What's new
What's new

Larger machine leveling

vmipacman

Cast Iron
Joined
Nov 21, 2014
Location
Virginia, USA
I have a best practice question. Leveling a small bridge mill, travels are X-90" Y-56". Using a Starett 12" precision level. There are 24 leveling screws.
If I level off the table in all areas of travel and checking all corners of the table, then put the level longways on the ways directly I get a different number. The table is level but the ways are ~.003 out of level over 12”. That’s about 6 ticks on the level. The level is .0005 per graduation. Should I be concerned about a difference between table and directly on ways or does un-even wear on the turcite (and generally an imperfect world) account for the table being level but not the ways themselves? It’s a bit like getting a wet noodle to fit a straight line but I don’t have a feel for "good enough". Leveling just the table, I have level reading ranging from dead on to .0015 per 12”(3 ticks) depending on where in the travel I am and where on the table.

Also, since functionally all the action happens directly under the spindle, should I really only focus on the checking level of the table UNDER the spindle rather than the extremities of the table?

Just let me know if i can stop overthinking this now :)
IMG_5818.jpg
 
First, I don't think level is so important on a milling machine, but it is important that the weight is evenly distributed across the 24 posts. Secondly, if the table is level under the spindle, but not on the ends, you have a bowed table and there's not much you are going to do about that anyway. I would advise to level at the table not the ways.
 
Contrary to popular opinion, level does not matter*. Flat and straight do.

There, I said it. :stirthepot:


*Only if you're using your machine for a bowling alley......
 
First, I don't think level is so important on a milling machine, but it is important that the weight is evenly distributed across the 24 posts......

The first part of this is true. It's not really level that you are after, it is straight. Level is just a nice means of checking/adjusting straightness. The second part is not true. Distribution of weight among the leveling feet will differ on a machine like the OP is asking about. The leveling screws under the bridge assembly will carry more weight than the ones at the ends of the table bed. How much difference will be determined by the stiffness of the structure and how the joint between the bridge and bed is contructed.

@OP: Your machine is best leveled off the ways. For the Y axis you should use a ceramic straight-edge something like....

Buy Ceramic Precision Straight Edge - CNC Specialty Store

to bridge the ways so you can remove the twist in the bed. For the X axis, level on the way and work out any crown or sag. Then go back and fix any twist you introduced. It's an iterative process and best done with the table removed. Getting the bridge aligned will depend on how the machine is built and the joint construction. To check and get that geometry adjusted correctly you should use a large granite master square.

Since it appears to be a used machine there may be issues related to wear that introduce odd measurements into the alignment process. Impossible to cover what one might encounter in a forum post. That said you may find it easier and adequate for your purposes to level off the table. The trouble with that is the table will "average out" straightness issues in the ways. This creates uneven weight distribution on the turcite.
 
Very Interesting your assertion to level off the ways! This was my thought as well as all of the above regarding straightness not level being the real goal. Still, level is a good place to start.
But the suggestion that it should be leveled sans bed is a bit off in my opinion. I know of no place that would level the machine by first removing the ~6Klb bed...
So tell me. What tolerance on level should I be OK with as a starting point? Dead on? .003" in 12"? Nothing is perfect, specially on a 15 yr old machine no matter how good condition it's in...
No telling what condition of level its been in for its life. A brand new machine may have been fine to level off table but uneven Turcite wear could possibly account for the difference I am seeing?

Thanks for the straightedge link.

So for this machine, new plan is to level off ways and then just let the table fall where it may. Good plan?
 
So for this machine, new plan is to level off ways and then just let the table fall where it may. Good plan?

Yes, that way you ensure the table will be riding on true ways and not worsening the existing wear. I wouldn't get too picky since its 15 yrs old. Up here in north country our concrete moves around quite a bit with season change, frost, etc. Don't know how it is around you but if I check machines every few months they always need a tweak.

My strategy is to get it pretty close (by then I'm sick of crawling around) and run the machine for a while then re-level after giving it the chance to settle.
 
.....But the suggestion that it should be leveled sans bed is a bit off in my opinion. I know of no place that would level the machine by first removing the ~6Klb bed...
So tell me. What tolerance on level should I be OK with as a starting point? Dead on? .003" in 12"? Nothing is perfect, specially on a 15 yr old machine no matter how good condition it's in.........

Leveling the bed with no table is how they do it to build the machine at the factory. Note that I said "best", not mandatory. With the table off you are not having to move it back and forth to get it out of your way as you straighten out the bed.

Moving the heavy table around will show or mask bed crown or sag falsely if the foundation is not super good. You can check for that by putting the level on the bed proper where you can move the table through its stroke without moving the level. If the bubble moves then the weight of the table moving is shifting the foundation. Almost impossible to truly straighten a bed out if that is happening.

How good it will end up depends on too many things to list. A new machine would get within 1/4 to 1/2 a tick on the vial with a .0005"/12" level. A good condition 15 year old machine can usually get to 1/2 to 1 tick. It is common to have to do it as close as you can now and redo it down the road as the machine "settles in".
 
Thanks for giving me some numbers to shoot for. I think I can hit close, just takes time. I was going to switch to a torque wrench to aid in evening out the loading on the bolts, keeping in mind loading will naturally be more in heavier areas.
Not to contradict, but I would think the slab would need to be especially springy if the table weight made much difference on level. I mean there is already 40k preload on the slab... but I guess if .001” deflection is two ticks on the level it would show up...

I went through this machine pretty well and refurbed or replaced all the typical wear parts. Wipers, way covers, paint, glass windows etc. Looking forward to being all done but not slacking off here at the end.
 
I have 2) 20,000+ lb Mazaks sitting on a 3.5" concrete slabs. 30 x 65 VTC. 9 levelers? Works for my needs. But I am a manufacturer of products.
 
Not to contradict, but I would think the slab would need to be especially springy if the table weight made much difference on level. I mean there is already 40k preload on the slab... but I guess if .001” deflection is two ticks on the level it would show up...

Very roughly, the modulus of elasticity for steel is about 6-7 times concrete, and for aluminum it's about 2-3 times, depending on grade of mix and the way it was poured.

"For example, the modulus of elasticity of steel is about 200 GPa (29,000,000 psi), and the modulus of elasticity of concrete is around 30 GPa (4,350,000 psi). The modulus of elasticity of aluminum is 69 GPa (10,000,000 psi)."

Modulus of Elasticity: Steel, Concrete & Aluminum - Video & Lesson Transcript | Study.com

So especially if the ground prep wasn't great, or the concrete thickness is not right for the bearing weight, your floor could be moving more than you'd expect.
 
...... just takes time........I mean there is already 40k preload on the slab... but I guess if .001” deflection is two ticks on the level it would show up...

When I was doing field service I would have quoted install leveling on a machine like that at 12-16 hours.

When you are trying to get the bed within 1/2 a tick on the vial, chasing 2 ticks worth of foundation shifting will drive you crazy.
 
level your machine on a few pads first, and then bring the other pads to touch. You can use an indicator on the bottom of the machine to see when you get enough load on the jack screw to move the machine a few tenths. That will be far less trouble than a torque wrench. you can also put your level on the foundation itself and move the table back and forth to see how much the foundation is moving, if any. when you put the level on the table you are taking an average reading of the ways under the table. That is why a small jog in the ways doesn't show up on the table.
 
I would get a procedure from the OEM.
Usually there is a specific sequence to leveling a machine with that many feet.
It may be in your installation manual if you have one.
 
I always did it graduatly
First I start with a decent optical auto level on bigger machines (I get it within a couple of 0.1mm with that)
Then get the bubble free eighter left or right With another less accurate level to start with even
Then sneek up to it one tick of the vial at a time No more One person reading the level and another on the bolts

AND CLEAN CLEAN CLEAN the machine and level every time you move it With your hand at best

Also make sure that the machine is not touched by direct sunlight
You will be chasing your tail then

Peter
 
I have a best practice question. Leveling a small bridge mill, travels are X-90" Y-56". Using a Starett 12" precision level. There are 24 leveling screws.
If I level off the table in all areas of travel and checking all corners of the table, then put the level longways on the ways directly I get a different number. The table is level but the ways are ~.003 out of level over 12”. That’s about 6 ticks on the level. The level is .0005 per graduation. Should I be concerned about a difference between table and directly on ways or does un-even wear on the turcite (and generally an imperfect world) account for the table being level but not the ways themselves? It’s a bit like getting a wet noodle to fit a straight line but I don’t have a feel for "good enough". Leveling just the table, I have level reading ranging from dead on to .0015 per 12”(3 ticks) depending on where in the travel I am and where on the table.

Also, since functionally all the action happens directly under the spindle, should I really only focus on the checking level of the table UNDER the spindle rather than the extremities of the table?

Just let me know if i can stop overthinking this now :)
View attachment 292558

.
worn turcite can be like a car with both front tires low on air. the road might be level but car with low tires appears to be leaning when sitting in the car but the car travels on a level road
 
I would level and align the base ways and ignore the table for now. As Tom said the table bottom could be worn. I would build a sled that looks like a T and fasten a precision level on it. As Glen points out start with a few levelers. If I was doing it, I would pick 2 under the center where the uprights are on opposite sides. The 2 on opposite side at the ends do a rough level say with in .001" at first and then tweek it in as close as you can get it. The bed is worn too, so you can't get it perfect. . Many times manufactures use 3 points. Do you have a manual? The sled shown is on a Lathe bed, but the principal is the same. The sled is one piece and as you can see we attached the level so it would not move. This is so much better the stacking up loose parallels and a level. If they move you will get a false reading. 20190916_131544.jpg?
 
I just wanted to close the loop one my leveling project. I did level off the ways and used a fixture plate I already had to mount the level to. Just like Richard shared. I just stuck bolts on three threaded holes and ajusted the feet for level. The whole process went very smooth after that. I used holes for feet that formed a right triangle rather than a equilateral. It make leveling the plate easier.

Got the ways level dead on perpendicular to X travel (across the ways). Parallel to X is one tick off, but consistent on both ways full length so I'm calling it straight (if not level) and not messing with it anymore.

Some final comments, its all stuff I "knew", in theory :). I found the tip to use a mag base and indicator from the machine base to the concrete as a way of visually seeing when that leveling screw had enough load. It worked really well. Very amazing how well that worked. I expected kinda spongy needle response but you could see the instant that foot lifted.
I also put the level on the concrete to test for concrete deflection as the table moved through its travel. I could not find any deflection at all. Worth noting that this machine is on a 4-5" slab, #4 rebar, 2" of foam under, and gravel underneath. Not exactly OEM spec.
Lastly, it is interested how little force is on each leveling screw once the load is well distributed. 40Klbs on 3-4 screws and you cant turn anything, but when its evened out each one feels like it has practically nothing. Again, its obvious but kinda cool to feel first hand.
IMG_5835.jpg
 








 
Back
Top