Lathe leveling - do you do it with the lathe fully dressed or just the headstock and ways?

[jwmelvin]

The additional bolts at the top of the frame seen in post #30 can be adjusted to twist with respect to the frame. They effectively are jack screws. Now the 1/4" may be a bit flimsy, but there is some adjustment besides the feet. Is it better to have one long thick bar picking up the two bolts, or two independent (cross) bars?

They are indeed jack screws. I am of the view that it is significantly better to have those screws bear on two different cross bars--one at the headstock end and one at the tailstock end of the bed. That way you are not adding to the twist stiffness of the structure that you are trying to twist.

Can you elaborate on this? Understand the jack screw part. And the plate positioning relative to the lathe. What are you recommending for the reinforcement of the existing plate/tubing setup? Do you think any of this can be done "in place", with simple levering, or do I need to totally lift away the lathe? By each foot, what are you referring to?

I meant each foot of the bed. That lathe bed has two feet, each with a single bolt through its center. So your task to impart twist is to apply a moment on one foot relative to the other. That requires supporting loads from the foot into some support structure, through the structure, and into the other foot.

You are trying to twist the frame. Ideally, a round section has the best ability to support that load. So I could imagine a length of pipe running along the bed along the top of the frame, with two plates that are independently adjustable, to which the lathe feet are bolted:


Now, I fully understand that is not a structure you are necessarily prepared to build into your frame. So let's think about the easiest way for you to achieve torsional rigidity. This is one design for the frame:


The truss structure gives stiffness to the two long sides; twisting the two ends relative to each other requires bending each side in the vertical plane, so that stiffness would give you more what you want. But wait, that's a lot of fabrication. So some have suggested sheeting your frame with a skin. That seems much easier, and accomplishes much of the same thing (a few braces along with the sheet cover would be better, depending on the thickness of the sheet). One thing I'll suggest is not leaving the front of the frame as just a simple cross member. That does give you some nice room for storage, but you give up too much stiffness in my view. I'd want at least the upper portion reinforced, as my truss structure does. It's not so interesting to draw a sheeted frame, but I'm happy to do some more sketches to the extent it helps.

To sum - there are two aspects. One is making a torsionally stiff frame. I think the easiest way for you to do that is add metal sheeting to the sides. I think you would get significant benefit from some additional tubes tied into your existing top tubes, but there's certainly diminishing returns. The second aspect is affixing your lathe bed to the stiff frame in a way that lets you twist one foot relative to each other. Jack screws and pads can be used to do that, but as RJ points out, you need the jack screws to bear on something stiff enough that it isn't undermining your efforts.

 

[WobblyHand]

They are indeed jack screws. I am of the view that it is significantly better to have those screws bear on two different cross bars--one at the headstock end and one at the tailstock end of the bed. That way you are not adding to the twist stiffness of the structure that you are trying to twist.


I meant each foot of the bed. That lathe bed has two feet, each with a single bolt through its center. So your task to impart twist is to apply a moment on one foot relative to the other. That requires supporting loads from the foot into some support structure, through the structure, and into the other foot.

You are trying to twist the frame. Ideally, a round section has the best ability to support that load. So I could imagine a length of pipe running along the bed along the top of the frame, with two plates that are independently adjustable, to which the lathe feet are bolted:
View attachment 385646

Now, I fully understand that is not a structure you are necessarily prepared to build into your frame. So let's think about the easiest way for you to achieve torsional rigidity. This is one design for the frame:
View attachment 385649

The truss structure gives stiffness to the two long sides; twisting the two ends relative to each other requires bending each side in the vertical plane, so that stiffness would give you more what you want. But wait, that's a lot of fabrication. So some have suggested sheeting your frame with a skin. That seems much easier, and accomplishes much of the same thing (a few braces along with the sheet cover would be better, depending on the thickness of the sheet). One thing I'll suggest is not leaving the front of the frame as just a simple cross member. That does give you some nice room for storage, but you give up too much stiffness in my view. I'd want at least the upper portion reinforced, as my truss structure does. It's not so interesting to draw a sheeted frame, but I'm happy to do some more sketches to the extent it helps.

To sum - there are two aspects. One is making a torsionally stiff frame. I think the easiest way for you to do that is add metal sheeting to the sides. I think you would get significant benefit from some additional tubes tied into your existing top tubes, but there's certainly diminishing returns. The second aspect is affixing your lathe bed to the stiff frame in a way that lets you twist one foot relative to each other. Jack screws and pads can be used to do that, but as RJ points out, you need the jack screws to bear on something stiff enough that it isn't undermining your efforts.

I greatly appreciate the detailed explanation. The figures help immensely. The skinning would be relatively easy, I could even add an X-brace over it, and some trussing. Not sure I could weld it straight (sure I could) but it might not be that straight afterwards. But that is not for discussion now.

Trying to wrap my head around the part that that lathe sits on, in my case, the 7"x46"x1/4" plate which is welded to a 2" square tube. In your tube/plate example, notionally, how would the tube attach to the frame? Does that need to have adjustments as well? At the moment, I'm still thinking if there is a simpler way of beefing up the lathe support.

 

[jwmelvin]

In my tube design, each lathe foot bolts to the center of a plate. The thicker the better. The trick, then is to mount that plate to a corresponding plate affixed to the frame. What I sketched was admittedly incomplete and, upon further consideration, suboptimal.

Each foot should get a bar spanning across the frame, bolted to the foot in its center. (Solid bar would be great, and we can consider alternatives and dimensions later, to maximize stiffness/$.) One one end of the bar you have two holes and on the other end one hole.

The frame gets threaded stand-offs (a.k.a. jack screws), which may be hex-head bolts with heads pointing up, corresponding in location to the three holes in the foot bar. Those threaded stand-offs are themselves threaded for a smaller fastener to screw in the end of the head after passing through a hole in the foot bar.

So now at each lathe foot, you have a bar attached to the foot and that bar sits on three points, each height adjustable. I’d put the two in back, and they can be used to twist the foot bar until it’s upper surface aligns with the lathe foot (before the two are bolted together). Then the front one becomes how you adjust rotation of the bed about its long axis. Because you have a foot bar at the headstock end and one at the tailstock end, you can apply a twist to the bed should you desire. When adjusting heights if any of the stand-offs, I’d loosen the bolt securing the foot bar to the stand-off.

A further improvement would be radiusing the top surface of each stand-off and using a hemispherical nut on the top side of each foot bar. The idea, which is consistent with securing the foot bar with only three points, is to minimize internal stress in the foot bar. Of course, using the foot bar to impart a moment on the bed means the foot bar will have internal forces, but I think it’s a benefit to prevent additional such forces, to make adjustment as predictable as possible.

I could imagine using just one stand-off at each end of the foot bar, and now that I think of it, I think it’s better. (Sorry for going indirectly here.) The foot bar will align with the foot’s bottom surface, and then one stand-off on each end will control the bed twist (leveling).

Making the pats for this will be fun. I may be way out of line to suggest the radiused head stand-offs and hemispherical nuts. You’d be fine just bolting with washers. The foot bar is going to be nearly coplanar with the three stand-offs before you tighten them together.

I can sketch for you. Just thinking it through a bit here first.

 

[WobblyHand]

In my tube design, each lathe foot bolts to the center of a plate. The thicker the better. The trick, then is to mount that plate to a corresponding plate affixed to the frame. What I sketched was admittedly incomplete and, upon further consideration, suboptimal.

Each foot should get a bar spanning across the frame, bolted to the foot in its center. (Solid bar would be great, and we can consider alternatives and dimensions later, to maximize stiffness/$.) One one end of the bar you have two holes and on the other end one hole.

The frame gets threaded stand-offs (a.k.a. jack screws), which may be hex-head bolts with heads pointing up, corresponding in location to the three holes in the foot bar. Those threaded stand-offs are themselves threaded for a smaller fastener to screw in the end of the head after passing through a hole in the foot bar.

So now at each lathe foot, you have a bar attached to the foot and that bar sits on three points, each height adjustable. I’d put the two in back, and they can be used to twist the foot bar until it’s upper surface aligns with the lathe foot (before the two are bolted together). Then the front one becomes how you adjust rotation of the bed about its long axis. Because you have a foot bar at the headstock end and one at the tailstock end, you can apply a twist to the bed should you desire. When adjusting heights if any of the stand-offs, I’d loosen the bolt securing the foot bar to the stand-off.

A further improvement would be radiusing the top surface of each stand-off and using a hemispherical nut on the top side of each foot bar. The idea, which is consistent with securing the foot bar with only three points, is to minimize internal stress in the foot bar. Of course, using the foot bar to impart a moment on the bed means the foot bar will have internal forces, but I think it’s a benefit to prevent additional such forces, to make adjustment as predictable as possible.

I could imagine using just one stand-off at each end of the foot bar, and now that I think of it, I think it’s better. (Sorry for going indirectly here.) The foot bar will align with the foot’s bottom surface, and then one stand-off on each end will control the bed twist (leveling).

Making the pats for this will be fun. I may be way out of line to suggest the radiused head stand-offs and hemispherical nuts. You’d be fine just bolting with washers. The foot bar is going to be nearly coplanar with the three stand-offs before you tighten them together.

I can sketch for you. Just thinking it through a bit here first.

Sounds interesting. Any sketches would be appreciated. Interesting to "hear you thinking aloud".

A question that I have is if the structure above (or something like it) can be "orthogonally tuned". Orthogonal tuning means the adjustments only weakly interact, or ideally don't interact at all. Sometimes having too many adjustments (done inexpertly, say by me,) is worse than a relatively fixed suboptimal solution. I know I'd like my lathe to be dialed in, but, if it is good enough to perform decent work, that's fine for me now.

 

[jwmelvin]

A question that I have is if the structure above (or something like it) can be "orthogonally tuned". Orthogonal tuning means the adjustments only weakly interact, or ideally don't interact at all.

What a fabulous proposal. From my prior world, this is one of the design axioms - to maintain functional independence. Here, I might define three functional requirements for adjusting how the bed affixes to the frame: (1) rotation about its long, z-axis, (2) rotation about the cross-slide, x-axis, and (3) twist.

It would be nice to have three knobs to turn, each adjusting just one of those requirements. Doing so makes the design a little more challenging. It’s possible but will require much more space and material, off the top of my head. Given how infrequently one needs to make these adjustments, I suggest that it’s not worth doing. But you could, for example, have the frame sit on a single foot in the center of the tailstock end. Adjust that foot for rotation about the x-axis. Adjust the other two feet (headstock end) in opposite directions for z-rotation. By sitting on three points, the frame never twists due to the floor and therefore should be much more predictable for twisting the bed. You could simplify the foot-bar adjustments by making one fixed and just one screw on one end of the other adjustable (that’s the twist adjustment).

This design’s getting better. And easier. Thanks.

 

[WobblyHand]

This design’s getting better. And easier. Thanks.

I'm sure everyone has run into a design or contraption that has 10+ degrees of freedom, or strongly interacting adjustments. From my perspective, they are a royal pain and truly not good designs. They may be inexpensive to manufacture, but they are dreadful for users.

 

[Larry$]

cutting 24" of steel would be a challenge with a 4.5" grinder

You will surprise yourself. A cutoff blade on a grinder is pretty easy to use.
I have a silly idea, just try using the lathe as is. Might be good enough for a start or even finish. Later you can beef everything up i you feel the need. Adding mass to the stand is a quick improvement.

 

[WobblyHand]

You will surprise yourself. A cutoff blade on a grinder is pretty easy to use.
I have a silly idea, just try using the lathe as is. Might be good enough for a start or even finish. Later you can beef everything up i you feel the need. Adding mass to the stand is a quick improvement.

Not such a silly idea. Nonetheless, it would be good to get the lathe moderately leveled. It isn't yet, but it is close.

 

[jwmelvin]

I'm sure everyone has run into a design or contraption that has 10+ degrees of freedom, or strongly interacting adjustments. From my perspective, they are a royal pain and truly not good designs. They may be inexpensive to manufacture, but they are dreadful for users.

Honestly, in many instances they are just more expensive to design, not necessarily to manufacture.

Here's a sketch of what we've discussed:


Blue frame, gold feet, gray stand-offs, red foot-bars. Here's a section through the tailstock foot-bar:


I drew it with just one of the standoffs (of four) adjustable; that one's the twist knob. (In reality, I'd probably make them all the same but just think about adjusting twist with one.) Then the three frame feet are responsible for leveling overall (one extra adjustment!). There is some interaction between them, and unfortunately the way to deal with that is to adjust in a certain sequence. Adjust twist and then overall level. When adjusting level, if you first adjust about z-axis (the two feet under the headstock end of the frame), then adjust about x-axis (the one foot at the tailstock end of the frame), it will minimize interactions.

Looking at this section makes me think it'd be worth skinning the inside of the frame too, in the upper section, to remove the frame's tendency to twist. As with some other ideas, this may be going well beyond what the lathe requires.

I'm sure there are ways to make this even easier, but this doesn't seem too bad in the current state. The foot-bars could be sections of rectangular tube, which would give you more stiffness for a given weight, but takes up more room and may cost more; it also has a tendency to crush if a bolt tightens through it unless you sleeve the hole. I'd go with solid bar, like 4" wide and 3/4" or 1" thick.

The chip pan will prevent you from accessing the screws that connect the foot-bars to the standoffs, countersunk into the foot-bars, so I guess I'd do those only moderately tight so that the adjustment standoff can still turn. Alternatively, one could thread the hole in the foot-bar and run a long bolt the full length of the standoff from underneath, tightening into threads in the foot bar. That leaves the head accessible notwithstanding the chip tray. Only the adjustable standoff would have to be like that. The other three could get bolts from the top, tight and with threadlocker.

 

[jwmelvin]

I have a silly idea, just try using the lathe as is. Might be good enough for a start or even finish. Later you can beef everything up i you feel the need. Adding mass to the stand is a quick improvement.

Not such a silly idea. Nonetheless, it would be good to get the lathe moderately leveled. It isn't yet, but it is close.

I agree and that's exactly what I would do. My first lathe was a G0602 (identical lathe other than motor & drive) on a stand the prior owner built that could be much better. I made no effort to measure bed twist or even level it overall. I learned a bunch using it and successfully made things that served their intended purpose. My current lathe has a built-in base and four feet, and I now have a sensitive level so I did adjust the feet to remove bed twist and get it overall level.

If you're going to put some effort into making the stand better, I think it makes sense to go about that in the most efficient way possible. Hopefully my ramblings give you some ideas to consider if you do so.