Evaluating wear on a flat-bed Atlas or Atlas-Craftsman lathe.

[wa5cab]

First, in a used flat bed lathe, the ways will show the tarnish and wear pattern typical of these lathes when they have had at least a little usage in the past few years. The narrow tarnished band is on the part of the ways not touched by either the carriage or the tail stock. The front and rear parts of the top of both ways should show no tarnish, indicative of a lathe that has had at least some recent use. If the ways are tarnished across their full width, the lathe probably hasn't been used recently. And if the entire surface is bright and shiny, someone recently polished it (or it is new). If you need to evaluate the bed wear, do so by measuring the way thickness of front and rear ways about every 3 inches from the right end of the bed up to the front of the headstock. The nominal new readings for Atlas built lathes would be 0.375" from end to end for the 9", 10" and early 12", 0.438" for the 6" and .500" for the late 12". Establish the actual new thickness from the readings at the extreme right end of the ways or under the headstock. Do the same measurements of the width of the front and rear ways. The width of the front way usually shows minimal wear because the reaction force is mostly taken only by the rear way.

In a used lathe, there will be wear in the cross feed nut and maybe the screw as well. To evaluate the wear condition of the nut and screw, first crank the cross feed nut completely off of the screw. Then properly adjust the screw end float by differential adjustment of the two nuts on the visible portion of the cross feed assembly. One nut holds the crank on. The other may be just a thin pattern hex nut, a threaded collar or the visible part of an internally threaded shoulder bushing. This end-float should be as near to 0.000" as you can get it without there being much drag when turning the crank. Then pull the cross slide back toward you while turning the crank CCW until you can just see the end of the screw with the cross slide chip guard removed. Take the slack out by turning the crank a little CCW and zero the dial. Turn the crank CW until you just feel the thread in the nut hitting the thread on the screw. The dial reading is the normal running clearance plus wear on the nut alone (since one never operates the lathe with the cross slide that far away from the operator). With a new nut, this should be 0.000" to 0.002". Crank the cross slide back to about where it spends the most of its working life and repeat. Subtract the nut clearance reading from the new reading. This figure is screw wear. Check it in two or three cross slide positions. What you do next depends upon you and how deep your financial pockets are. But note that if you use normal SOP, a worn nut or screw should have no affect on the accuracy of your work.

To evaluate the condition of the lead screw and half-nuts, with the motor off, first crank the carriage the full length of the bed and back again. There should be to sign of the half nuts tending to drag on the lead screw.

-more later-

 

[Provincial]

I completely rebuilt an Atlas QC42 (the 10F with a factory-installed quick change gearbox, introduced around 1947) almost forty years ago. It had been somewhat abused, and the old fellow had it bolted to a 2x12 plank in an unheated shop. Pretty much all the Zamak gears and the half nuts were completely worn out, except the quick change gears had survived quite well. The lead screw threads were worn considerably in the usual area.

The spindle and its bearings were in perfect condition, which made the project worthwhile. I wanted to keep the lathe in service out of respect for the original owner, so I put more time and money into it than most others would. At that point, you could still get most everything from Clausing. I bought the last lead screw they had in inventory!

To get to the point, I have some input from experience about the ways. Use a straight edge to check for flatness end-to-end. Mine had about .015" sway-back (low at the mid-point), probably from being clamped to an uneven surface that put a bending force on the unsupported area.

Sway-back doesn't affect straightness of cut as much as wear on the rear "outside" face of the ways. This is because it makes the contact point of the cutting tool raise and lower. Geometry means that as the tool drops down, it takes less of a cut. This amount is pretty small, and .015" up and down results in only a few thousands difference in the finished diameter of the work.

I took the stripped-down bed to a precision grinding shop in a larger city a couple of hours away. They took surprisingly little off the surfaces of the ways to get them square and uniform. The lathe has cut perfectly straight ever since.

To sum up: put a precision straight edge end-to-end on the ways to check for "sway-back."

 

[wa5cab]

Supposedly, you can compensate for bed twist, sway back and humpback but it depends upon how the bed legs are made and how stiff the surface that the legs attach to is compared to the bed. It would certainly help to start with a flat bed.

Incidentally I once wrote a paper that I think is in the sticky area that compared finished part diameter accuracy to errors in tailstock height errors. The conclusion was that the overall error from tailstock to headstock varied inversely as the square of the nominal part radius (as best I recall). But the reason that I mentioned it is that assuming that everything else was correct, the diameter error would also disappear for part diameters above about 5/8" with your bed as it was originally. By "disappear" I mean be good to 3 decimal points. The thing was actually prompted by the question raised 8 or 10 years ago as to how interchangeable tail stocks were.