Lathe setup, correcting taper.

The Starrett #98 is definitely not sensitive for machine leveling, it is intended for millwright work; the 198 is 10 times the sensitivity. The 98 is good for getting the job withing the range of the 198. The ways should be level, if the lathe still cuts with taper, other means should be pursued, such as scraping for alignment under the headstock or shimming. It is best to not have a twist in the bed.
 
Before checking for turned taper, I check for perpendicularity of the cross slide travel to the spindle axis. If the travel isn't perpendicular, you will cut a dome or a dish depending upon which way the the misalignment runs. Since the carriage isn't moving on the ways, a twist in the bed will have minimal effect on this cut. On my lathe the only way to adjust this is to adjust the headstock.

I check for proper alignment by mounting a boss on my faceplate several inches from the center. I set up a dial or test indicator on the cross slide and rotate the faceplate so the boss is to the front and horizontal with the spindle axis and adjust the indicator to intersect the boss with the cross slide retracted. I zero my indicator and rotate the faceplate with the boss to the rear and horizontal and move the cross slide so the indicator again intersects the boss.. If the reading is zero, the cross slide is perpendicular with the spindle axis. If not, the difference divided by the separation between the two measurement points is the amount of taper. To ensure hitting the same point on the boss , I place a small marker dot on the boss.

Once this is done, I next adjust turned taper, checking by Rollie's Dad's method or by the two collar test. Finally, I adjust my tailstock using a modification of Rollie's Dad's method or by the two collar test.
 
Not to hijack the thread but the discussion about the accuracy of the Starrett 98 vs 199 got me searching around. While the graduations on the 98 are .005" the spacing between the graduations is quite large. I'm sure one can easily use that to a finer degree than .005".

Poking around for used 199's I found you can get a brand new Mitutoyo 960-613 for $375 which is cheaper than the other two less accurate models. It's accuracy is .00017". I can't imagine trying to level a lathe with that. I'd be chasing my tail for hours. LOL
 
Poking around for used 199's I found you can get a brand new Mitutoyo 960-613 for $375 which is cheaper than the other two less accurate models. It's accuracy is .00017". I can't imagine trying to level a lathe with that. I'd be chasing my tail for hours. LOL

The Mitutoyo 960-613 has a sensitivity of .00024"/12" as compared to the Starrett 199 with a sensitivity of .0005"/12". I'm not sure what the accuracy spec refers to; possibly the flatness of the surface?
 
I'm not sure what the accuracy spec refers to; possibly the flatness of the surface?

Me neither!

Am I wrong in interpreting that to mean it's twice as sensitive as the Starrett? And, it's under 8" vs 15" which might be a bit more conducive to leveling a lathe.
 
I
Me neither!

Am I wrong in interpreting that to mean it's twice as sensitive as the Starrett? And, it's under 8" vs 15" which might be a bit more conducive to leveling a lathe.
It is twice as sensitive as the 199. A problem with a shorter base is that it will be more prone to error due to imperfections or debris under the base. My level has a sensitivity of .0005"/10" and is slightly less sensitive than the 199 but it has an 8" base. For measuring across the ways, it is fine as the span is less than 8" on my lathe. If I were leveling a mill table, I would prefer a level with a longer base.
 
The sensitivity of a spirit level refers to the change in slope required for the bubble to move one unit on the vial. The more sensitive the level, the less slope required to move the bubble one unit. For example, my Kinex level is sensitive to 0.02mm/M, or 0.0002"/10". In contrast, a Starrett 199Z Master level is sensitive to 0.0005"/foot, so about half as sensitive as my Kinex. It can take up to a minute for the bubble to settle on one of these levels so be patient when using one, and try not to have other machines running nearby when you do.

In general, I follow a sequence when setting up or checking my lathe
  • I start by confirming that the headstock is aligned with the ways. You can do this with test bars or Rollie's Dad's Method or whatever you like. I prefer to prove the position with test cuts on a 4-6" long bar supported by the chuck alone. If the diameter all the way down the bar is the same, the headstock must be aligned with the ways.
  • Level the cabinet next. I move the saddle to mid-position and adjust the feet on the cabinet so the lathe is zeroed in both X and Z.
  • Level the lathe. I use my Starrett 98-6 sitting on the cross slide, perpendicular to the spindle centerline and set there with a square. I start at the headstock end and adjust the studs under the headstock end to get it level, then do the same at the tailstock end. Once completed, I lock the lathe down to the cabinet. Then I switch to my Kinex and repeat, this time adjusting the feet of the cabinet. Because the lathe is mostly leveled already, it takes just a minute adjustment of one of the feet to bring the lathe into level according to the level.
  • Then I do a 2-collar test to finalize the level. Most of the time, this just requires a tiny tweak on one of the feet and the lathe is level.
  • Finally, I use a test bar that was cut on this lathe to align the tailstock to the spindle and I'm done.
Regardless of how you do it, I suggest you begin by making sure the headstock is aligned with the ways. Everything else that follows depends on this one thing being right.
 
Reading these replys which reinforced my thoughts on twisting the bed is not fix and merely a bandaid, I'm back to leveling the machine today and adjusting the headstock to remove the taper. Thanks as always for the input.
 
Reading these replys which reinforced my thoughts on twisting the bed is not fix and merely a bandaid, I'm back to leveling the machine today and adjusting the headstock to remove the taper. Thanks as always for the input.
Twisting the bed needn't be a bandaid. The important takeaway is that two wrongs don't (necessarilly) make a right. All other options should be exhausted first and bed twist used as a final tweak. In using a level to adjust for taper you are assuming that the lathe had been made correctly to begin with and that wear during use had not created a condition that leveling couldn't cure.

In the end, the ultimate test is whether or not you can make parts to your specification. If you still have a taper when turning after leveling, it does you no good to have a level machine. OTOH, if your machining operations are cutting true, why should it matter if you had to twist the bed to make it happen.

It is wise to remember that any and every manufacturing operation has some error associated with it. Tolerances are set up to define the product and if the product is within the tolerance specs, it passes out the door. Good manufacturing has tighter tolerances, poor manufacturing has looser or non-existent tolerances.

Another important point is that we must analyze what surfaces are important to our alignment. For example, my lathe has a prism on the front way and a flat on the back way that define the location along the bed. I cannot assume that the flat just in from the prism is parallel to the prism. Nor can I assume that the top of the prism is parallel to the working surfaces of the prism. Nor can I assume that the prism itself was ground parallel. Setting a level on the wrong surfaces and leveling might not produce a straight cut even though the level indicated that the lathe was leveled.
 
Project complete. Re-leveled the lathe again. Probably in better level condition now than it was before. Adjusted the headstock to .0002 in 9". I'm happy with it and will sleep better knowing the bed is not twisted to a point of excessive stress. Going to use it as is for a while to let things settle in and then recheck.
 
Back
Top