View Full Version : lathe leveling

12-20-2002, 07:03 PM
I just got a craftsman 12x36 lathe and I was wondering about the importance of having it leveled. The stand that it is on is a large I-beam with legs that is 75-100lbs. Is this going to be stable enough or should it be bolted to the floor? I know that I should at least shim it to be certain that the bed is not twisted. Is a carpenters level accurate enough for this or do I need to find a precision one?
Thanks for the help.

12-20-2002, 07:27 PM
I try to bolt every thing down in my shop that has moving parts or excess weight. I had a piece of equipment decide to walk will I was doing some buffing, I put a 100 pound lead weight on it, and it still walked. I mounted it to bench which was mounted to the wall, quess what it doesn't walk any more. My wood lathe is mounted to a work bench, while my 7 x 10 metahl is not, the metal lathe is heavy enough not walk or move.
I try to keep things level, I have a small (18 inches)carpenter level, as well as multi bubble 6 foot one. I try to level and true every thing up. I find that if the machine is level and I am trying to do a measurement while working on the piece, it may not be the same size relative to the machine or the work surface.
In my recent purchases I got a nice HF laser level, and had a friend machine a block with a one millimeter hole thru it. When the lase is level, and it shines thru the little hole, the piece I am working on is level.


12-20-2002, 08:01 PM
I have 2 lathes like this. One Craftsman bench and one Atlas floor model. I've never had either one bolted to the floor. The Atlas never gives a problem with chatter. The Craftsman is mounted on a wood bench and has always given problems with chatter. The wood just isn't sturdy enough, even though its heavy. I would make sure the stand sits on the floor and doesn't wobble. I would shim the stand level, then shim the lathe. This will take any extra twist out. When I leveled my lathes I used a good carpenter's level, then checked it with my Starrett and it was on the money. A good level is a good level. Hope this helps!

12-21-2002, 01:05 AM

Many find it difficult to level a lathe properly even with a precison level. Check some of the past posts for a link to "rollie's dad's method" which seems to work for most people.

I use a Starrett #199 Master Precision Level and have never had problems leveling machines (lots of practice). Important thing to remember is level the headstock laterally first then level the bed.

12-21-2002, 08:38 AM
I wouldn't worry about bolting this machine to floor.

Use the best level you have or can borrow easy. Make sure the bed doesn't have twist in it, even the careful use of a carpenters level is better than nothing. Starrett mechanics levels are just about right for this sort of thing, the higher precision levels are a pain. Loaned my .0006 per 10" Mitutoyo to a buddy to level up his Logan, readings varied per time of day, probably sunlight through a window, gravitaional pull or some such, just awful sensitive.

But do try to level as best you can.

12-21-2002, 09:00 AM
Halfnut's right -- the key is to be sure there is absolutely no twist in the bed, but don't bother with a precision level. You don't need one. See http://home.attbi.com/~wasser/NEMES/RDMLatheAlignment.html which is what Thrud is referring to.

As long as you don't try to turn seriously unbalanced work (and you shouldn't, anyway), you shouldn't need to bolt the lathe to the floor.

12-21-2002, 02:53 PM
Thanks for the tips everyone. I will try to level it this afternoon with my carpenters level. Thrud, I did check out "Rollie's Dad's method", I will see if I have any bar stock long enough and try this method also. Hopefully by next week.
Thanks again

12-21-2002, 11:58 PM
The only thing I can see with NOT using any decent level is that all the measuring in RDM is "internal". That is, nothing is an external reference, so how do you know you don't have offsetting errors within the machine?

You might chase around trying to get it right, and later find the headstock was out because a chip is under it, etc.

Seems you would want a level to "keep you honest" (a reality check) and then go on by using a test bar or RDM, etc.

This was just discussed on an email group. Everything from "don't bother leveling" to "you hafta use a 'master precision' level" was posted as a reply. What the guy who posted the original question thought. I don't know.

[This message has been edited by Oso (edited 12-22-2002).]

12-22-2002, 01:33 AM
Remember to blt your lathe to the floor if
you live in earthquake country. I remember
seeing photos of a hsm's shop near the
epicenter of the 1989 quake in the SF Bay
area (the one that interrupted the World
Series). All his tools were lying in a heap
in his shop ....

- Bart

Bart Smaalders

Forrest Addy
12-22-2002, 02:39 AM
To level or not to level a lathe? Whatta question! Heck yes level it.

A new lathe should be leveled. The bed ways represent the path of the saddle. When it was set up at the factory the two flat ways (or some other bed way reference) was leveled crossways and longitudinally. as eeach piece was added the effect of its weight was compensated for by tweaking the leveling as necessary.

All other references were taken from the saddle axis including the spindle. The underside of the headstock was scraped to align the spindle axis to the ways as represented by the axis of a proof bar in the spindle taper. Same with the tailstock.

Theoretically, a new lathe properly leveled with turn and bore straight right from the start.

An older lathe will have some wear. Chances are it will turn some taper after even the most careful leveling. Never-the-less you still level the machine first. That's the starting point.

After the machine is leveled you may still have to tweak it. How much to tweak depends on the leveling arrangements in the headstock pedistal and whether it's possible to "kick" (force a slight bend in the horizontal direction) the machine.

Kickers are short lengths of barstock with a hole tapped 90 to the bar's axis. They're installed at the corners of the pedistals. Jacking screws are installed in the hole to bear against the pedistal. The kickers are grouted into the machine's foundation with the jacking screw aligned acoss the bed axis and a height to engage a solid part of the pedistal. In order for kickers to beeffective the machine HAS to be bolted to the floor.

Clever use of the levels, the kickers, and a sense of visualization in three dimensions will lead to straight turning on even a machine with considerable wear.

Then there's the tailstock. If the tailstock is pointing at the moon instead of nicely along the spindle axis you have to expect some problems turning striaght and drilling deep holes. Sometimes a little scraping to make the quill axis parellel to the spindle axis and shimming it to the spindle height is all you need to do.

If the quill bore is bell mouthed and you want to make straight parts between centers you're running a foot race with one foot in a bucket. Every time you set the quill clamp the quill is forced sideways in rough proportion to the torque you applied to the lever. It'll drive you nuts until you finally figure out what's wrong.

Here's a quick and dirty test. Clamp your tailstock, put an indicator against the quill and set a zero, and manipulate the clampfrom loose to tight. 0.0003" movement is almost new machine performance. An older lathe, veteren of hard service might have 0.005 quill clamp offset. That's 0.010 mystery taper.

Leveling the machine at installation is a simple way to ensure the alignments at the purchaser's shop are the same as established at the factory.

The machine can be aligned out of level and still run within specs but leveling is far quicker. How do I know quicker? I set up a small horizontal boring mill on board a ship and the ship was waterborne. Levels don't work on board a floating ship. Whatta nightmare. The machine sat on its own steel foundation so it was rigid but getting the twist out of the bed required about 47 mile of piano wire. It took days monkeying with sag tables and exotic apparatus because all the shop's optical alignment equipment was monopolized at the time by the ordinance dweebs. After I got the bed straight and flat all I had to do was square it.

Hm. What ship? I think it was a subtender. Memory's going along with most other powers.

Don't be dismayed if getting a worn machine to cut straight takes some time and fiddling with the adjustments. Even hard core millwrights have been known to lose two falls out of three rassling a worn out machine into some semblance of alignment.

OK, somone explain the use of a home turned proof mandrel made of iron pipe as an aid to secure horizontal and vertical spindle alignment to the saddle motion.

[This message has been edited by Forrest Addy (edited 12-22-2002).]

12-22-2002, 11:22 AM
Forrest says: "Don't be dismayed if getting a worn machine to cut straight takes some time and fiddling with the adjustments. Even hard core millwrights have been known to lose two falls out of three rassling a worn out machine into some semblance of alignment."

I was assisting getting a Landis Grinder set up. One of those "we have ahd the machine for years, haven't got around to installing a foundation yet. But We NEED it yesterday, for a very import emerecency that will occur tommorrow if we don't get it done now" . Meaning "if Headquatrters finds we have not installed this machine for years, and that we can not grind landing gears like we said we could, some bodies lying ass is gonna burn- probably starting at the top and comming down. Rember S**T rollsdown hill".

ANyway we mounted that thing on a concrete floor, very thick (big airplanes rolled on it before it became a machine shop), first moved it from across two slabs of concrete and put it all on one slab. Used optics, piano wires, mirrors, levels to level and aligne. TO list some of the things that affected the machine: It had a bed (13 feet long i think), grinding head travesed and the ways sagged. Put bow in ways so the thing ran level. Before we put it all on one slab, we could actually measure from slab to slab the tides (both slabs did not floatthe same). Even on one slab, a fork lift on adjacent slabs would rock the machine. Had to block sky lights to keep heat effects to a minimum. Finally got it to meet specs (.ooo1 inches end to end). Sent a crew to calif where the men did the work routinely. Discovered they were unable to meet spec also, so (being the weapons system mangers, they just re wrote the specs to match what they could do- but failed to tell us.

A few jobs like this one is why I say "imagine your machine is made of rubber and look at every thing as being a possible source of flexing". Forrest seems to have been down the same road, under even worse conditions- his big advantage was that he "KNEW" nothing was stable on a ship. We had to discover that even the earth moves a little. Also I sure learned that you can spec something out so NO ONE can do the job in a reasonable amount of time. So even the specs are not trustworth.

I think some times what appears to be competance is just paranoia. Trust nothing, no one, check it twice and expect t even the gods will diddle you just to see you sweat!.

Dean Head
12-28-2002, 07:43 PM
A newbie question...I used Rollie's method and got the same results in both the vertical and horizontal...about .0065". What does this mean? I expected to see different results in the vertical as opposed to the horizontal measurements.

Herb Helbig
12-28-2002, 08:56 PM
Both Forest and Docsteve mention piano wire. I'd be happy to hear how to use it in squaring up a lathe.

I was happy to hear Forest mention a lathe on shipboard. When I inquired about making a precision level (not on this BBS) for leveling a lathe, I was told I didn't need one, that lathes were run quite nicely on ships.

Thanks to all for all this wisdom!


12-29-2002, 10:39 AM
Herb: The ship board mounting of lathe is a very old story, its told mostly when some one says you MUST level the machinenery. The end of the story is usualy that level is not important, not being twisted, foundation being rigid ARE important. IMHO, leveling has many adavantages and should be done- but if the level is not at hand- then level as well as possible (which is all you can do with even the best of levels) with what you have.

All the work we do with machines depends on IMAGINARY straight lines from which we measure. The level gives you a easy to transport straight line (or axis) referencedto the earths center (which is pretty stable for most purposes). Some one in a previous post mentioned a bucket of water hung from the ceiling- this makes and excellent plumb bob and since(as theposter mentioned) you can spin the bucket and the water stands still, you have a plumb bob that does uses its inertia to keep twisting and untwisting the line, is easy to use and fast. so here we have two axis easy to get, one vertical, one horizontal. problem is measuring from them to align other things.

But what are you aligning? as some one else says, probably the bed of a lathe is best placeto start. so you put a straingt edge on the bad and it flat! yep till you move the carraige, then it sags- every one sags a little due to the down pressure. can you measure the sag? given the right tools you can. Is it important? depends on the work you haveto produce. Thus we conclude every lathe will have SOME error. If the error is significant we say we have a lathe with a flexible bed (Atlas is a good lathe but very (IMHO) flexible. then when you put cutting tool to work you induce more or less deflection. (I guess rapid deflection is what we call chatter). The work deflects away from its imaginary axis and more erros occur. machinist rests his foot or arm on the tool and more or less deflections occur.

The Rollies dad method (which i used for years before i discovered it had a name- which name i found here) lets you find those axis with precision, but the axis will change under loads so you can't trust even that. Its all a paradox- you gotta have SOMETHING you trust, you can't trust nothing nohow! So i say, think of your equipement, no matter how srtong, well mounted, well aligned, as being made of rubber. It will stretch, bend and return to its original shape. But while in use it is bent some how- all you can hope for it that the bend is not significant or is compensated for.

The piano wire is samll diamter, will take a lot pulling before it breaks. small diameter means little mass, so yu can pull it tight and have very little sag- but it will sag a little. Nice thing about piano wire is the you can pull it at any angle and still have a "very straight" line (depending on what you mean or need by "very straight". Then you measure from or along the piano wire (as best you can) to see if other things are where you expect them to be. Tow or more piano wires crossing each other (which only makes you have two or more straight lines) will show yu that the surface you have is flat or bowed or twisted. Thats especiall useful on a machine where you cant lay a real straight edge on the surface or where the intersection is over a hole.

IMHO optics makethe best straight lines mirrors make best squares. but every light beam has some diameter, so its a long way from perfect.

I guess the REAL starting point is how much tolerance (error) can you tolerate? My advice (and I know no one asked my advice http://bbs.homeshopmachinist.net//smile.gif ) is get things solid enough for you to use. That means "if it looks good nuff for you then try it". Once you turn out a few things that amaze you cause you thought YOU (not the Metal mite, Thrud, John S) could never make you should be happay a while. then you will want better and better and you become obsessed (like some (MOST?) of us http://bbs.homeshopmachinist.net//smile.gif). Then you tweak, tug, twist measure and ASK QUESTIONS. Start out trying to makesome thing rather than installing a machine, then when your desires seem to exceed what your machine will do, first blame the operator and learn how to do it "right" (aint no such thing a "A" right way. Then Blame the guy who set up the machine (that YOU BUD http://bbs.homeshopmachinist.net//smile.gif ), then blame the cutters grinds (that s you ) the cutter materials, last blame the machine and buy a new one. Do it in that sequence and you will cure most problems quickly, cheap and be happy.
Reminds me of the old story about the violin being sold, no one would bid, cause it was ugly. A "MASTER" played it and it sold. The story is usualy told in a religious context. What I always think (having a pragmatic sarcastic turn of mind) is that all though the violin could make good music in the right hands, those who could not see the worth inherant in the violin, no violin would play in the nahds of the man who bought the violin- and the "master" did not buy the violin himself.

Thats why we here can only advise buying fixing trying. The real worth is in the operators hands- and machinst and pilots have two things in common- a mike will show how close he came to perfection, but like landing a plane, if you survive, it was good nuff in the long run.

So sawward- stop if it don't worry you, looks safe, you are happy, make a few chips. ANd from yor question, I bet you have not made many. I had a lot of trouble getting my first chips made right. Tried to show son in law how bad it was, what mistakes I made. First lathe was a good one, presnet lathe is not very good shape. I still could not make the same screwed uo cuts i made on the goodlath. Hands just would not do what ever i had done so worng before. So cut your chips and correct YOUR mistakes then whereevere you see movement, get rid of it and you will progress. Aint gonna make progress till you make chips.

12-29-2002, 01:08 PM
Excellent post Docsteve66. I never know what I am going to learn when I sign on here!

12-29-2002, 05:28 PM
...that and "never fart in a ski-doo suit". http://bbs.homeshopmachinist.net//biggrin.gif

12-30-2002, 12:49 PM
I got a E-Mail reminding me that the question was "how to square a lathe with piano wire. http://bbs.homeshopmachinist.net//smile.gif Us old codgers do ramble on!!!!

One way to check for twist in a bed is to run 2 piano wires from each ofthe four corners (corner to corner). you must raise the one wire above the bed by a known amount and the other wire a known amount above the lower (first) wire. You raise the first wire to clear vee's, dings etc. The second wire is raised to clear the first wire a known amount.

so use four spacer blocks, exactly same height. Clamp your piano wire ends to the spacers and the spacers to the bed. the wires should cross (they make a x some where in the middle of the bed). if the bed is flat (no twist) the wires will touch and each will be deflected 1/2 the wire diameter- which you can not really see with fine wire. So slide feeler gauge leaf same thickness as the piano wire plus .001 on each end of the top wire, back on the blocks. There should now be .001 space between the wires where they cross if there is not, you have some twist.

If the space is less than .001" (maybe the wires still touch), swap the wires (top to bottom etc). You must have a space between the two wires. Course you can get the space with out swapping by deliberatly twisting the bed and putting the warp in the opposite direction. Remember, if the bed has no warp the wires will be .001 inches apart. So twist the bed by jacking some where (the floor is (IMHO), the better place to jack. Floors don't bed. benches sometimes do.

Thats all there is to it. I tried (started) to write how the geometry works, 2 points make a line, three point a plane and flat planes intersect in straight lines , but it got unwieldy. But what you are doing is very simple if you think a moment and do it by thinking - not by following instructions blindly.

12-30-2002, 01:04 PM
If anyone wants to talk about "Rollie's dad's technique" with Rollie himself, I'm pretty sure he's be at Cabin Fever in a few weeks with the contingent from the New England Model Engineering Society. If he runs true to form, he'll have his Congreve clock, 1/4-scale Bentley BR2, some other models, and be happy to talk your ear off. Really nice guy.

Herb Helbig
12-30-2002, 09:03 PM
Thanks, docsteve, for the wire explanation, and thanks to the E-mailer. I was content with the dissertation you provided - figured you wanted me to work out the wire problem for myself. But now I can worry about how to clamp the wires instead of what to do with them!


12-31-2002, 11:30 AM
Herb: clamping those wires is some times a problem. If what you are measuring has its reference surface at or above the rest of the parts, its easy to clamp your wire to any lower part, lay it on the referece surface and stretch it to the next reference surface. when the is no clear shot across you need shims (blocks or some thing).

BTW the reference surfaces do not need to be at the same hieght on both sides. you do need to have those "imaginary lines" (the lines not made by the wires) to be in the plane you are trying to establish). Study the job. Remember "two points" make a straight line. A third point makes a flat plane (no skew). Four points can make a skewed plane andthe skew is what you are usualy trying to get rid of. One of the posters has advised deliberatly skewing for solving a job problem.

I wanted, but never did, to insulate on of the wires and put a bulb, battery, buzzer to let me know when the wires barely touched. Still think it would be nice. Its my personal opinion that a skewed bed on a lathe is of little importance, excepting if the bed is skewed( TWISTED) what do yu have to trust? Tools are like a marraige, if you can't trust your partner (tool or spouse) you will do lots of un-needed worrying. The untrusted tool (or spouse) can not possibly be responsible for ALL the bad things. But it sure makes a convient whipping boy!

I spent much time taking away the excuses for pooor skills. When a man admits his tool are capable of doing the job, and he has materials then the only thing left is the worker. Faced with that, the worker eiather learns or leaves.

BTW when Forrest says a new lathe shouldbe leveled ASAP I agree completely. A new one will have stresses that work out over time. If its level and de-stressed, hopefully the lathe will relax and stand properly. I think you can "lock up stresses" and permanately distort a machine. But I would hate to have to cite any proof of that statement.

01-04-2003, 01:33 AM
on using the wires: i am not getting how the wires are supposed to be orientated or positioned relative to the bed.

would someone who knows mind taking a picture or making a sketch and posting it to like the metalweb news drop box?