PDA

View Full Version : Questions about lathe leveling & runout



berol
07-18-2006, 12:04 PM
I leveled up my Sakai lathe in both directions, as good as my patience would permit, using a 10" Starrett precision level. After leveling I inserted a 12" L x 1" dia test bar which has an MT2 taper on one end and a center angle on the other. I cleaned the taper and inserted the bar into the headstock and into the tailstock using a fixed center. Using a .0001" Interapid indicator at the top center of the bar I indicated the bar near the headstock to be .002" higher than it indicated near the tailstock over an 11" travel. On the side of the bar the indicator showed me the bar was .001" closer to me near the tailstock than near the headstock. Would additional leveling improve this runout or is this about as good as one can get. What kind of runout might a person expect to find on a well set up Hardinge or other good quality lathe? Thanks! Berol

SGW
07-18-2006, 12:33 PM
If I'm understanding the setup correctly, what you are mainly testing is the alignment of the tailstock to the headstock. This may or may not accurately reflect the overall lathe alignment. The tailstock my not be aligned correctly. It's quite easy for a slightly offset tailstock to pull a bar in, out, up, or down by several thousandths of an inch.

Back off the tailstock center and take the readings again.

J Tiers
07-18-2006, 12:33 PM
First, you are mixing two separate types of measurement....

The bar you have should be for use WITHOUT the tailstock in it.... i.e. "unrestrained".

That will allow you to determine that the carriage is moving parallel to the spindle axis, and finish shimming (assuming the bar is long enough and accurate).


The, separately, you put a test bar that you can make, between centers, which allows you to set the tailstock properly in line with the spindle axis. (or you can set the T/S so it goes nicely into the center on the end of the bar you have......).

By trying to jam the bar you have into both the spindle taper and the T/S center, you get bad readings. If there is any error, you may bend the bar, or just get an inaccurate reading.

This is because the flex in the bar vs the bed will fight it out and come to an average position, which is not anything you need to know about.

You want to know how the bed and spindle are alined AS THEY SET, and then whether the T/S is in line with spindle.

Properly set up, a good accurate machine should be well down in the tenths on an unrestrained test bar over a foot or so. Very good machines will be lower.

It should be possible to be in the same accuracy range on a part between centers, meaning in teh tenths over the whole length of the bed. Of course that can be adjusted with T/S.

The unrestrained bar measurement is a measurement of "inherent" accuracy as-set-up.

berol
07-18-2006, 06:50 PM
If I'm understanding the setup correctly, what you are mainly testing is the alignment of the tailstock to the headstock. This may or may not accurately reflect the overall lathe alignment. The tailstock my not be aligned correctly. It's quite easy for a slightly offset tailstock to pull a bar in, out, up, or down by several thousandths of an inch.

Back off the tailstock center and take the readings again.
Thanks for the help. Should the MT2 taper hold that 1" dia bar straight with a 12" overhang?. Guess I can't do much side to side as the Sakai tailstock is not adjustable. Regards, Berol

SGW
07-18-2006, 07:26 PM
My surmise: The side-to-side shouldn't be a problem, assuming the bar is straight. You may get a little sag, but vertical error is less critical to accuracy.

Personally, I'm not a big fan of test bars. I think it's more to the point to chuck a piece of bar, maybe 1" dia. x 6" long, in the chuck (only), and turn two collars on it at the same setting, one out at the end and one up near the chuck. Take a light skim cut over both, to minimize deflection caused by tool pressure, then measure to see if they're the same diameter. Adjust the bed leveling until they are.

Once you get that correct, you'll know the headstock is aligned with the bed. Then set up a bar between centers (NOT holding in a chuck) and repeat the exercise. Adjust the tailstock left/right until the lathe will cut a parallel bar. Then you know the tailstock is properly aligned. But since that's not adjustable on your lathe, the best you can do is know how far it's off. One thing you could do, if it's "unacceptable," is contrive an offset tailstock center.

wierdscience
07-18-2006, 07:32 PM
.001" vertical and .002" horizonal over 12" isn't much at all.Believe it or not the heat from your hands can cause that much influence on your test bar.

Maybe try letting the test bar sit in the machine for awhile and run the checks again and see what you get.

Also try rotating the bar 90* and checking it again,the error could partly be the test bar.

In any rate .001" is 1/3 of a human hair so it's not too bad out:D

Al MacDonald
07-18-2006, 09:43 PM
For checking my lathe for straight (substitute level if you're not on a ship) I have a precision level supposedly good to .0002"/ft. This is a good time to be fussy. Instead of a test bar I checked out my lathe by centering a 3" diameter chunk of mild steel by about 18" in my 4jaw chuck, did a very light test cut and miked both ends of the cut. My headstock will swivel so I kept moving it to compensate for the error and did another test cut. I eventually managed a .0001" taper over 11". For a smaller lathe you could substitute a smaller diameter testing piece but it needs to be big enough to remain rigid without support at the tailstock end. You could accurately check the tailstock alignment for side to side by using a test bar on centers, doing a light cut at each end and measuring the cut diameter. To see if the tailstock is the correct height one might be able to chuck a dial indicator and run it around a center or something similar set in the tailstock taper. Cheers, al.

Millman
07-18-2006, 10:12 PM
{I eventually managed a .0001" taper over 11".} Now you know we can't "tolerate" accuracy such as that. Blasphemous!