Lucas Horizontal Boring Mill going to Tuckahoe . . . - Page 11

# Thread: Lucas Horizontal Boring Mill going to Tuckahoe . . .

1. ## Bent like a banana . . .

[I posted two posts back-to-back, so if you want to know what happened to the Lucas at Tuckahoe and why I am discussing bent tables, please read the post #200 on the previous page. Wow, I just noticed that this thread went double-century . . .]

Now, I fully expect to start a polite discussion on why machine tool tables bend. I am hoping we can "get to the bottom of this" without too much heat. Like all of us, "I know what I know":

o The older the mill, the more curvature you are likely to find in its table.
o This curvature will be such that the ends droop lower than the middle.
o There is a tendency to ascribe this "droop" to gravity.

I hope I can provide both theoretical support and some direct observations on the Lucas table to figure out what is going on. Here we go:

o The Lucas table is clearly drooping some .018" in its four-foot length.
o Most of the length of the Lucas table is supported by its saddle.
o There is only about five inches of overhang on a five-inch thick table.

Given this situation, I find it hard to believe that gravity is the culprit. I will add that the tendency of metals and other materials to slowly move under load is called creep. I wrote my required paper for my MS on a survey of creep and I was always a bit skeptical about the commonly-held idea of mill tables responding to gravity over time. I discussed this with a close friend who got his PhD at Stanford University specializing in creep. He was adamant that most of the structural metals, including cast iron would not creep in measurable amounts in a human lifetime when cantilevered the sort of distance found in milling tables. For measurable amounts of creep, the ambient temperatures would have to be sizable percentages of the melting temperature. Please allow that the statements I have just made are to show that I have tried to investigate the theoretical end of mill table droop, not to shut off any discussion of theory or practical experience.

So, if not gravity, what causes the droop? In posts on the PM forum and from numerous old-timers, the answer is that normal use (& abuse) of the mill results in a large number of small and not-so-small dings on the top surface of a mill. This peening places the top surface of the table in compression, which makes it get longer, while the lower surface stays the same. For an example of this in practice, one way of straightening bent shafting is to peen the inside surface of the curve.

Back to the Lucas table: My concern is that if the top surface that is in compression is machined off, there may be a tendency for the table to change shape back. Clearly this would undermine the scraping I have done. Even folks that agreed about the peening from use causing a mill table curve were not able to agree on what would happen if this surface was machined off.

I think I have the makings of a good experiment:

o I have just rough-scraped the table ways, so I have a good reference surface.
o I will be planing the top surface of the table enough to clean it up flat.
o I can then use the straight-edge to determine how much the table changes shape.

Stay tuned . . .
Last edited by Archie Cheda; 05-16-2010 at 11:19 AM. Reason: add info . . .

2. Titanium
Join Date
Jul 2004
Location
Reddington, N.J., U.S.A.
Posts
3,262
Post Thanks / Like
Likes (Given)
113
257
Do you liken the effect of the peening to that of a Blacksmith, stretching hot steel by hammering it? Is it simple elongation or are some stresses being induced into the table. What would happen if the table was put though a heat treatment cycle to stress relieve it? Would some of the droop disappear?

3. ## Trapped stresses . . .

Reggie,

I think that the blacksmith comparison is good, but the blacksmith is working on a high-temperature work-piece so hammer blows on one side pinch the whole work-piece in compression and, hence, make the whole work-piece thinner and longer. (I think that there is some curvature anyway and the blacksmith flips the work-piece regularly to minimize this effect.) With mill tables the peening only affects a small depth on one surface -- this causes a curvature because the top surface's elongation is not matched by the lower surface.

The compressive forces generated by peening would be considered to be trapped compressive stresses induced into the top layer. The rest of the table would be resisting these stresses and because there is a much larger amount of material trying to keep the table straight, the stresses resisting the change in shape would be tensile stresses at a much lower level. All of these trapped stresses would balance each other out with the table taking a slightly curved shape.

All of these stresses would be well within the elastic range and this is the reason I am concerned that if I remove part (or all) of the peened material, the stresses trapped in the table will be unbalanced and the table will take an upward curve -- perhaps I should say an "lesser downward curve". I will be inspecting the way surfaces that were scraped flat before the planing, so I should find out how much change in shape results from removing the upper surface of the table. I expect some of the droop to disappear -- this is the reason I interrupted my scraping. (Upon more reflection, I doubt you could stress-relieve anneal out all the curve because the higher stresses in the thin layer are not symmetrical with the resisting forces. Still, I expect some change -- we will see . . .)

A stress-relieving anneal would have the effect of reducing all of the trapped stresses. If all stresses were annealed out, the order of operations would be much less critical. It is clearly a good thing for the machinist to work with stable materials, but life is never perfect.

Archie

P.S.: When making parts out of cold-rolled steel if one does more machining on one side of a thin (<1/4") part than on the other, the part will curve because the cold-rolling process that finishes bar stock traps stresses in both surfaces of the stock. When faced with parts like this when I ran a surface grinder, I would constantly flip the part, so as to minimize the curvature -- if I timed it right, I got to final thickness with a small enough warp that it was in tolerance. My experience with thin cold-rolled stock is that you can never get it perfectly flat by only surface grinding, no matter what your technique. This is where a stress-relief anneal is called for.

Let me also mention that the magnetic chuck is your enemy in this process. The chuck will pull the thin part flat during the grinding, but upon releasing the chuck's magnetism, the part will curve. You can measure its thickness and it will be the same everywhere, but it will not be flat. In these cases, I did most of my grinding by using double-backed masking tape to hold the part. If you could grind off enough material to get through the thin zone where the stresses were trapped, then you could use the magnetic chuck and grind it like hot-rolled. Some electric magnetic chucks can be set to lower power to reduce this problem, but you need to have some blocking pieces with less thickness than the work and larger areas to keep the part from sliding around.

4. Cast Iron
Join Date
Nov 2006
Location
Maryland
Posts
302
Post Thanks / Like
Likes (Given)
0
3
The process would be more like peening a rivet than drawing out material. Light blows expand and mushroom the material to a shallow depth. In this case we also have expansion inside the T-slots from people cranking down hard over the years. These stresses wont be removed. The geometry of the tables and use of the top surface and T-slots lend themselves to causing the slight rainbow arc. Stress relief would help, but some of the dents and deformation are plastic rather than elastic.

5. ## New info . . .

Jacob,

Thanks for contributing -- I never heard the T-slot analysis, but I am ready to agree that it would be a serious contribution to putting the top layers of a table into compression and that machining the top would not remove that portion of the curvature.

I would expect that the two effects (peening damage & T-slot damage) would have a wide variability depending on the type of work and training of the operators. The Lucas table is 20" x 48" and would most likely have large work-pieces dogged down around the periphery, while I would expect that a typical mill table would have two zones between the center and ends where most of the T-bolt action occurred.

There are good and not-so-good ways to use strap clamps. The best way is to have the fulcrum block as far from the work as possible and have the T-bolt touching the work if possible. This reduces the stress that the table sees and reduces the chance of a break-out of the T-slot, while concentrating the clamping force on the work-piece. The "Rite-Hite" (by Progressive Pattern) clamps are convenient, but you pay a price for that convenience because it the T-bolt is often not as near to the work as might be desired -- when you see them being used backwards, you know that little thought is being given to stressing or damaging the mill table.

Archie

P.S. for the engineers in the audience: To add to the elastic/plastic point: All deformation starts with elastic deformation, most of which will be recovered when the load is removed -- the linear portion of the stress-strain curve. When there is sufficient deformation (past the elastic limit), the plastic flow is permanent is not recovered and the geometry of the table is changed permanently.

When peening is the type of deformation, the small dents have zones of trapped compressive stress around them which add up to creating a overall effect of a stored elastic compression in a zone perhaps 1/8" deep. This zone would have the most stored stress at the surface and the stress level would fade deeper below the surface. This means that machining off the surface layer should remove some of this stress and some portion of the "drooping" should be recovered -- the table should droop less. The T-slot stresses would still be present, so that portion of the droop would not be recovered, unless the T-slots are cut.

I will report what I learn this Saturday when I complete the cut on the planer. I plan to do a clean-up cut with a carbide tool, but once I have a surface clear of rust, I will finish up with a positive rake tool that will leave less machining stress in the final surface.
Last edited by Archie Cheda; 05-18-2010 at 12:36 PM. Reason: Add info . . .

6. Hot Rolled
Join Date
Apr 2007
Location
A little north of Houston, TX
Posts
798
Post Thanks / Like
Likes (Given)
3
22
Great info, thanks for sharing.

7. Stainless
Join Date
Sep 2006
Location
Massachusetts
Posts
1,422
Post Thanks / Like
Likes (Given)
52
98
Hi Arche,

The other thing it could be is peening of the "T" slots from the "T" nuts . this would be on the surface opposite the table top inside the "T" slot. The theory is the same as your previous statement. Lots of pock marks put the surface in compression, which essentially makes the top of the table longer than the bottom, and the result is droop.

Dave

8. May not have anything to do with what is being discussed here...but,

I had a old Cincinnati "swap" belt planer 6 foot stroke, I bought many years ago. When I first set it up and got it running, the table had a .008-.010" bow high midways down the table. As long as I knew this, it would plane straight within .001" in 42", since this was the only straight edge I had at the time to measure with. I assume the straight edge was straight, always hung vertically the years I had it.

Also had a old No. 3 B & S mill that I rebuilt. The table on it had a bow of nearly .060" and it wasn't wear. In fact you could see it eyeballing it end to end. I took about .065-.070" to clean up on a 12 foot stroke Rockford planer. The mill table was about 54" long from what I can recall.

But what I'm getting at, machinery builders back then and now, tried their best to machine slides "straight", but what if one miss-clamped the table down not taking a .018 "bow" in account, the surface is machined, clamps released, the "bow is now transferred to the just now machined surface.
Then this slide is matched it mating base, guess what, the base now has that .018" bow in it too! I've seen this scenario happen many times over the years.
I doubt Lucas would let this happen, but what if it did happen?

Ken

9. We had a No. 5 Reed Printice verticle mill years back. If you weren't carful of clamping down parts, the table in fact would deflect and put it in a bind. The table was nearly 5" thick. Of course I sure the true table thickness was no more than 2" thick in the hollowed out area of the casting.
Archie does has a good point on properly placing hold down bolts on any machine table, any person needs to pay attention to this. Too much stress in a casting will break.

Ken

10. ## Progress report . . .

Dave(s),

Thanks for posting. Jacob's T-slot effect is an extension of the general situation that the top of the mill table gets almost all of the damage and it all tends to induce stored compressive stresses that make the top longer than the bottom, creating the bend. In most cases, including the Lucas, taking cuts on the T-slot surfaces is difficult and undesirable. I am planning on making up a set of T-nuts for the Lucas that will be longer than normal and, I hope, be able to bridge over the damaged areas and spread out the load. There are no breakouts and the top of the table looks pretty good now (see second pic below).

First a pic of yesterday's work. After watching the iron oxide surface of the Lucas table take the edge (& a lot more) off of my HSS tool bit, I gave up on being "authentic" and switched to carbide which did fine. Unfortunately the Rockford planer has one operating speed and that is not for carbide. Each of the two roughing passes took over an hour. (Remember, I have a 2-1/2 hour drive each way to get to Tuckahoe). I was almost finished the second roughing pass when we had our late lunch at 1 p.m. and barely got the finish pass done in time to head for home.

I calculated that the Rockford had a cutting speed of 30 surface feet per minute -- appropriate for cutting cast iron with high carbon tool steel and also for keeping its table from going off the end when cutting long strokes. The belts were squealing when reversing the table with the extra weight of the Lucas table on board. The Rockford managed four cycles of the table per minute and the feed for the roughing was 1/16" per stroke -- if you do the math, it adds up. My depths of cut were .010" -- this relatively light cut was chosen to minimize inducing additional stresses in the table and to keep me busy all day so I would not have to any really hard work.

Below is the final result -- I will be scraping the table for final flatness and museum appearance. I was pleased that the familiar-looking pattern of divots in the table were actually the oiling holes with their screws still in place. I'll show the detail of the rudimentary oiling system at some later date. I also found the bung where the hole in the table for the pivot point for the optional rotary table is plugged to keep chips from falling directly on the table lead-screw.

I brought the table home to work on it for a week in my own shop so as to make more progress in the run-up to this years show at Tuckahoe -- I will do a survey a bit later today and report the details.

Archie

P.S.: Ken, I agree with all your statements. The Lucas saddle and table were worn to match each other. I think that as the table gradually took an increasingly larger and larger curve, the upper surface of the saddle wore to match. This wear was definitely encouraged by lack of oil getting to the ways. The table ways were worn much less -- I think that they must have been oiled directly, which would be relatively effective because there were no wipers fixed.

11. ## Survey report . . .

OK -- I laid my four-foot straight edge on the four-foot long table way on which Greg & I had completed rough scraping. (At least we thought we had . . .) This way was at about 10 points per square inch. I had only started scraping on the other way, although it was getting straight.

The first observation is that the straight edge would spin freely about its center, showing that the way that was flat before the planing operation on the top surface was now bowed in the opposite direction as the table was before planing. This is what is to be expected if removing material relieves trapped compressive stress -- so far, so good . . .

In order to measure the curve, I placed the two ends of the straight edge on .010" feeler gages and then used a set of feeler gages to see how much clearance the center of the straight edge had. The result was .005", which (when subtracted from .010") gives a .005" recovery from being curved .018". This is the end of my experiment, but if I were willing to cut more off the table's top surface, I would expect to be able to get a bit more recovery. I also think that the T-slot damage discussed earlier was a major contributor to the "droop" of the table -- maybe as much as half of it.

At this point, the table should be stable and I can complete the scraping of the ways. I brought the table home with me so I can make more progress than I make working only on Saturdays. By not planing the table top first, I have added a lot of extra work -- I hope everyone reading this remembers that the first thing to do is a complete survey and then a careful choice of the order of operations, based on the survey and an understanding of the history of the work-piece.

12. Originally Posted by Archie Cheda

P.S.: Ken, I agree with all your statements. The Lucas saddle and table were worn to match each other. I think that as the table gradually took an increasingly larger and larger curve, the upper surface of the saddle wore to match. This wear was definitely encouraged by lack of oil getting to the ways. The table ways were worn much less -- I think that they must have been oiled directly, which would be relatively effective because there were no wipers fixed.
Archie,
Good point, I didn't think about the lack of lubrication causing abnornal wear...Looking nice!.
Ken

13. ## All too normal . . .

Ken,

I am afraid that wear from poor operator lubrication is very normal -- this is why machine tools evolved in the direction of automatic lubrication.

Archie

P.S.: I presume that the "looking good" refers to the freshly-planed surface. I forgot to put my straight edge on that surface today before I flipped the table, but after the last roughing pass, I found a .002" concavity overall in that pretty surface -- I presume that this was generated by the planer. With this in mind, I will scrape for flatness first, then make it pretty.

14. ## Lucas

Archie,
I've ran a few Lucas mills before, never as small. Never as nice as this one will be!

NEW CARLISLE JOHN

15. ## Awww . . .

John,

You're going to make me blush . . .

I appreciate the encouragement and any constructive criticism -- it makes it easier to keep at this long project. After the Lucas, almost anything I have planned should feel easy . . . I do now hope to have it all back in one piece in time for next year's show at Tuckahoe, although I will still be working on tooling for another year.

Archie

16. Hot Rolled
Join Date
May 2004
Location
Plymouth, Devon, England
Posts
922
Post Thanks / Like
Likes (Given)
3
70
Archie

I spent a few hours yesterday reading through this excellent & informative thread, many thanks to you for your extra efforts in sharing such detail of your very fine work & processes.

Your conclusion that the bow in the table was caused by compressive stress in the table top through the peening effect is very interesting & something I have often wondered about but not been in a position to prove.
The table has bowed in much the same way as an Almen test strip (as provided to prove saturation shot peening specs) I have often wondered if a new & seasoned cast iron table of good quality was saturation peened on all surfaces the tendancy to bow over long periods of use would be reduced if not stopped. Doubtless the peening intensity etc would have to be varied in different areas of the table but there is plenty of knowledge of this within the quality shot peening companies.

regards

Brian

17. ## Fighting fire with fire . . .

Brian,

I appreciate your thought process -- peening as a manufacturing process was pretty far out on the periphery of the processes I taught, so it was mentioned, but not covered in any depth. You obviously know more about it than I do.

I think "preemptory peening" might be a usable process based on your mention of saturation, but I have the following concerns about it being applied to a mill table:

o The saturation is at the surface and the compressive stresses induced into the surface falloff as the distance from the surface increases. I think that the average hit that a table takes from abuse involves more energy and would be creating deeper effects, with a different stress distribution.

o The complex geometry of a mill table not only includes the dovetails, but often there is a "honeycomb" of cored pockets to reduce the overall weight. This problem could be possibly be overcome by designing the table with the peening process in mind, but I think that the mill manufactures would rather sell more new mills to replace the old worn ones.

o I think that the actual abuse would vary widely in both intensity and location. It would be hard to compensate in advance for this. All the same, preemptory peening might well improve things and would be unlikely to hurt anything.

If I were designing a mill table, I would include a sacrificial top layer that was designed to be replaced. It could be made up of thin (< 1/4") hardened strips of hardened steel, covering the strips of the table exposed between the T-slots. Almost thirty years ago I put a 1" thick sacrificial aluminum top on a brand-new Bridgeport CNC that was donated to our school by SME. All our Slo-Syn retrofitted Bridgeports had divots on their table when the students did not hit the "Slide Hold" button in time. It turned out that the much more reliable control (& perhaps some improvement in teaching students how to avoid crashes), resulted in the aluminum layer never getting cut. The aluminum layer had a grid pattern of alternating threaded and reamed holes rather than T-slots. After installation, I fly-cut the surface from end to end with a monster flycutter made by putting a boring bar cross-ways in a boring head -- a wooden stick on the diagonal damped out most of the vibrations from a lack of rigidity.

Archie

P.S.: Discipline is a low-cost way to minimize peening the top a mill's table -- the old-time teachers I learned from in the 1960's laid this on thick: nothing but the work-piece and the clamping accessories were to ever even touch the mill table. They provided a plywood tray that could be parked on the end of the mill table during setup for all other items.

18. Hot Rolled
Join Date
May 2004
Location
Plymouth, Devon, England
Posts
922
Post Thanks / Like
Likes (Given)
3
70
Archie

I most certainly do not wish to dilute this thread, nor your efforts in rebuilding the Lucas, with peripheral issues that probably belong in a seperate thread.

So very briefly

I am indebted to the excellent blokes at the Metal Improvements Co who have given me sterling service & advice for over 20 years

Metal Improvement | Home

As far as I "know" a metal surface that has been saturation shot peened will contain to some extent the effects of a heavier local impact in much the same way as it contains & resists propogation of a surface stress crack.

I have also watched them peen form an Airbus wing with varying intensities & shot size to suit cavities etc.

The side benefit is the minutely sized indentations which provide excellent lubricant reservoirs in much the same way as the flaking operation you apply.

But enough of this methinks......back to your HBM.

regards

Brian

19. ## Adding to the thread from a parallel universe . . .

I received an e-mail communication from a friend following this thread and thought that this portion of his message and my answer might be of interest:

I have taken a basic scraping class from master rebuilder Richard King here in Minneapolis (several years back). When I was reading about the table for the Lucas, it brought back memories of Richard explaining that machine tables will act like a banana when they are machined releasing the stresses - much like yours. He talked for a while about the sequence that should be used. IIRC, he felt that the ways should be machined first, then the top surface, then everything checked prior to commencement of scraping. If needed, a second machining should be performed. He also said that any iron member should be "peened" prior to each scraping cycle by which he meant bopping it with a dead blow mallet while suspended. This is supposed to "relax" the metal.
The memories you shared about order of scraping ring quite true to the experience I am having on the Lucas. I should have done my way planing and top surface planing the same day and then proceeded to scraping. As it is, I did learn that only a third of the original curvature was regained by the planing I did on the top surface -- useful information, gained at the cost of a large number of extra hours of scraping.

The dead blow mallet blows you mentioned would send shock waves through the table. My discussion of creep, made the point that cast iron would not flow appreciably at room temperature, even over centuries, assuming that there are no shock waves. Shock waves would give the cast iron the ability to recover some of the warp caused by trapped stresses. The problem is how many blows and how big a tap -- without a major amount of research, one is likely to end up spending a lot of time and only undoing a small amount of the warp. (Maybe this explains the large number of peening dents at the far end of the Lucas front bed way . . .)

It just came to mind that I have heard that a straight edge that is true can loose that truth if it is bumped too hard during handling. I try to handle mine very carefully for this reason. I am not certain if anyone has done serious research in this area -- although I know that individual companies have worked on solving their own stability problems, most of the time they do not share that information by publishing it, hence making it available to their competitors.

Archie

P.S.: I have more theoretical knowledge than practical experience in metallurgy. Although I have never specialized in metallurgy professionally, I have a master's degree in it and have used my knowledge many times in working as a manufacturing engineer (three years) and in teaching that same subject (20 years). I am always interested in understanding and explaining actual observations of how & why metals behave the way they do. Please consider this to be an open invitation to post a question in this thread or any other I have started. (Actually, post questions for me anywhere you want, but send me a private message to call it to my attention as I am often not catching all new threads that come along.)

20. ## Weekly progress report . . .

Or should that be "weakly" ? ? ?

Nothing dramatic enough to justify a picture, but I got the previously rough-scraped way scraped straight again. As mentioned earlier, the planing operation on the top of the table had resulted in a .005" recovery measured on the previously straight way. Because my planing setup was not absolutely perfectly parallel with both of the ways, there was about .002" more to take off from one end to the other. (As we will see below, the planing was parallel to the other way.) This meant that I had to remove an average of about .003" from the 3" x 48" way -- this works out to .43 cubic inches of cast iron dust on my floor. No wonder it took three days. (The total material removed from both bed ways was only two cubic inches.)

I am still hand-scraping, but I would not want to use a power scraper on these ways, which have a one-inch guide-way perpendicular to one and a one-inch gib-way next to the other. The only difference between the two is that the gib-way uses a set of screws to support the gib and is not a critical surface -- it still could break a carbide blade that hits it. With hand-scraping, it is not so bad, but it still slows things down a lot, plus you need to reach over the 20" width of the table to scrape clear to the perpendicular way, which also slows things down.

The other way has only had a small amount of scraping on it and it has .011" of excess cast iron at both ends and .016" in the middle. An average amount of material to be removed is about .013" -- that is almost two cubic inches, so it would be prudent to haul the table back to Tuckahoe and put it back on the planer. I could have done this before scraping the first way, but when the Lucas table was on the planer table right side up, it was not as easy to get good measurements as you might think. Now that I have the top side of the table machined and one way rough-scraped flat, I can put the Lucas table on the planer up-side-down, get it leveled so that I can sweep an indicator along the scraped way and along the machined top surface (which is now on the bottom -- I will put the table on blocks so that this is possible). I will also sweep the guide-way because it will be a reference for machining the sides of the table.

Once I have the Lucas table leveled and blocked and clamped (from its ends), I will be able to do the following planing cuts:

o Plane the second table way, leaving a scraping allowance.
o Plane the two sides of the table parallel with the guide way.
o Plane the bottom edges of the table where the fixed gibs are bolted.

The first cut will save me days of rough-scraping time. The second cuts are partly for appearance, but having the sides of the table absolutely parallel to the guide-way allows these surfaces to be used as references during setup. I will also consider dressing the vertical sides of the T-slots, but this decision will be put off because it involves flipping the table . . . Finally I knew I would have to plane the bottom-side horizontal surfaces that the gibs bolt to, but wanted to do this using the scraped way surfaces as a reference. The amount of material to be removed is at least that of the total wear and scraped material removed from both the upper saddle ways and table ways they mate with. These surfaces are relieved so that it will be relatively easy to file and scrap them for final fitting.

Archie

P.S.: The show at Tuckahoe is a little more than a month away, so stop by the Machine Shop Museum if you attend and I'll be happy to see you in person and give you a grand tour. Other than a few breaks, I should be near the Lucas for most of the show. Drop me a private message if you wish and I can be looking for you.

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•