The Z-axis;
planning stage
5 Bears Home Homebrew CNC bench mill

Column layout; top viewEnough playing with the X-axis! It's time to actually do some work and get this thing a bit close to the fateful day when aluminum chips begin to fly. The CNC motion is hypnotic. It's very easy to sit at the bench for far too long, playing with servo PID parameters and G-code, watching a stupid cast-iron table move back and forth... back and forth... back and forth... (sleepy yet?)

I have already done a bit of work on the Z-axis - two fairly thick and heavy side plates have already been milled true and are ready to mount onto the 90mm x 90mm column. These side plates contribute two critical functions; first, they add necessary stiffness and beef to the vertical column. More importantly, they form a mounting surface for the 25mm NSK rails. I thought at first that I would mount the rails directly to the column, but this is impossible, as the column is not wide enough. The spacing between two rails mounted in this fashion is too tight and a ballnut will not fit between them. Additionally, a greater rail to rail dimension adds stability to the assembly. Just like widening your stance for balance, a wider set of rails is less prone to flexure and displacement.

The first step is to lay out the major components and visualize exactly how this beast will be put together, with an eye on accuracy, maintainability, and ease of assembly. I need to consider the dimensions of the spindles that I will use, the rail/block assembly stroke, the ballscrew stroke, and the distance from the spindle nose to table. All of this needs to come together without any wasted axis travel or other dimensional weirdness.


The side plates have been temporarily secured to the column section, and laid on my big mill table. You can faintly see the 8 smaller holes on the near side plate, which are for 1/4" SHCS to T-nuts. The mating surface of the side plates was trued via face milling.

The Z-axis way consists of two rails and two blocks, type NSK LH25. This is an especially heavy-duty block and rail, and there is no need for more than two blocks... these two alone could probably support a couple of tons, although the aluminum cannot, of course. These are the long NSK blocks. There is a shorter variety of this block which won't handle the loads of this long variety.

My last eBay THK ballscrew is nestled between the rails. Note the commercial bearing block for the ballscrew. That was my favorite ebay buy, the "$5 NSK bearing thing" that costs $250 or so new. The clearance generated by the side plates (rails/blocks vs ballnut) is evident. Nothing is yet attached - it's all just sitting there.

The servo motor, ballscrew, fixed-end bearing, and simple (bottom) end bearing will be mounted on a 3/8" thick aluminum plate which will nestle between the rails, and be attached firmly to the column with a series of T-nuts and bolts. This makes the ballscrew assembly adjustable vertically, and is easier to construct, as the screw assembly can be constructed away from and off of the column, and be attached in moments. Also, the entire motive train of components can be removed from the column as one unit for maintenance or alignment.

At the near end, a brass bar represents the top of the work table, the measurement having been transferred for visualization. The rails are not long enough to occupy the entire column... ideally they would have been, but that's how long they were from the eBay purchase, so that's what I have to work with. The entire system is limited by the ballscrew stroke anyhow, and these rails and blocks have a stroke equal or greater than the ballscrew.

Maximum stroke possible with the THK screw = 11.7" / 300mm. A good system demands as much Z as you can, because Z is rapidly used up by tooling. A neat little 3" Anglock vise (eBay once more) will be the primary workholding device.

My goal here is to find the optimum rail position, relative to the ballscrew and spindle, so as to deliver the spindle nose to the table surface at the bottom of the travel.

One aspect of this task which will make my life easier is the fact that the rail mounting holes are regularly spaced... I am already planning on drilling and tapping extra holes for the rail mounting into the aluminum sideplates, so that I may adjust the vertical position of the rails by simply unbolting them and relocating them up or down. This will also allow me to install longer rails at a later date if needed.
After checking the basic layout and clearances, I then checked the 3 spindles that I will possibly use; L to R, an R8 "mini-mill" spindle, a KAVO 4041 HF spindle, and a Sherline ER-16 industrial spindle.

The distance from the back (mounting plane) of the spindle to the nose is important with my mill because of the limited Y-travel, thus the R8 spindle will probably be unuseable. That leaves the KAVO and the Sherline. Of course, I plan on making the back plate to be universal, and so allow other spindles to be mounted in the future.

A piece of clear plexiglass was set on the NSK blocks, and my primary spindle, the KAVO 4041, has been placed on it for rough meaurements. The KAVO spindle will be gripped with a clamping block around the shiny stainless steel housing. Note that the nose of the spindle, when set so that it is in approximate contact with the "table", is too far below the bottom of the NSK rails, indicating that the rails will have to be mounted approx 3" below their current position. We want to support the cutting end of the spindle as much as possible.

Fortunately, the Sherline spindle has rough dimensions very similar to the KAVO, so that a common backplate for the two will be easy to do.

Many might consider the Sherline spindle a bit small for this mill, but it makes up for size and power by being both attractively priced, and very flexible in use because of the ER16 collet nose. I have tested it under power, and it is a gutsy little spindle.

I haven't decided if I will begin with the ballscrew plate, or the Z rail mounting. One or the other will be next.