Question about lathe levelling

All discussion about lathes including but not limited to: South Bend, Hardinge, Logan, Monarch, Clausing and other HSM lathes, including imports

Moderators: GlennW, Harold_V

atunguyd
Posts: 199
Joined: Tue Oct 20, 2009 9:39 pm
Location: Durban South Africa

Question about lathe levelling

Post by atunguyd »

Hey guys, I want to test my lathe for bed twist.
I don't have a machinists level neither do I have anyone who does have one, but I have seen some interesting videos of people doing this with a plumb bob. In my mind this makes a lot of sense, if your vertical rise is high enough you can get a very sensitive measurement from a simple tool. Worked for the ancient Egyptians so why not?

Anyway my question is not about using a plumb bob vs a machinist level, but rather that for both processes I have seen that everyone measures directly on the lath ways.
Is there any reason why one could not put the measuring device on the cross slide carriage and measure with it near the head and then near the the tail stock?
In my mind this makes sense as the carriage will be carrying the cutting tool anyway so ensuring that the carriage does not twist as or travels down the bed is really what we want to measure anyway isn't it?

Sent from my SM-N975F using Tapatalk

pete
Posts: 2518
Joined: Tue Feb 10, 2009 6:04 am

Re: Question about lathe levelling

Post by pete »

Yeah I've watched that Dale Derry YT video about using a plumb bob, and yes with enough height it should be a very usable method. It's one draw back is the time spent waiting for that plumb bob to fully stabilize, and there must be as little air movement as possible within the room it's being done in. Your correct in that the carriage can be used and it's what I do with a proper machinist level and for the reason you mentioned. I can't recall if Dale did so or not, but I think I'd start out machining a wide base 60 degree point to set just below that plumb bob point to have a visual check in two directions of where that plumb bob point is located once it's stabilized. Done correctly with only a few thou between the two points most should be able to detect by eye a deviation to within a thou or two.But even on brand new unworn lathes, that "level" condition no matter how it's done is only a very close condition. Any lathe I've done always required extremely small adjustments very slightly away from dead true to compensate for minor deflections within the machine under it's normal cutting loads. So cutting tests and those slight readjustments are then required.To save a bit of time I think I'd get the lathe as close to level as possible with a standard carpenters level and then start using the plumb bob. And don't expect those level adjustments to remain stable even with a concrete floor. After adjustments, the bed does take awhile to fully stabilize to that new position. I'd recheck after about a week, then a few weeks later and a month or two after that. At a certain point you'll stop seeing any further changes as the bed readjusts to it's new position. At that point you can probably start doing yearly double checks just to be sure nothing is changing. Even heavy concrete floors can move enough to be measurable over some time.
atunguyd
Posts: 199
Joined: Tue Oct 20, 2009 9:39 pm
Location: Durban South Africa

Re: Question about lathe levelling

Post by atunguyd »

Thanks Pete for the detailed answer.

The lathe had been in its location for about three years now. The guys I bought it from did level it when I bought it but as you said it would have settled by now.

The room is small with only one door at the opposite end of where you the lathe is so still air is easy to achieve

My intention is to use the plumb bob method to get it more out less there and then use a cutting test to finalise it.

I have seen some suggest to replace the plumb bob by something much heavier and then put a test dial indicator up against it to measure deflection. I am battling to determine in my mind whether that would work. My initial thought is that the pressure exerted by the indicator would skew results but then I guess that pressure is applied when testing at the head and tail stock so they should balance each other out. I guess though that such a contraption would require that the dial indicator contacts the same area of the plumb every time, which is difficult to do given that it is hanging from a line (I considered making the line rigid but that brings its own set of complications).

Doing a bit of trig, a plumb bob set at 1.2m high (about 4 feet) will detect a 90 second twist (0.022 degrees) by deflecting 0.5mm (just under 20 thousandths). That is easily detected by the eye, actually even half of that should be detected by eye given a sharp enough plumb bob and well defined markings (was thinking of putting a steel rule under the plumb bob).

Being able to mount it all on the carriage and just transverse that makes for less error in my simple mind.

Will try this weekend if family life allows me the time.

Sent from my SM-N975F using Tapatalk

pete
Posts: 2518
Joined: Tue Feb 10, 2009 6:04 am

Re: Question about lathe levelling

Post by pete »

Well as I said, if you can add another point looking directly up at the plumb bob's point and with the bare minimum of clearance between the two, then a very few thou should be quite observable if your eye sight is even half decent. Magnifying that with even a 2x-3x hand held magnifying lens would obviously help. I bought a fairly cheap USB camera that if I remember correctly will go to either 30x or 40x magnification that works with my laptop as a stand in for a proper tool makers microscope. At that level then low 10ths deviation should be easily observable. But please post how well this works for you since I'd be real interested. Some day when I get enough motivation I want to test a plumb bob against my machinist level. Other than the far greater amount of time invested with the plumb bob, I think it's a great alternative if you don't have and especially where you are can't easily get or afford a proper machinist level.

Given how delicate and the amount of deflection force required (basically low thousandths or even millionths of a gram) to affect that plumb bobs or any weight that's logically usable, there's no way I can think of that any spring loaded dial indicator or dti could ever work at all. I used to work in a coal testing lab and know exactly how little 1/1,000 of a gram really is. I've weighed out thousands of samples to that level of accuracy. The only way it could be actually measured would have to be some kind of extremely expensive non contact laser or machine vision system.I really don't think that in an average shop condition anything better than one of those USB cameras would ever be needed. As you said, your doing this to get to an inital starting position and then making your final adjustments by measuring what the lathe produces under actual cutting conditions. But your fortunate the lathe has been in that position for as long as it has. The floor will have bent (and even concrete bends) about as much as it's going to. But it still takes time for the machine to fully stabilize. It can and will take a "set" within the metal that takes a certain amount of time to relax into it's new readjusted position. So expect to make further adjustments as that happens.For anyone who thinks that's not possible, even thermally stabilized high precision gauge blocks don't and can't hold there size. Mitutoyo themselves told me there or any all steel gauge blocks can change up to + - 1 UM per inch/ per year. Everything moves whether we think that's so or not. I've seen more than one post mentioning seeing measurable amounts of deflection movement in 6" thick steel reinforced concrete floors just by moving your body weights position on it.

I've mentioned this PDF quite a few times here before. If you don't already know of it? https://www.google.com/url?sa=t&rct=j&q ... aHqDvetyil I think that link should work, if it doesn't? Google for "PDF of Testing Machine Tools" and it will come up. Non optional reading for what your starting to do. As that PDF shows, lathe bed leveling is in actuality only part of a full series of tests and checks, but it can certainly help to correct a whole lot of minor issues a twisted lathe bed also creates. The rest can be worked on as and when you have the time.

Edited to add, I just checked that PDF link and it does work for me.
atunguyd
Posts: 199
Joined: Tue Oct 20, 2009 9:39 pm
Location: Durban South Africa

Re: Question about lathe levelling

Post by atunguyd »

Progress so far:

WEnt to go look at plumb bobs at the hardware store - best one I could find was a stanley branded one, sharpish point and brass weighing in at 226grams but not exactly what I was after.
I figured that when one has a lathe, one does not go buy a plumb bob so I took some 30mm diameter stainless and turned my own. I knocked up the design in solidworks in order to determine what dimensions will give me the desired weight, solidworks told me the below would be 270 grams and it ended up at 280 grams (granted solidworks did not have 303 stainless in the material library and I may have left the length slightly longer.
I also put a die grinder on the tool post and ground the point as sharp as I could to increase precision.
Unfortunately the roof of my workshop is really low so I could only get a vertical rise of the plumb bob of 950mm

Here's the plumbob setup on the lathe:
20201017_140524.jpg
I placed a steel ruler just below it and some magnets around the rule to hold it in place but still allow me to move it to zero the reading.
20201017_140555.jpg
Next I 3D printed a plastic mount to mount my soldering microscope to the carriage way you can see it to the left of the top image.

Here is what I can see on the magnifier:
20201017_140630.jpg
The graduations on the ruler are 1mm, I think I can easily line up the point to within 1/10 of the graduations like this. By my calculations (simple trig) then I should be able to detect a angle difference of 0.006 degrees or 21.6 seconds which I believe is pretty much on a par with the cheaper machinist levels. Even if I can only detect a 0.2mm difference that is still 42 seconds which may be good enough for a hobby machinist.

Where this really battles is time - Any movement and I pretty much need to leave it overnight to stop that tiny swaying motion. You may also see a sharpie marker on the plumb bob - this is to ensure that it has not twisted between readings as the knot in the string is not perfectly spherical so I suspect that the plumbob would sit at a slight angle (evident by the fact that when I first hung it and the twist in the string caused it to spin it had clear runout).
John Hasler
Posts: 1852
Joined: Tue Dec 06, 2016 4:05 pm
Location: Elmwood, Wisconsin

Re: Question about lathe levelling

Post by John Hasler »

atunguyd writes:
> Where this really battles is time - Any movement and I pretty much need to leave it overnight
> to stop that tiny swaying motion.

Suspend the bob so that it is partially submerged in a pan full of clear oil or water. That will quickly damp the oscillations without affecting the steady-state reading. You can put your ruler at the bottom of the pan and still read it.

> You may also see a sharpie marker on the plumb bob - this is to ensure that it has not twisted
> between readings as the knot in the string is not perfectly spherical so I suspect that the plumbob
> would sit at a slight angle (evident by the fact that when I first hung it and the twist in the string
> caused it to spin it had clear runout).

Since you only want to see deflection in one direction put a crossbar up top and run a string (piano wire would be better) from each end of it to opposite sides of the bob.
pete
Posts: 2518
Joined: Tue Feb 10, 2009 6:04 am

Re: Question about lathe levelling

Post by pete »

Once you get the bed leveled to as well as you can, the next step would be adjusting your tail stocks Morse Taper side to side until it's centered to the head stocks C/L. A dti "can" be used to do so since effects from gravity will or should be equal. What you can't do using that method is check the tail stocks vertical alignment, depending on just how rigid the set is, you can see anywhere from .002" - over .020" due to indicator sag just from the effects of gravity. But once your tail stock is centered then cutting tests can be done with a bar between centers with further adjustments made until the lathe does turn parallel. At that point the bed will be as correct as it's going to get. Then it's time to indicate your tail stocks quill while it's extended and has the quill locked. Indicate along the top and front of the quill using the carriage to move the indicator. At that point my guess is your going to have a few more questions about correcting what your probably going to find. :-( This might give you some answers or at least one method, https://www.youtube.com/watch?v=v6dtsHu ... Eq8EpVXaqS
atunguyd
Posts: 199
Joined: Tue Oct 20, 2009 9:39 pm
Location: Durban South Africa

Re: Question about lathe levelling

Post by atunguyd »

Thank Pete - used the two lines idea and it resolved my twisting problem. Also I exaggerated about needing to leave it overnight - about an hour seems to work.

I have had some good success with this method so far - with the carriage at the headstock wide I zeroed my plumbob out by sliding the rule till the tip was directly in the middle of the 10mm mark on the rule. Moved over to the tailstock and after settling the plumbob was reading 9.2 or thereabouts. I added some feeler gauges directly under the lathes feet on the operator side as indicated in this image:
SIEG-SC8-750mm-HiTorque-Gear-Head-Lathe_1564168.i.jpg
SIEG-SC8-750mm-HiTorque-Gear-Head-Lathe_1564168.i.jpg (9.17 KiB) Viewed 8699 times
Through trial and error with different feeler guages I have managed to get almost the same indication from the plumb bob as was on the headstock side.
I must also add that with the plumb bob and the microscope I could easily see the difference between using a 0.33mm feeler gauge and a 0.35mm feeler guage so a difference of 0.02mm (about 8 ten thousands) under one foot is detectable (even loosening the lathe mount bolts causes a deflection on the plumb bob of about 1mm).

I think the addition of a microscope makes a really big difference here, that and being able to just take photos with my phone to compare the measurements greatly enhances this method.

I found that adding a block of steel near the lathe about the same height as the leadscrew pillow block and placing a 3-4foot long steel bar over this block and under the pillow block allowed me to just leaver the tailstock up by the 0.5mm or so I needed to slide the shims under it.

Now that this is done I want to do the tests that are outlined here:
http://homews.co.uk/page319.html
I will then move onto the tailstock.

I must however point out that I undertook this not because I was turning a taper - I honestly never tested for this (do 99% of my work up near the headstock), but just a desire to do it as I don't believe that the guys who sold me the lathe really did anything to level it correctly.
So I cant feedback if this method has improved anything but I can say whether the lather is turning parallel.
pete
Posts: 2518
Joined: Tue Feb 10, 2009 6:04 am

Re: Question about lathe levelling

Post by pete »

Square non twisted lathe ways are still very important even with short parts for a couple of reasons. It's debatable how much effort gets expended while setting up the bare lathe beds on a surface grinder with the smaller off shore hobby level machines. But ideally and in simple terms the lathe beds should have the mounting feet machined and/or ground on the better industrial machines while the bed is in a relaxed or neutral state, it's then set on those feet and the ways are ground. So running the machine with a bed in that non neutral or twisted condition accelerates wear to the moving parts. But that twist also affects the tools being used due to how the misalignments get amplified. With turning tools locked into a tool holder, any bed twist will move the tool either into or away from the desired tool location due to the amplification or tool height above the bed ways. Visualize that twist in the ways as rolling the tool tip in a partial arc into or away from the part depending on it's direction. It's that effect that causes a part to be tapered from end to end on a misaligned lathe bed because the lathe carriage has to follow and will replicate what ever alignment there is on the ways. But even with short parts the tools used for internal machining, drilling and reaming from the tail stock, boring bars held in tool holders are usually fairly long. Any twist can then easily result in taper drilled, bored or reamed holes.

This might sound highly critical, it isn't meant that way, I'm trying to be factual due to what I've personally experienced with some of what I've bought. A whole lot of these off shore machines look fairly good on the surface. In some cases there at best a lathe shaped object once you start actually checking what the factory produced using the tests and methods in that PDF link I gave you. EVERYTHING does need to be verified on these machines so you do know what it's alignments are. My Sieg C6 lathe as just one example came with the tail stock barrel pointing up hill .009" in just over 2 inches. Add a morse taper mounted drill chuck to that tail stock and even a short center drill and the length and therefore misalignment is over double that. It was impossible to use any center drill from the tail stock since it would instantly break the tip off. Reamed holes were seriously tapered even with a properly leveled bed. While a vertical misalignment at the tail stock has far less effect in creating a taper turned part for logical reasons, due to that serious misalignment it still refused to turn exactly parallel no matter how I adjusted it. So the reality's are that bed leveling is only the first step in what's required as further checking.

Myself I'd never run parallel turning tests between the head and tail stock using any lathe chuck on the head stock end. It's far better to hold a piece of at least 1" diameter scrap material with what ever chuck you want to use and turn a proper 60 degree point on it. That way it's as concentric and with a true running point on the head stocks exact C/L as the spindle bearings can produce. Use one of the sides of the a chuck jaw to drive the lathe dog that's attached to your test piece. Leave that 60 degree point undisturbed in the chuck until all your turning tests and adjustments are complete.

Once the bed is known to produce a properly parallel turned work piece you then have a starting point for further tests. At that point I'd probably verify the tail stocks quill and how parallel it is in both the vertical and horizontal directions. I'd then start checking the lathes facing alignment. All lathes should be set up to face concave and never convex. It's an extremely small amount on even the best manual and cnc lathes made today. That Schlesinger PDF will show the numbers. From memory it's only .001" - .002" concave over I think a 24" diameter work piece for a tool room lathe and a bit more allowable deviation for lesser quality lathes. No your lathe can't turn anything of that size, but there's a trick to amplify the measurements on smaller diameter parts. Did you get a face plate with that lathe? If so spotlessly clean the spindle nose and face plates mounting surfaces. Mount it to the lathes spindle and make a light but full facing cut from the outside of the face plates rim to the inside hole in the face plate using a new carbide tip or extremely sharp high speed steel tool. To use a clock dial as the best way to visualize this you'd be turning from the 9 o'clock position to the face plates center. Once that face is fully cleaned up withdraw the cross slide to your starting position, now set a 10ths capable or metric equivalent dti's tip just on the far side of the face plates center hole. Use the cross slide to run the indicators tip to the back side of the face plates rim or it's 3 o'clock position. Doing it this way will show DOUBLE the measurement readings of what the facing alignment is. Using the same path the tool took to make that facing cut ( 9 o'clock to center) will only follow the exact same path the tool took and your readings will show zero so it tells you nothing.

Well made high end lathes are purposely set up to have the head stock pointing towards the operator by approximately .001" in 10"-12" from dead true on a test bar and another roughly .001" upwards over the same distance. Those purposely set misalignments are done to help compensate for work piece weight and turning pressures. The lathes cross slides are also aligned to help the lathe face very slightly concave and never convex. That's done so the machine starts to slowly wear towards flat facing cuts and not immediately away or towards a highly undesirable convex facing situation. No I don't expect your lathe to be that accurately made, I'm providing those numbers of what a proper lathe should have. What you measure with that lathe might be just about anything but correct. But the Schlesinger numbers can be at least used as rough guidelines. What's generally known as the Two Collar Test can be used as another rough guideline for your actual head stock alignment, mistakes and faulty test results can easily be made just due to a slightly dull tool, taking more than the barest minimum of cuts, and trying to make cuts on a test piece that's too small in diameter. Any flex at all in the work piece while it's being cut and the test is worthless. For that lathe I think I'd chose to follow the maximum accepted diameter part length to it's diameter without tail stock support. 3-1, or no more than three times the part diameter extending outside the chuck jaws. Something like a piece of 3" diameter 6061 aluminum with a 9" length between the two collars being cut would be the maximum I'd feel confident in trying. Again it takes a minimum depth of cut, very sharp tools and a very good trustworthy micrometer to be 100% sure of your test results. Because of it's much lower cutting pressure requirements, then aluminum and not steel would be my choice of material.

What you find during these tests, what your willing to accept or not, your skills or how much effort you want to invest in correcting anything that isn't quite right is going to dictate how far you go with this. But for information about correcting for wear or built in alignment faults then I don't know of a better YouTube channel than this one, https://www.youtube.com/channel/UCD1jVj ... PHw/videos Those Schlesinger test methods in that PDF aren't the only methods of course. There's others where people have come up with different ways to get to the exact same point as Schlesinger does. I chose to use his methods since there proven to work and are still being used today by just about any manufacturer I know of. I hope some of this will help.
User avatar
Bill Shields
Posts: 10459
Joined: Fri Dec 21, 2007 4:57 am
Location: 39.367, -75.765
Contact:

Re: Question about lathe levelling

Post by Bill Shields »

i guess the thought of putting the plumb bob in a small tank of water with the scale at the bottom is.....dark side of the moon?
Too many things going on to bother listing them.
atunguyd
Posts: 199
Joined: Tue Oct 20, 2009 9:39 pm
Location: Durban South Africa

Re: Question about lathe levelling

Post by atunguyd »

No the thought of putting a container of water ontop of cast iron ways did not suit well with me and my luck.

Since we ate on lockdown here and I am working from home it was not a problem to take a 5 minute break every hour to go see the new reading and make adjustments.

The plumb bob method with microscopic is really sensitive, I don't have a machinist level to compare with but I would not be surprised if the two are very close (obviously the plumb bob is not giving me absolute level wrt earth but the two ends of the ways relative to each other are really close now

Sent from my SM-N975F using Tapatalk

User avatar
Bill Shields
Posts: 10459
Joined: Fri Dec 21, 2007 4:57 am
Location: 39.367, -75.765
Contact:

Re: Question about lathe levelling

Post by Bill Shields »

Lightweight spindle or cutting oil would also work. Of course if you are like me...having a pot of cutting oil on the lathe is a guarantee of spillage...
Too many things going on to bother listing them.
Post Reply