12" machinist level

[Dabbler]

I agree. After moving I suggest checking a week, a month and 6 months after.

 

[Tozguy]

The 2-collar test allows us to dial in that last bit of resolution we need to make us feel comfortable that the lathe is as level as we can get it.

If the work you are turning is the best indication, what does checking periodically with a high resolution level reveal that the dimensions of your work piece do not?

 

[mikey]

If I detect a taper that is significant, leveling it with the level first is much faster than trying to dial it out with a 2-collar test.

 

[GeoffO]

This is an intense thread.

perhaps a laugh?

 

[GunsOfNavarone]

This is an intense thread.

perhaps a laugh?

I'm sure its a bit rough for Mikey, but I don't think there are any hurt feelings anywhere. Not speaking for Aukai here, but trying to explain levels that a gnats hair can throw into a tailspin, IS rough for both sides. This has been a lesson in how incredibly tight tolerances can be in machining. I thought I got it.... I didn't. I need to start at a lower level of accuracy, this is not where I'm at yet. My table that my lathe SITS ON TOP OF is level (ish?) My lathe passes the collar test at .003 difference. I'm gonna stay here for now...and get a less sensitive level of course!

 

[Aukai]

I had not realized that there was a vector of level, and the standard level is a gross representation of "level". My collection of bubbles, and needles are in the 5 pound short handle sledge hammer realm of "level". My CHEAP .0002 level is in the jewelers hammer realm of "level" , and I need a persuader that will connect the gap. Accuracy is a challenge for me, if it fits it ships. Learning to come in on size means something now, and trying to get there is frustrating. It will take seat time to learn. My tool height bar came in at .906 on a .910 "0" + or - tolerance, that .0004 is not going to grow back on

 

[mikey]

I'm sure its a bit rough for Mikey, but I don't think there are any hurt feelings anywhere.

Nope, no problem on my end. Dabbler already knows I have nothing but respect for him, always will.

What is frustrating for me is that there is nothing actually difficult about this level calibration thing. It's just setting up carefully and turning an adjuster. Leveling a lathe is also not difficult if approached methodically but that whole thing was a digression.

I was trying to make this simple but it seems to have gotten much more complicated than it should have been.

 

[graham-xrf]

I was trying to make this simple but it seems to have gotten much more complicated than it should have been.

[Sigh!] I only just came across this thread, and it has still 5 little squares of past postings to click before I get through it all. It has explored the leveling deal to the limit of what can be done with bubbles on glass. What to do if the thing you are working with shifts about just because you walked across the room, or did six weeks of heavy rough cuts?

I get it that true level is a beast to get right, and has to be addressed, because the stuff we turn/mill will show the effects, depressingly confirmed by the double-ended test bar taper test. It leads me to ask..

1) If you have leveled (say) across the ways at a headstock, and the tailstock is supported in the middle on a third point, and you check level at the tailstock end, and you discover it is out, does that mean there is a true twist that has to be scraped out? Can you just deliberately set the supports to "untwist" it?

2) Suppose you adjust from under the whole lathe supports (4 points this time), and you tweak away until it is all level. Have you really left one support with zero or some little force under it, and in effect set up force to "untwist"?

3) If you do tweak on the leveling supports, do the twists and bends from the forces you set up take the iron out of shape immediately? In 5 seconds? In 10 minutes?

4) Suppose the machine has been "out of level" for ages, and has been turning tapers. Then you set the level to what looks correct. Has the iron "taken a set" ? Did the twist become permanant? If not, how long does it take to "relax out"?

I am embarrassed to ask these questions. In my recent work, I have used absolute encoders and electronic clinometers which work between 50 and 400 milli-arc seconds, but I still don't know how the restoration fine piece of semi-steel lathe bed from mid previous century is going to behave. I am getting it that my shop/outbuilding/shed/man cave (early construction stages) had better have concrete under it, and if a wooden floor, then I had better cut through and set blocks down to substrate.

 

[Dabbler]

@graham-xrf Well you have identified a bunch of factors affecting lathe bed twist, which is great. It signals that you understand that these questions are just the beginning of the things that affect lathe bed twist.

To address your main points: A solid foundation is essential if you are trying to get a consistent and accurate two collar test, else the world will shift and put your lathe out.

The world of accuracy has to be tempered with acceptability. Chasing tenths on the 2 collar when the lathe only weighs 800 lbs is like shooting a firearm at a target 500m away with a loose barrel. On my smallest lathe I only go for .002 on 12". On my bigger, heavier lathes that would be the beginning.

Now if your primary hobby is to rebuild the lathe and chase the ultimate precision, have at it. I like making parts, and even with a curvy lathe bed it is possible to do good work.

Most of the previous thread was diagnosing the calibration of an ultra sensitive level, which did not suit the intended purpose and current skill set of the owner.

 

[graham-xrf]

@graham-xrf
Most of the previous thread was diagnosing the calibration of an ultra sensitive level, which did not suit the intended purpose and current skill set of the owner.

Thanks @Dabbler . You get it immediately what I was after. I never had a problem with calibrating. Finding the line that is level on a flat surface by rotating the level is easy. Turning the level through 180° and set back against the line provides enough. It can be used like that, if you keep in mind the offset. Or, you can start tweaking to get the bubble centered if it isn't already. I have three levels. The best is shown in the picture. The other two are lower cost Chinese with aluminium body. All are 0.02mm/m or about 4 arc-seconds). I need to get at least one a bit less sensitive.



What you say about the size/strength of the machine in relation to the accuracy it can hold is interesting. We come to the question of whether the machine simply "rests" on its supports, as opposed to gaining strength/rigity from being anchored solidly onto bigger stuff.

Starting with a lightweight machine, and by your guideline of 800lbs, mine are definitely featherweight, it will have some intrinsic strength to maintain it's shape even if just plonked down on a level surface.

It gains a whole lot more if it is hard bolted to "A" frame or other steel structure hidden in the bench under it. If the final anchor to the floor is by Lewis bolts, or resin-anchors into concrete with steel mesh and some rebar, the lightweight machine has become a heavyweight, and if the steel between it and the floor is designed right, it can also become a very stiff thing. I would not bother to go to such lengths. The accuracy(s) you mention are quite enough for my level of skill. Even so, any place I am going to bolt down a real machine is not going to rely on a springy set of floorboards!

The bench, if wood, has to be substantial, laminated, stiff, and of good design. I would not trust a lathe set down on a wooden bench that was not at least capable of having a level read the same from day to day, if set down against the same pencil line. As for floor. If it is wood, then consider at least setting the machine down on it's own strong single board, and perhaps putting some strategically placed bricks between floor and terra firma.

In some ways, the bigger and more mighty the castings, the more one has to pay attention to the huge force effects of it's self-weight. I would guess that setting up "level" again when the machine has been working "bent" might look OK only for a short while, as the iron "un-distorts", and you may have to keep doing it over in smaller increments until things go stable.

Maybe someone with experience of having his iron distort from getting out of level can say what happens when he tries to get it to "go back".