. Note the nice surface finish which the face mill delivers. It was cut to 6.5mm thickness, but the exact thickness is not critical in this example so long as the opposite shelf be identical.

My depth guage shows a depth of exactly 0.255", or 6.477mm.

The opposite shelf was likewise milled to this thickness, as best as I could manage.. my quill has an LCD DRO of the capacitative variety, and probably not accurate much past .0005" at best.

After the plate was removed from the mill, my finest Mitutoyo digital mike showed both sides at exactly 0.2549". Deviation of 0. This is blind dumb luck but I'll certainly take it! I would have been happy with any measurement deviation of less than 0.001".

As expected, with the shelves cut, the ballscrew assembly no longer rests flat on the plate; rather, the ballnut conflicts at its bottom edge with the longitudinal upper surface of the plate. The ballnut was placed on the plate, and rough measurements were made to determine the depth and dimensions of the long cut, "C", which must be made to create clearance. Like the pad relief areas on the bottom, this is not a critical cut dimensionally. All we must do is remove enough aluminum so as to allow the ballnut unimpeded travel along its intended path.
The cut was made with a 3/4" rougher; it measured 0.100" deep, and 1.25" wide.
Finally, although not required, the rims of the cut were rounded concave with a 3/4" ball end mill, as shown. This is purely cosmetic, but it allows the cut to match the contours of the ballnut, and is pleasing to the eye.

The backplate is complete except for the ballscrew-securing bolts.

Note the longitudinal cut, which now clears the ballnut nicely.