Homebuilt Metal Cutting Lathe: The Bed

My lathe will have hand feed at least, preferably by both rack and pinion for manual advance, and by lead screw which will eventually be driven by the spindle for thread cutting. For the bed, however, I've chosen to model mine partly after the Popular Mechanics design (hereafter referred to as the PM lathe), and partly after a Taig. The PM design is based on angle iron -- and since I had a scrap bed frame in my basement, it supplied the 1 1/2" angle iron to make the bed. However, I like the Taig idea of cantilevering the bed; this prevents flex in the table, bench, or board the lathe is attached to from translating into flex in the ways. I'll find out after the bed is assembled if it's stiff enough this way (I can easily convert to having both ends supported if that appears necessary), though the plan at present is to use a steel plate for the ways, similar to the Gingery design. My lathe, however, will be larger than the Taig or Gingery, and much larger than the PM lathe -- I'm designing a 24" bed, which is the same as the Gingery, but more swing and a wider bed; I plan to make the bed at least 4 1/2" wide, versus the 3" width of the Gingery. That will let me maintain better rigidity in the cross slide when machining the larger work, because the cross slide won't have to hang unsupported off the edge of the bed for large diameter work pieces.

September 1, 2001: Here we have the rails for two lathes, cut from the bed frame, but not fully dimensioned for the lathe bed. The longer pieces are actual angle iron, 1 1/2" wide and 3/32" thick, and will form the beds for both lathes; the shorter parts are steel folded to the same width as the angle iron, but 1/16" thick, and will be used for the bases. They'll be bolted together with aluminum spacer blocks to form the cantilevered bed assembly. These parts weigh ten to fifteen pounds per lathe.

September 25, 2001: There's been a slight change of plan. Mark remembered he had almost five feet of aluminum square tube, 3" across with a 1/4" wall. We had been concerned about the stiffness of the angle iron bed after reading reports of problems related to rigidity in Gingery lathes (which use a 1/4" plate ways on top of a cast aluminum channel bed). The aluminum square tube will be much stiffer than the angle with spacers at the ends, and will simplify construction as well. Best of all, we can borrow a page from Taig and, once the screws and nuts are all in place, fill the bed tube with concrete or something similar to further stiffen it and absorb vibration. The end result will be heavier than the angle iron bed once the concrete is in place, but not a great deal heavier (especially if we can add foam beads to the concrete or otherwise lighten it in some way).

Here's a preliminary assembly of the new bed design. Well, really, at this stage it's just setting the longer tube on top of the short tubes that will form the feet. The feet you see here will be augmented with small angle iron feet on the sides to give a base about six inches wide; those angle iron feet can, in turn, be drilled to allow bolting the lathe to a table or bench, which should greatly improve stability. The lower tube sections will be bolted to the upper sections, and the ways bolted to the top of what you see here. Current plan is to use a Cold Rolled Steel plate 4" x 24" x 1/2" thick, secured with flathead machine screws countersunk just below the surface. We're still trying to decide whether to thread the aluminum or put nuts inside the tube -- the latter promises to be a pain, but once the assembly is done, will be much stronger, and reattaching the ways to the bed after the concrete is in would be a major pain if the screws stripped out of the aluminum.

It's been a long time since we had any progress -- Mark was pretty busy with classes, and I was pretty busy just trying to keep my head above water. We finally, however, did manage to get together today and get some progress made. In the course of about two hours of actual working time, we laid out and drilled all the holes to fasten the feet to the bed as seen in the image above from September, and drilled the holes in the angle iron feet to fasten them on the sides of the short tube sections. We haven't got those holes transferred to the short tubes or drilled -- those will be tapped in the aluminum, as they're both accessible even after full assembly, and don't take much stress. We plan to use #8-32 socket head screws for that task, and 1/4"x20 bolts and nuts to fasten the short tubes to the main bed tube. The ways will then be fastened to the bed with 1/4" x 20 flathead screws countersunk in the ways plate, or (if we can figure out a clean way to do it) hex head bolts counterbored into the ways -- either way, they'll be held by nuts on the inside of the bed tube.

Today we got together again and completed the drilling of the holes to mount the feet on the legs, and fastened the bed parts together. This completes the structural part of the bed construction, though we aren't yet done with the bed -- although the bed is all there now, we still have to scrape the top of the bed tube flat enough that when we fasten on the steel plate ways, the ways will require less (or no) scraping. All the scraping we can do on the aluminum will mean that much less to be done on the tough cold rolled steel of the ways. So, the next step is to get some kind of surface reference, an indicator of some kind (red lead in linseed oil was traditional, but Prussian Blue oil paint has also been suggested, and is much easier to get), and start flattening the top of the bed. Once that's done, we'll be ready to clamp on the ways, then drill and countersink the holes for the bolts to fasten the ways to the bed. Once that's done, the ways will be complete, and I'll start the next chapter of this account.

Meantime, here are some pictures of the most recent stage -- this is essentially how the bed will look until the lathe is finished and the tube filled with dampening material.

We started the process of flattening the bed tops in preparation to fitting the ways plates. The first step is to file off the major highs and lows; for this, I bought a very large file labeled as suitable for aluminum. It has a very coarse cut, with a finer and much shallower second cut across the first, almost as if a rough file had been cut on a blank already partially cut for a finish file. With an appropriate angle of cut and effort expended, it cuts pretty smoothly and takes off the metal rather quickly.

Here is the first bed, with filing just started. You can see how the file is cutting mostly at the edges, the corners of the tube. There are two reasons this might happen. One is if someone is leaning too hard on the ends of the file and flexing it; the other is if the process that formed the bed tube had left a slight inward bow in the material. The latter is why we go through the process of filing and scraping the bed before attaching the ways: because the manufacturing of aluminum square tube takes no account of its use for precision machines like a lathe; it's enough if the tube is straight and sturdy enough for structural use. And, given that the same phenomenon occurred with both bed tubes and with two different people using the file with significantly different techniques, there's reason to believe that fabrication is the cause in this case.

Here we see the bed tubes, one with initial filing completed, the other with the surface as found -- extruded and then weathered for an unknown length of time under unknown conditions. Notice the smooth reflection on the right side, aside from scratches and corrosion pitting. The scratches come from pinning, where chips from filing lodge in the file's teeth and gouge the smooth surface; this phenomenon is pretty much restricted to soft materials such as brass and aluminum, and is worst with "gummy" alloys such as the 6061 the tubes were made from. "Gummy" metals are those that are both ductile and malleable, and which can work harden. The combination of these properties can lead to chips that lodge in file teeth, and are harder than the surface they were cut from.

Fortunately, even very deep scratches make no difference in this process, as the heavy steel ways plate will even out any small imperfections and make contact with the bed only at the high spots. The next part of the process, scraping, is intended to ensure that those high spots are evenly distributed and comprise a large enough percentage of the total surface that the parts don't deform when clamped together by bolts -- or, in the case of sliding surfaces, to ensure that the two surfaces can slide over one another without changes in height or orientation.

No new pictures, but we started the process of scraping. We're using a 24" carpenter's level with a milled face as the reference flat -- this is suggested by Gingery in his book on lathe making, and we've had confirmation from at least one person who used Gingery's methods to complete a lathe that it can produce a more than adequate surface. The procdure is to apply "bluing" (which appears to be Prussian Blue or phthalocyanine blue, aka Thalo Blue, oil paint, at least to this one-time painter's eye) to one surface and then put it in firm contact with the other surface, giving a light tap or slight rub to transfer the dye. In our case, we applied the blue to the level, marked the work, and then scraped away the blue marks and an area about a half inch to an inch around the marks.

It didn't take long, starting from some good Web accounts of scraping with hand and power tools, to get a feel for the correct tool angle, stroke length and speed, and pressure needed to make clean, smooth cuts, and after five to ten repetitions of the marking and scraping cycle, we could easily see how the contact points were spreading. With a couple hours of work we were both able to advance from limited contact to fairly broad areas of contact; probably one to three more sessions will see these beds ready to attach the ways. It was also pleasing to see the scratches from filing slowly disappear as the metal around them was scraped away, and see the surface start to pick up the characteristic appearance of a scraped surface as scraping in different directions after each marking led to areas where marks overlapped and intersected.

We did find that our scrapers will likely need to be stoned once or twice per work session, but I have a fine grained sharpening stone that I originally bought (some twenty years ago) to hone my straight razor that does a very nice job of smoothing and cleaning up the edges without taking off material so rapidly as to make it easy to (incorrectly) reshape the cutting edge of the tool. If necessary, I also have a much coarser 2 in 1 stone that I bought for beveling the edge of my telescope mirror, and it is coarse enough to reshape the edge with a bit of work -- but I don't expect to need that.

February 22, 2003

No, it hasn't been six months since we did anything on the lathe; in fact, we've been getting together every couple weeks for most of the intervening time and working, but I've been too busy with other things to update this page. In the interim, however, we completed the beds; all the parts are drilled, all the scraping is completed, and Mark went ahead and painted his (I prefer to wait on paint until all the work that might scratch up a paint job is completed -- most likely I'll either leave the lathe in natural metal or disassemble it for painting when it's complete). Here are some pictures of the process.

This is the surface plate we improvised to make it easier to mark the bed -- the bed surface is several times as wide as the level we had been using, and that made it somewhat difficult to get a good mark across the full width of the bed. This plate is a piece of float glass, orginally a product shelf from a store, frosted with 80 grit loose carborundum using the bottom of a jar, ground wet to give an evenly textured finish (if I were doing it again, I'd probably use finer abrasive or even leave the glass with its original polished surface; the frosted plate was hard to clean and couldn't be used for very critical fine scraping because it marks too coarsely). Under it is a swatch of deep pile carpet to give it full support without stressing the glass. A frame is built up around the glass with strips of lauan plywood that I got from the dumpster at work (a sheet of it had been under the payload on a pallet), stapled to a piece of 1/2" waferboard which is in turn epoxied to a scrap of fiberglass faced aluminum hex cell board I got at Boeing Surplus. The construction gives a stiff, reasonably flat base, and the carpet prevents irregularities in the base from transferring to the glass. The top layer of the frame laps over the glass, with about 1/16 inch clearance (that is, the frame captures the glass, but doesn't touch it when the plate is sitting flat as shown in the image), and there's about 1/8" or more of clearance around the edges of the glass; when we move the plate, the glass shifts quite noticeably in the frame, and that's our assurance that the original flatness of the glass is preserved.

Unfortunately, we found as we were scraping that the glass in fact isn't quite flat; there's a significant low area (probably .002" to .005" lower than the rest of the plate) inboard of one end; we worked around it by marking with the bed in both orientations on the glass, to ensure we weren't missing high spots on the bed. This would probably be less of a problem with glass that hasn't been in use for years as a shelf; the glass can acquire a set in that kind of use, but new glass wouldn't have had time to take a permanent curve. I've since learned that it would have been even better to use an offcut of granite counter top (a sink cutout, for instance); those are made the same way, on the same machines, as commercial granite surface plates, but aren't thick enough to support themselves. The same kind of support we made for our glass plate would let such a granite plate provide the kind of accuracy one would expect from the very heavy plates, starting at inches thick, commonly used for surface marking and comparison.

Here we see the bed in place on the surface plate for marking. The bed surface is first coated with bluing, then the bed is laid on the glass and rubbed back and forth a bit -- a fraction of an inch movement is plenty -- to ensure that the glass rubs off the blue at the points of contact..

Here we see one of the beds, marked and ready for a scraping session. This was quite hard to photograph; I have a half dozen shots that otherwise look much better than this one, but don't show the marks in the bluing at all. Here, however, you can see where contact with the suface plate has rubbed off the bluing in a large area near the center of the bed, and in smaller regions along the lower left side and near the right end. Those are the areas that will be scraped, and then the process repeated until about 50% to 75% of the surface makes contact, evenly distributed across the entire bed. That amount is plenty to assure a flat base for the 1/2" ways plate, though one might want 75% as a minimum if this were to be a primary reference surface in the lathe.

This is the end of the bed section -- when scraping was completed, the bed was finished and it was time to move on to operations on the ways.