Foundations of Mechanical Accuracy by Wayne R. Moore

[hmw1972]

Very interesting, Rich. And worth replicating and measuring the results to actually see what is going on.

So, based on your 50 years of experience, tutoring by the Red-Headed Irishman, and common sense, how long would it take to establish a “set” in a cast iron casting and after the set was established, how long would it take to resolve?

I would propose an experiment along the lines of my testing thermal expansion using the mu checker loaned (and now given) to me by Carbidebob. I could set up a thermally stress-relieved grey iron 36” camelback casting on 2-4-6 blocks with careful attention to 3-pointing it. Then that setup could be established as “zero” deflection. Then I could load the casting with something like it’s own weight and observe deflection. The setup could be allowed to sit for hours or even a week and the the load removed to see if it springs back and by how much. That would show how much set the casting had taken and then observations made over time to see it relax to zero.

I am thinking loading and unloading with water poured into or drained out of a suitable container could ensure no disturbance of the setup due to load manipulation.

Anyone interested? Actually observing the behavior of cast iron taking a set might be really helpful to our cumulative understanding.

The checker is quite stable with typically only 1-2 millionths drift (likely temp related due to elongation of the probe) over days’ time.

Denis

I would have to agree with Richard on this matter. I have seen it happen on my lathe. I have moved my lathe three times and each time it sat for a while whichever way it was set down on the floor (i.e. no attempt to level or properly support it). Usually within a week or two I would get around to levelling it (taking the twist out of the bed). All three times it behaved exactly the same way. I would level both ends of the bed (using a Starrett 199 level) so the bubble was in exactly the same spot. Leave it for about four hours and check it. Both ends would have moved off of level and not necessarily in the same direction. Readjust the jacking screws until both ends were level again and leave it for four more hours. Once again it would have moved. I would repeat this procedure until the lathe stopped moving and stayed level at both ends. From day to day you could see that the amount the level changed would decrease. Usually it would take about two weeks for it to stabilize. I would check it at six months and it would be very slightly out and I would correct it. At one of the locations I checked it about two years in and it hadn't moved since the six month check. I don't know how long it takes for the lathe to take a 'set' but my experience was about two weeks to get rid of almost all of it and six months (or less) for it to disappear completely. Ken

 

[dgfoster]

I would have to agree with Richard on this matter. I have seen it happen on my lathe. I have moved my lathe three times and each time it sat for a while whichever way it was set down on the floor (i.e. no attempt to level or properly support it). Usually within a week or two I would get around to levelling it (taking the twist out of the bed). All three times it behaved exactly the same way. I would level both ends of the bed (using a Starrett 199 level) so the bubble was in exactly the same spot. Leave it for about four hours and check it. Both ends would have moved off of level and not necessarily in the same direction. Readjust the jacking screws until both ends were level again and leave it for four more hours. Once again it would have moved. I would repeat this procedure until the lathe stopped moving and stayed level at both ends. From day to day you could see that the amount the level changed would decrease. Usually it would take about two weeks for it to stabilize. I would check it at six months and it would be very slightly out and I would correct it. At one of the locations I checked it about two years in and it hadn't moved since the six month check. I don't know how long it takes for the lathe to take a 'set' but my experience was about two weeks to get rid of almost all of it and six months (or less) for it to disappear completely. Ken

So, you’d have to wonder which was taking the ”set”—-the floor or the lathe? Since it is commonly observed that simply walking around a lathe placed on a 4” slab with soil beneath the slab will distort level readings significantly, the floor is likely a source of time-related deflection. Placing a large weight on any section of concrete slab always results in some compression or “settling.”

Denis

 

[IceCzar]

Placing a large weight on any section of concrete slab hound always result in some compression or “settling.”

The extreme and accelerated example being a former employer installing a 100lb Little Giant power hammer without a proper foundation and destroying the floor inside 6 months

How bad deflection and settling gets depends on both age and soil. Then add vibration (or impact as above) I’ve been doing stone masonry as a secondary profession for 4 decades. It can last centuries if you consider soil and drainage. But water always wins.

 

[eKretz]

It's a tricky thing using anecdotal experience, because we don't always have a complete picture of what is going on. Did the machine de-stress? Did the spring thaw hit and then cause the floor to shift? Hardly anybody ever checks all the possible variables.

 

[hmw1972]

So, you’d have to wonder which was taking the ”set”—-the floor or the lathe? Since it is commonly observed that simply walking around a lathe placed on a 4” slab with soil beneath the slab will distort level readings significantly, the floor is likely a source of time-related deflection. Placing a large weight on any section of concrete slab always results in some compression or “settling.”

Denis

I agree the floor can move, I have seen it move while standing next to my older lathe that was on a 4" floor. However, the lathe in questions was sitting on a 24" thick slab that I poured for it. I could see the bubble as I approached and I would stand perfectly still once I got to the level and would watch for any signs of movement. None occurred. The other thing is that the lathe got stable with time, not unstable, indicating that both the floor and the machine were stable. Unless by some extreme coincidence the floor was moving just the right way to counteract the movement of the lathe. Ken

 

[hmw1972]

It's a tricky thing using anecdotal experience, because we don't always have a complete picture of what is going on. Did the machine de-stress? Did the spring thaw hit and then cause the floor to shift? Hardly anybody ever checks all the possible variables.

I agree there are a lot of variables at work that can cause a lathe (or other machine) to move, and a person may not consider (or even know) all of them. However, in my case the lathe got more stable over time, not less stable. Unless the many variables all conspired to work together to make my lathe more stable with time I think the likely answer was that both the floor and the machine were stable over a period of several years. Ken

 

[dgfoster]

I agree there are a lot of variables at work that can cause a lathe (or other machine) to move. However, in my case the lathe got more stable over time, not less stable. Unless the many variables all conspired to work together to make my lathe more stable with time I think the likely answer was that both the floor and the machine were stable over a period of several years. Ken

Ken,

I would think that initial instability to some degree is to be expected. Your relatively massive foundation did much to reduce the random movement seen in lathes parked on 4" slabs. However, soil is a pliable material and a new placed chunk of concrete necessarily will compress it. After a while, IF moisture levels in the soil are constant (clay has amazing expansion pressure and contraction capacity with moisture changes) you would expect general stabilization. Now if you lived in an area that allowed the foundation of the lathe to be placed on bedrock (Granite would be best<smile>), that would take the moisture and compression out of the equation.

Example: For a long time I wondered why a new construction site might have a huge pile of random soil maybe 30 feet high and hundreds of feet in both dimensions piled on it and then a year later scraped away and building commenced. I learned that is a "cheap" way to squeeze water out of soil and compact it improving its structure support capacity. In some sense you made that compression stack using concrete.

I am not diminishing your more-than-usual care in siting your lathe and doing so three times no less. But everything is rubber, especially soil. Still, your lathe, having settled in, is now likely to be much less "squirrelly" than most.

Denis

 

[lucky7]

I’m no expert on bigger machine tools, but I do find it fascinating that the manufacturer provides only half (or less) of the rigidity when they sell you a machine. The rest come from a properly engineered, reinforced, isolated, concrete foundation.

 

[CarbideBob]

I would have to agree with Richard on this matter. I have seen it happen on my lathe. I have moved my lathe three times and each time it sat for a while whichever way it was set down on the floor (i.e. no attempt to level or properly support it). Usually within a week or two I would get around to levelling it ..

There is a company named Toyota. I think they make cars and are somewhat decent at.
They do these things called cells and move machines way more often than my wife moves the furniture.
Settle time? Kiss that goodbye.
We are used to the old way of it sits here to be stable. Others laugh at us and make more parts per hour.
Then we wonder why we are no longer king of the hill in manufacturing.
We are the best... but some old ideas might be holding us back.
All the work and "Wayne r. Moore" so superman in it's time and thought as the bible and tell all.

 

[eKretz]

Yeah there's a reason most manufacturers want a big monolithic block of concrete under the machine. Separate from the rest of the slab.

 

[John Garner]

As mentioned earlier in this thread, by others, there have been a number of highly-regarded machines built with plop-'em-down-anywhere "three-point" supports. That said, those machines have been much closer in size to home-shop machines than they have been to a common-size horizontal boring mill.

My own thought is that the bigger-than-your-average-bear machine tools aren't built with plop-'em-anywhere geometry has two prongs: 1) The full costs of the additional materials would FAR outweigh the total costs of a properly-constructed foundation. 2) The bulk and mass of a plop-it-anywhere large machine tool would be FAR more difficult to transport from the builder's plant, IF it could be transported at all.

To paraphrase something that was also said earlier, there are sound reasons for a multitude of the adjustable support points on large machine tools . . . and assembly fixtures . . .and other equipment that needs 'em.

 

[hmw1972]

I used to work at a gas turbine over haul company. I was back there visiting a few years ago and they had just poured the foundation for a new CNC HBM that would re-machine gearbox housing bores etc. The tolerance on the hole position was very tight (.0002" or so over 36"). The old machine they used for this had finally wore itself enough that it couldn't hold those tolerances anymore. Because of the tight tolerances the manufacturer of the new HBM would only guarantee those tolerances if the foundation was built to their specs. Their requirements were to excavate a hole 12 feet deep, from the bottom of the hole drive piles to refusal (I believe it was a lot of piles as well) and then pour a concrete block on top of the piles 12 feet thick. There were other requirements for steel content etc. but I don't remember what they were. Regardless, for tight tolerance work on large machines the foundation has to be rock solid.

CarbideBob mentioned Toyota and how they constantly rearrange their factories. This gas turbine shop was like that when I worked there. They were constantly moving lathes, mills and grinders but the one machine that never moved was the HBM that did the gearboxes.

 

[magisterFactotum]

As mentioned earlier in this thread, by others, there have been a number of highly-regarded machines built with plop-'em-down-anywhere "three-point" supports. That said, those machines have been much closer in size to home-shop machines than they have been to a common-size horizontal boring mill.

My own thought is that the bigger-than-your-average-bear machine tools aren't built with plop-'em-anywhere geometry has two prongs: 1) The full costs of the additional materials would FAR outweigh the total costs of a properly-constructed foundation. 2) The bulk and mass of a plop-it-anywhere large machine tool would be FAR more difficult to transport from the builder's plant, IF it could be transported at all.

To paraphrase something that was also said earlier, there are sound reasons for a multitude of the adjustable support points on large machine tools . . . and assembly fixtures . . .and other equipment that needs 'em.

At work, we've just received an NHX 8000 horizontal. Tech was looking forward to an easy time leveling it due to the main machine casting actually resting on only three points, with some outrigger supports for the pallet and toolchanger portions. Sitting, as noted, on a foam isolated massive foundation.

Of course, didn't go quite so easily as hoped. Tech had to have some of the alignment spacers reground to dial in the last few microns of alignment.

 

[MCritchley]

At work, we've just received an NHX 8000 horizontal. Tech was looking forward to an easy time leveling it due to the main machine casting actually resting on only three points, with some outrigger supports for the pallet and toolchanger portions. Sitting, as noted, on a foam isolated massive foundation.

Of course, didn't go quite so easily as hoped. Tech had to have some of the alignment spacers reground to dial in the last few microns of alignment.

Curious where the alignment spacers were installed, table shims?

 

[MCritchley]

There is a company named Toyota. I think they make cars and are somewhat decent at.
They do these things called cells and move machines way more often than my wife moves the furniture.
Settle time? Kiss that goodbye.
We are used to the old way of it sits here to be stable. Others laugh at us and make more parts per hour.
Then we wonder why we are no longer king of the hill in manufacturing.
We are the best... but some old ideas might be holding us back.
All the work and "Wayne r. Moore" so superman in it's time and thought as the bible and tell all.

Toyota probably bought machines that don’t require a foundation with their constant moving in mind.
They most likely also have a very thick slab in the entire building.

That’s smart factory planning, I wouldn’t put it past Toyota.

 

[magisterFactotum]

Curious where the alignment spacers were installed, table shims?

I believe they were between the column and base.

 

[John Garner]

I've put a fair number of one-piece steel shims between monolithic foundations and the surface-ground bottom surfaces of Unisorb Lev-L-Line machinery leveling wedges over the last forty years.

Many of those shims were epoxy-bedded on release-coated concrete so that we could return the facility to a flat floor; others were bedded in gypsum cement or even plaster-of-Paris.

As a point of interest, the Loma Prieta earthquake cracked and buckled the vinyl floor tiles covering the joint between a 50-some foot square inertia block foundation supporting our test apparatus and the foot-thick building floor. The test apparatus, with its test object installed, itself rode out the earthquake, then settled down within five arcseconds of its as-aligned position.

 

[MCritchley]

I believe they were between the column and base.

Interesting, it’s not going to be as rigid as a scraped interface.
Don’t get me wrong that’s a great machine, and it will make great parts.

Shims are probably the difference between a Mitsui Seki and DMG Mori.

 

[magisterFactotum]

Interesting, it’s not going to be as rigid as a scraped interface.
Don’t get me wrong that’s a great machine, and it will make great parts.

Shims are probably the difference between a Mitsui Seki and DMG Mori.

You'll love what they do for the microprecision upgrade option then:

My assumption is there probably is a loss in absolute rigidity, but that's accounted for in the design specs. For production level machine tool building, likely worth the tradeoff.