Help With Physics: Column Placement on Perfectly Level Floor

[Ranger Mike]

Ok folks i thought i would use this forum for once.
if i have a 2 foo by 2 foot by 10 foot long metal column
and a perfect level floor, and place the column on end so it is 10 feet tall it should be pretty straight and square to the " perfect flat and level floor " surface..right! assume the column is flat straight and square to 0.0005" overall run out..
question, if the one end of the column has an imperfection of say 0.001" out much out of plumb or out of square ( non perpendicular ) will the end of the column be ten feet abvoe the ground?

 

[Ranger Mike]

never mind..i worked it out..

 

[FredGarvin]

I just saw this. I ran into some definition problems from your question. Do you still want a sanity check?

 

[Ranger Mike]

Hi Fred
i dragged out the old surface plate, put a precision height stand on it and put a dial indicator on the gage point two feet up...i slipped a 0.005" feeler gage under thestand and zeroed the dial indicator..the dial read 0.010" change when i pulled the feeler gage out so i got .050" change at 10 feet
unless i am doing something weird??
whadya think?

 

[FredGarvin]

I was just wondering where you were measuring the height. Were you measuring it at the point where the axis meets the top surface or the highest point on the top surface. I would say that either way, the height would have to change slightly if the bottom is out of perpendicularity. If the top is out, then it would depend on the measurement location.

 

[Ranger Mike]

at axis point Fred
the whole thing was to determine if a gantry cmm has non squareness becasue the column are lagged into concrete..and concrete ha curing problems
also the cmm in my opinion has metrology problems because the gantry design has two banks of columns that are not tied togeather
see attached jpeg
what is your thoughts

 

[FredGarvin]

Since you are measuring at the axis point then the perpendicularity issue will definitely shorten the vertical reach of the column. It has to by the cosine of the angle the axis makes with the base. Over that long of a distance, it can go pretty far out.

I would expect that this should be designed with some form of leveling ability between the gantry and the columns. It's tough to see what is going on there in the picture. I have never seen the actual process of setting up a CMM but I do know the two we have immediately at our disposal took the better part of a week to have up and running. I would imagine this would be a longer set up.

I can't see an issue with the columns because they do look like they are tied together at the tops and bottoms. Is that not the case?

 

[Ranger Mike]

The gantry has to be set up with a laser. spirit level will introduce bend due to curvature of the earth. my point is many things effect the accuracy of this system.
one- the requirement of using concrete to tie the two column banks togeather may cause the 0.005 inch variation due to concrete curing. anyone who every built a house sees this after a few months..
two- the metrology ring is not stable since the two columns are tied togeather by the bridge. this bridge has a floater side and the drive side. the drive side has air bearings on top bottom and both sides to control crabbing error.
three- heat is present 20 feet up ( 2 meter Z axis model) and this causes thermal expansion thus more geometry errors.
four- this cmm has concrete, steel, ceramic (z axis) and aluminum ( bridge X axis) all of which have different thermal expansion rates.

 

[Q_Goest]

Hi Mike. I've never seen a CMM mounted to concrete, but then again I've never seen one that large before. What kind of tolerances are you trying to achieve? And how are you getting the concrete so flat? From my experience, a tolerance of 0.005" over a span of 10 feet is at least an order of magnitude flatter than carefully leveled concrete. Once concrete cures, it shrinks, and I've seen concrete pads collapse as much as 1/2" over that distance.

 

[Ranger Mike]

depends on size of the cmm
all cmms have a machine inaccuracy and a length dependent inaccuracy expressed in Microns
so the typical large cmm has a Volumetric accuracy statement of 10 + L / 150


so we have 10 micron Machine uncertainty plus a length uncertainty of Length / 1000mm or
10 micron + 3000/150 or
10 + 20 = 30 micron of 0.0011 " bandwidth on the entire volume.
various Industry standards are used the ISO 10360 being most often used accuracy standard.

and the gantry design does bolt into concrete,,one big potential problem i think

 

[FredGarvin]

The gantry has to be set up with a laser. spirit level will introduce bend due to curvature of the earth. my point is many things effect the accuracy of this system.
one- the requirement of using concrete to tie the two column banks togeather may cause the 0.005 inch variation due to concrete curing. anyone who every built a house sees this after a few months..
two- the metrology ring is not stable since the two columns are tied togeather by the bridge. this bridge has a floater side and the drive side. the drive side has air bearings on top bottom and both sides to control crabbing error.
three- heat is present 20 feet up ( 2 meter Z axis model) and this causes thermal expansion thus more geometry errors.
four- this cmm has concrete, steel, ceramic (z axis) and aluminum ( bridge X axis) all of which have different thermal expansion rates.

A lot of these things there is no way to know about without an intimate knowledge of the design which you don't get from a picture.

I can see the issue with the concrete base. I would imagine that the time should have been spent to wait for total curing and size change, but even then, I would still expect it to move over time. What do they plan to do when the concrete reaches the end of it's useful life in 10 years or so?

 

[Ranger Mike]

i did not know the duration concrete was good for..where can i get more info on diemsional shift and life span..of concrete?
good input Fred..why would you go with the design with a ten year life then need to dissasemble the cmm and pour a new slab and then install and recalibrate it?

 

[Q_Goest]

Hi Mike. All the CMM's I've seen are much smaller and sit on an exceedingly smooth and thick block of granite. I'm sure you're well aware of all that. But what may be needed here isn't a smooth block of granite or slab of commercial grade concrete. It sounds like you're into a very demanding type of foundation that might be better suited to a modified concrete. It needs to be longlasting, but also minimal shrinkage, thermal contraction, etc... Have you looked around to see what kind of specialized concretes are available?

 

[Ranger Mike]

there is no doubt that construction grade concrete will not work...One huge problem is deflection caused by part weight like 40,000 pounds..
metrology concrete must be used with the gantry design..i have a perfect solution but need to explore the disadvantages of the gantry design.

 

[Ranger Mike]

i found a British study of gantry machine tool foundation structural stiffness. seems that you need a 500 ton foundation to give stability to a 60 ton machine tool traversing over the length of the X travel to meet maximum deflection of 40 micron.

 

[Q_Goest]

there is no doubt that construction grade concrete will not work...One huge problem is deflection caused by part weight like 40,000 pounds..
metrology concrete must be used with the gantry design.

Interesting about the concrete. I did a Google search on metrology concrete but didn't come back with anything. Might want to check the spelling.

I guess I'm not too surprised about the 500 tons needed, but that sure is a lot of concrete!

I wonder why these kinds of techniques have to be used at all with the technology we have. Isn't there a technology that can locate a point in space without there having to be a rigid structure to locate that point? Take GPS for example, which uses radiowaves. Does a similar design using lasers exist? Just curious...

 

[Ranger Mike]

metrology is the science of measurement
if you figure 3300 # per cubic yard, it does not take long to get to 1,000,000 pounds
this was as paper done in 1994 where a gantry milling machine weighing 30 tons moved over the foundation and deflection was recorded. to maintain the required straightness ( 25 micron per meter i think) the foundation had to be that hairy...
there is a better solution but looking at all the drawbacks to concrete based cmms
who ca n tell me how to calculate service life of concrete?

 

[FredGarvin]

I have been looking at any references I have on concrete life calculations and can't find anything as of yet. I am interested in this as well so I am going to keep looking.

In what I have seen, I am somewhat amazed on the amount of "on the spot" testing has to be done for high end concrete QC. I can't say I have ever seen this.

 

[Ranger Mike]

i think these are valid