"Leveling" a lathe.

[T Bredehoft]

While this deals with a PM product it is not about that product but a procedure involving that product.

I recently (last year) acquired a new PM10-30 lathe, installed it in my shop and began enjoying it. I'm not working for NASA, so I wasn't concerned about 'leveling' it. Today I got bored and wondered just how out of level it is. I have a Starrett level, but it only has two lines on the bubble, (not a .0001 reader).
I checked the lathe on both ends and found that the two ends are off by two thicknesses of one of the lines.

Just how far out is it? Well, the level is very nearly 8" long, and a one thousandth (.001) shim under one end moves the bubble one thickness of the line.

One thou, devised by 8 is .000125 per inch. so my lathe is out of parallel by .00025.per inch. Remember, this isn't even gov't work.

So, friends and colleagues, just how bad is my lathe out of parallel?

 

[mikey]

So, friends and colleagues, just how bad is my lathe out of parallel?

Dunno', what does your 2-collar test show? Bubble readings have little value. The 2-collar test will show you, in real numbers, how much taper you have to eliminate and that is the goal, right?

 

[benmychree]

I think likely your level is .005" per foot graduation; is it a #98? I think the 98 is or was available with graduations of >005" per foot or with just two lines. Generally speaking, the #98 is not sensitive enough for final lathe leveling, which requires a level with greater sensitivity, usually .0005" per foot per graduation. The #98 is sensitive enough for rough leveling and then going to the two collar method, which has been described on HM many times previously.
The bottom line is that a lathe does not need to be level, it just needs to be in alignment, with any twist removed from the bed, by whatever method is appropriate, such as jack screws, shims, or whatever.

 

[silence dogood]

On the low side of the level, take a feeler gauge and pull what ever thickness feeler that puts the bubble back between the lines. Read the thickness on that feeler. Now you know how thick your shim will be. I did that to my lathe. My brother-in-law owns both a Starrett and a Southbend levels. Used both and both came out the same accuracy. Have to say, the Starrett was a bit touchier. Had to wait a few seconds for the bubble to settle down. Before adding a shim, my lathe was 4 thousands off. According to both levels, it's right on.

 

[T Bredehoft]

I haven't had the material to make a "two collar" test. Retirement's great in some ways, others (cost of materials) it sucks. I did inquire of the PM tecs about parallelism of the ways to the headstock, was told not to be surprised if it were no better than .0008 per inch. That's about what mine is.

 

[wrmiller]

Dunno', what does your 2-collar test show? Bubble readings have little value. The 2-collar test will show you, in real numbers, how much taper you have to eliminate and that is the goal, right?

This...

I use a level to get the major twist out of my lathe bed. Once close, I then start cutting on a piece of 2 or 3" diameter stock about a foot long to really dial the lathe in. No way I'd just trust a level, but that's just me.

 

[silence dogood]

Tom, after reading these comments a second time, I realize that my comment was basically a repeat of what you stated in the beginning. Dope slap time. :***** slap: However, even both level readings and two collar test may not be enough. My brother-in-law bought a Southbend when the company was still making lathes here in U.S. He dutifully leveled both ends and shimmed the tail stock end when needed. Ran a test bar, both ends were right on. But the middle was off, Don't remember if it was less or more, but it was way off. Turned out there was a double twist in the opposite directions in the bed. Even a 2-collar test would not be enough, that's where wrmiller has what I consider the best answer. A happy ending, Southbend had Dave's (my brother-in-law) lathe sent back and repaired free of charge. Dang it, wish that company was still in business.

 

[wrmiller]

While out in the shop today, I looked at my 'test piece'. It's a 12" piece of 2" round that I skim cut as much of the length as I can and then measure 4 or 5 places along it's length. Not sure why I started doing that though, but maybe I wanted to see if the saddle was wandering on the lathe bed?