View Full Version : Other Questions on Leveling (a Lathe)

Paul Alciatore
04-12-2008, 03:25 AM
After reading the other thread on levels, I was reminded of a couple of questions I had when I attempted to level my lathe.

First, almost all descriptions of how to level a lathe bed state that the level is placed across the ways. Sounds OK till you look at the ways. Most lathes have Vee ways and perhaps one or two flat ways. But the level will set on top of the Vees, not on the flats which are lower. Now, the descriptions of scraping I have read talk about scraping the working surfaces: the flats and the 45 degree Vee sides. But never the tops of the Vees ----- where the level will rest. So, what good is it to use a super accurate level on a surface that is not all that accurate? Or do all scrapers also do the tops of the Vees but never discuss it?

Second question. With a level that shows 0.0005" per foot, and a lathe with a base that is 6-8 inches deep (front to back) you would be seeing errors in the range of 0.0002" or 0.0003" between the front and back of the base. Where the heck do you get shims that are only a few tenths or that differ by only a few tenths. My shim kit started at 0.001". Aluminum foil is about 0.0007".

Or is this done by the torque on the bolts? Or what?

04-12-2008, 04:03 AM
Q1 - I use a pair of short parallels or match up a couple of 1/2" square toolbits with a mic, set them on the flats and have at it.

Q2 - I don't worry about .0002 errors in leveling - life is way too short. I make the fine adjustments by changing the bolt torque.

J Tiers
04-12-2008, 09:39 AM
There are always assumptions........

The tops of V's may or may not be finished evenly along the length. The way lesser machines are generally made is to gang-mill the surfaces and then grind or scrape to finish. In that case, it is reasonably likely that the tops are even the whole length.
It might have been planed and then ground, especially if larger, which may involve independent adjustments, but likewise an even height the whole length.
One might then use the tops as a reference, since they probably have zero wear aside from dings.

Or, one might make a pair of "saddles" that ride on the v-ways and provide a reference that is identical to that which the carriage will see. Use of the carriage (no topslide) is equivalent.
But of course, wear at the headstock may add an effective "twist" which might be as much as 10 or even 20 thou at the head end, but zero at the tailstock end, in rather bad cases. (front way normally wears more).
Un-twisting this "effective twist" will distort the bed. NOT un-twisting it will mean an error equivalent to a twisted bed.

One might use two parallels on a machine with V-flat type ways. Or possibly a "saddle" for the front and a matched parallel for the back. Use of the carriage itself is similar to that, so long as you remove the topslide and use the top of the crosslide ways (to eliminate extra variables)
These have the same problems of wear as a twin-v machine.

Bottom line is that if you have significant wear, no method will really work to 0.001.

If you do NOT have wear, the use of the carriage, or the use of purpose-made saddles is probably best.

04-12-2008, 04:53 PM
The tops of the V ways are not working surfaces, and I would never use them as a reference for the level, it serves no purpose. If the lathe has 90* V ways, use a set of V blocks, if the lathe has only one flat way, always keeping each block on the same way when moving from one end to the other. If the lathe has 2 V's and 2 flats, I would use matched parallels on the flat ways, observing the protocols as above.
With regard to your second question, I wouldn't worry about the height difference you are referring to, unless I'm involved in a reconditioning project, and then it becomes very important. The only time I would use shims, is to get the bubble in a useful position in the vial in the roughing stages. That practice is not recommended when one gets close to finishing; it then becomes a liability.

04-12-2008, 08:03 PM
I don't see the need to level the lathe. As long as the turning axis is parallel to the ways, what difference does it make if the ways are not parallel to the floor? If the machine is "square" relative to it's self, leveling shouldn't enter into it. Right?


04-12-2008, 08:13 PM
As long as the turning axis is parallel to the ways, what difference does it make if the ways are not parallel to the floor?

It doesn't make any difference, as long as the spindle (and tailstock) are aligned with the ways. The problem is, unless you have a laser autocollimator, a precision level is the easiest way to determine if things are aligned.

04-12-2008, 08:48 PM
" "Square" relative to itself" pretty much is the definition of level, and that apllies to very few lathes once the manufacturer is finished with the lathe. The Hardinge, Monarch 10EE, CVA and such are built upon a 3 point foundation. They can't be leveled in the traditional sense because you can't twist the beds into alignment. The only reason they are leveled is for lubrication purposes, and I do find it easier to use a lathe that is reasonably level. It's the rest of the lathes that leveling becomes very important.

If you are so fortunate to be able to plop a lathe on the floor, and it self aligns that's great. I was lucky one time in being able to do that, but when I relocated the lathe, my luck ran out. I had to bolt the lathe to the floor and twist it into level. The fabricated steel base on which the lathe was mounted got out alignment in the 20 mile move. It was a 13" Harrison, and practically brand new when I bought it, and still in excellent condition when I sold it, 14 years after the 20 mile move.

J Tiers
04-12-2008, 09:23 PM
The tops of the V ways are not working surfaces, and I would never use them as a reference for the level, it serves no purpose. If the lathe has 90* V ways, use a set of V blocks,

The "pro" argument is that the tops were likely finished in a way such as to keep them parallel to the ways, and they are NOT wear surfaces, so they are still as they were.

"Con".... as you are checking to less than 0.001, it is asking a lot for them to be that perfect, since they are not "functional" surfaces. They may be off several thou without affecting the machine operation.

As an added point, any wear on the machine will easily be more than that 0.001 you are measuring to, regardless of where you reference. But the tops might be at least good for a preliminary "survey" leveling check as the machine is set up...... and as a "sanity check".

Significant wear on the V and flat ways will also produce an error, whether you reference the tops, OR use blocks etc to reference from the ways..... You are unlikely to be able to twist the machine in a way that will correct a "dip" at the headstock, so you will be stuck with errors any way you go on a worn machine.

As far as the V-blocks, that is fine IF the ways are perfect 90 deg ways. But not all are, nor may they "still" be 90 deg after however much wear and possible prior re-grinding etc. So you may have to make your own. My Logan, for instance, does NOT fit a set of Brown and Sharpe v-blocks, the blocks don't "seat" on both slopes, and can shift, so obviously there is a slight error off 90 deg.

If you have an Atlas, you are set.... flat ways. No problems to worry about with the level, except the bed wear issues. But the wear on the wide flat surfaces of the Atlas will probably be less than on typical V-ways, unless the past owners were unacquainted with oil or cleaning. It is usually the wear on the tiny EDGES of the bed you need to worry about.

So NO way is perfectly suitable, although the V-block way is the generally *best* one IMO.

Most generally *usable* would be removing the crosslide and setting the level on the top of the crosslide way. That gives you a set of "v-blocks" that (presumably) fit right, and also means you do not have to disturb the level (with possible errors) when checking all along the bed.

04-12-2008, 09:56 PM
The tops of the V ways are not working surfaces, and I would never use them as a reference for the level, it serves no purpose.

In Machine Tool Reconditioning, Connelly describes using "bearing" and "non-bearing" surfaces as level references. Figure 26.28 shows using a precision level across the tops of a lathe bed with prismatic ways, but I think you'd want to check on a "lesser" machine (i.e., not a Hardinge or Monarch) if the tops of the prismatic ways were accurately ground. He explains that if you need to use "non-bearing" surfaces for the level, that you need to measure level at several points, average the reading, then "rough level" (my term) the lathe. Connelly says that the "final level" should only come from the level resting on precision, bearing surfaces, and that if you need to use the tops of the prismatic ways, that they should be spotted and scraped (after the rough leveling) such that level measurements at several points all give the same reading. But then again, Connelly was a perfectionist... :)

As a possible short-cut, Connelly instructs the reader to scrape the tailstock base first (since it mates with the prismatic ways), and using that as your reference artifact for all subsequent measurements. That seems like a lot less work than scraping the tops of the ways.