Mill column design questions

In the process of designing a small, precision 3-axis mill for processing aluminum. While I can do FEA to evaluate most any design feature, it really is not a substitute for experience. I have seen a few 'rules of thumb' quoted, such as thread depth = 2.5 diameter of the screw, and imagine (and hope) you can share more.

In particular, consider a mill box-column with sides and cross-braces made from 1in x 4in C1018. Deflection calculation seems to depend mostly on Elasticity, which appears to be roughly the same for most steels. Static analysis indicates that worst case deflection should be about 20% of my target. Cannot find any data on dynamic response to differentiate C1018 from any other steel. Or data on dimensional stability over time. Is there some other basis for choosing a different grade of steel or is C1018 a reasonable choice for the intended purpose?

Should the ends of the cross-braces be designed to slot into the vertical sides and, if so, by how much? Maybe 0.1in (as shown)? Static analysis suggests that the worst case forces at the brace end will be 70lbs or less in all directions. So, seems like it would be a good idea -- help to establish and maintain alignment -- but quite uncertain what it will do when dynamically loaded, or whether the slots will cause the sides to warp over time.

Also, how many screws, and what diameter, should be used to connect through the side and into the end of each brace? Cap-head and lock washers or just flat-heads?

Recommendations?

Thanks in advance.

I think you might have a "diving board" effect on the horizontal element if you don't add some sort of bracing under it. Cold rolled steel of any type has a lot of mechanical stress built into it from the rolling process. A slot milled in it will cause it to deform as any machining will. That's why we invented stress relieving. LOL

Dick Z

That was just a detail. Here is a more complete view of the proposed box structure.

Any further suggestions and insights regarding my questions would be appreciated.

Working to refine design. Questions originally posed remain.

i think you would get better rigidity by welding two 6 inch schedule 40 pipes at a 90 degree angle, and then welding on flanges and whatever to get you square mounting surfaces. stress relieve it in a furnace and then scrape the surfaces flat and square and bolt the rest of the kit together.

i see no purpose in making something as heavy as you've designed it, but then consigning yourself to fitting the Z axis rails inside a 4 inch by 4 inch by 4 inch cube either.
consider bolting the Z axis rails to the 4 inch by 2 inch columns for the Z axis (which look to be about 10 inches apart), and turning the bottom 4 x 2 tube 90 degrees to increase rigidity in the major axis.

or consider bolting two 4 inch square tubes together, instead of the 100+ bolts needed as it appears now.

take a look at the parts breakdown for the TAIG milling machines for other ideas.
i would also consider copying their design, but using 4 inch square steel instead of 2.5 inch square tubing.

Hi, if you make a bell from hard brass and then strike it with a hard object ever so lightly, it will ring........the bigger the bell the lower the ring tone......this becomes a factor for resonant frequency, where a small cutter can set up a vibration pattern that defies damping no matter how rigid the structure.

This is probably why cast iron is better than steel for a mill column, but as it's just borderline a composite structure would be more suitable.

This means that if the column was cast out of a solid piece of iron it would resist any form of pressure but would resonate at the slightest opportunity and continue to do so while the force that caused it is present.

You can move a heavy punch bag with the tip of your finger and with continuous small nudges get it to swing in quite wide arcs.

You can knock a brick wall down if you get a continuous force to push it lightly at the right moment if you apply a frequency that the wall is resonant to, whereas a pile of bricks won't move at all as the resonant frequency is variable and always changing.

Making a mill column is the taking those factors into consideration.

The strongest column has to be the pyramid or equilateral triangle, where the two sides are alternately either in tension or compression and brace each other and will resist any attempt to move either any which way.

Why then, with that knowledge, do designers of fabricated mills always design the columns as straight up pencil like structures that flex when the forces are right.
Ian.

Ian, the strongest column is circular cross section. closed boxes can be stronger than I beams in bending, the same material by weight invested in an I beam may be stronger than a square steel box in bending, but if there is the slightest amount of torque, the box is stronger with the circular cross section strongest, and if its death by compression then the tube/pipe will buckle last.

I recently welded an "L" from 3.5 inch schedual 40 steel pipe. The horizontal part of the " L" where the XY axis of my taig will soon reside, is about 12 inches long extending from the inside corner, the vertical pipe is 24 inches.
I happened to find out that the rigidity due to torque exerted on the pipe is approximately the same as the rigidity in bending, when the same load in foot pounds is instead applied to 24 inches of said pipe. This supprised me by about an order of magnitude. I expected the deflection due to torque to be negligible, however, the pipe is thinner than I assumed.

this discovery promted me to look at OP's design again. it has nearly no resistance against forces in the X axis on the spindle, particularly at high Z axis positions.
he would need to weld the ends of those box sections to a steel plate to link the two together at the top, bolting from the inside of the tube won't cut it.

Agreed with Eldon_Joh:
- It would help to space the Z rails apart more. A spacing similar to or somewhat less than the spindle-to-column distance is helpful, so cutter forces are not excessively amplified at the rails.
- 1 big tube is ~always stiffer than 2 little tubes (as long as the big tube's wall thickness is not too thin).
- Copying proven designs is usually a good thing to do, unless you're smarter than the pros.
For a stiffer and simpler structure, just make an "L" from two really big A500 (hot-rolled) steel tubes, forming the bed and column. Bolt or preferably weld together (plus stress relief). Lots of examples of that design.

As Ian alluded, the main goals of professional VMC designers are to maximize both stiffness and damping. Cast iron has pretty good intrinsic damping...
Cast iron VS mild steel - Page 2
but it is difficult or expensive DIY. Two good DIY approaches are epoxy-granite and constrained-layer damping. Both are well suited to a steel tube structure. See the archives for E/G and the Bamberg thesis for CLD.
But a steel frame with neither can work just fine, too, if stiff enough. David deCaussin's little Fadec uses a 6x6x0.5 column and 10x4x0.5 base; he makes no mention of any damping features. Yes, theoretically, there will be some frequencies where it will resonate wildly, but if they are high and narrow enough, not a big problem.

gbrose85, I'm not clear if your goal is more to machine aluminum, or if you really want and have time to make your own CNC mill? If the former, I'd just convert BF20, RF45, or similar. If the latter, here's one good link from the archives:
http://www.cnczone.com/forums/vertic...-cnc-mill.html

Ian and Eldon, I'm assuming you are thinking "stiffness" when you say "strength". To clarify on pyramid vs circular cross section, Ian is speaking of the overall layout. Eldon_Joh is referring to the cross-section of each member. Yes, for bending moments in random directions, round tubes have the best strength/weight. But a heavier frame is good, and round is a hassle for a CNC frame. Yes, square or rectangular tubing is the way to go, as they have great stiffness for both bending and torsion.

Ian, the pyramid design makes good sense for stiffness; I'd assume it's not common mostly because of construction complexity. To make it work, the base-ends of the pyramid need to stiffly tie in to the bed, and I think that part is more difficult (precision design and construction) than the pyramid itself. So a simpler approach is to size the tubes for the highest bending and torsion moments (near the "L" joint), and live with extra material at the ends of the "L". The extra heavy end of the 'pencil' does lower the resonant freq (vs a pyramid), but the mass of the Z axis + spindle dominates anyway. So overall, it's a pretty good design.

Hi, by a pyramid design I didn't actually mean a pointy column as such, but you could see a truncated cone as a pyramid with the top cut off.

My preference for a column is one that has the sides sloping so that lateral forces that attempt to sway the column left to right or forward and back are resisted by the sloping sides being in tension and compression alternately.

A long straight column that is as wide at the bottom as the top has no bracing integrity to flexing and will bend like a spring, even if ever so slightly, which will manifest itself when it vibrates due to the weight of the head being half way up the column and acting like an upside down pendulum.

Torsional twisting is not such a problem as it only needs thicker material for the column sides, and eventually with enough material in the column sides the anticipated strength condition that normal CNC milling needs can be met.

If a round tube is used it would have to be quite a large diam and with a thick wall, and this makes attaching any slides a difficult process unless a large amount of metal work is attached by welding to make a flat area to bolt linear slides on and clear the bulge of the round tube.

A large square tube can be achieved by welding two heavy section channel irons together with a thick steel plate between them to give an internal longtitudenal rib effect.

I think it would be simpler to fabricate from the word go a column to a suitable design by dedicated separate pieces of steel plate as opposed to making a pattern and having the same cast in one piece from iron.

This would also allow you to get away from the constraints of aspect ration as the column attains it's ideal height and base area.

But as a simple plan is mostly a simple exercise to achieve, using off the shelf heavy sections of square steel tubing will eventuate in a design that is quick and easy to build without huge amounts of welding as David DeC showed in the UMC-10 build.

My opinion is that if you skimp on the tubing wall thickness for economy, you will end up with flexing problems later on.

Only for the fact that it is difficult to machine, a column and base in one piece is better than a bolted assembly any day.
Ian.

Stopping resonant frequency in a box column design is too simple. Fill it with concrete(or EP). End of resonance.
If you want to save a bit on weight, put a removable core in the column before pouring the concrete. Now, next question.

Example.Attachment 240698

And before anyone comments on the rails being all the way down, I'm well aware of that, but I just quickly drew this and need to be off to work. This is a 6x6 steel tube with rails and filled with concrete. the inner void is where the core would of been. It leaves a perfect place to run wires through. The angle irons can be bolted or welded before concrete job.

Hi Henry.....I think a 6"X 6" steel tube is too much like a lolly on a stick and would not have much at the base to resist deflection.

At an aspect ratio of 4:1, this would give you a column height of max 24".....any higher and it becomes too slender etc....add some weight at the top, as in a Z axis head and you will get the column moving cement or no cement.

I think you need to lay out the required dimensions of the respective travels first and then dimension the various sections around them.

Once you have a skeleton frame designed you can apply forces at various points of maximum travel to see the reaction.

A starting point could be....table length 400mm......table width 150mm...... cutter to table clearance 400mm etc etc.

With those dimensions there are definite slide way dimensions to carry the table on the slides for both X and Y and with linear ways, to keep them always on the slides as opposed to manual milling where the male dovetails can stick out of the female dovetails, called overhang.,,,,stick out with linear ways is not so forgiving.
Ian.

Took a while to get back to this. The mill is intended for "hobby" use, primarily machining of aluminum. The design is somewhat constrained to components that I already have and the fact that I do not have direct access to any "professional" level machines.

The existing components are
(1) a steel plate 33x12x1"
(2) an XY table mounted to the steel plate - the table surface is a T-slot plate that is 18x6"
(3) two linear rails - THK SHS15 380mm, ea with two bearing blocks, wide type.

Based on the thread comments, and the need to simplify, my thoughts are to weld a 12x6x1" column support extension to the base plate. The column would be a 6x3x24" steel tube with a 1/2" wall thickness. A bottom plate welded to the column and bolted to the base plate will permit alignment adjustment.

I am presuming that the welds, properly done (shopped out), can be made to full thickness, will be as strong as the underlying material, and not unreasonably subject to fracture. I am also presuming that the materials will not warp too much due to welding.

The broad question is, is this design reasonable?
More specific questions include whether epoxy leveling -- under the rail supports and at the column/base mating surface -- is a reasonable alternative to machining flat surfaces.
Size and type of bolts? Most column designs that I have been able to see have used threaded holes rather than through holes and nuts. Is there a preference for threaded mounting holes?

I don't understand the flat plate on the bottom.
What you want is another closed box about the same size and shape as the Z column.

Here is how I built mine:
Attachment 283310
more photos here:
frankenTaig - Album on Imgur

I used the steel I had on hand, didn't buy a single piece.. that's why there's some holes that have been patched in that Z axis.
stress relieving was done in my wood stove.

Here is a google skechup file that is fairly accurate:
millingmachine.skp
if you don't have google sketchup, download the old version 7 or 8, its free.

There is about 1/8th inch of 'system three mirror coat epoxy' between the C channel box and the 20mm rails, as well as beneath the Y axis extrusion.
when i scraped the epoxy to bolt the rails to, all i had was the 12 inch long straight edge from the taig z axis, and the rails themselves (which happened to wring together like gauge blocks.. they were dead flat in all axis. but maybe i got lucky)

i don't recommend system three mirror coat, but it is the cheapest option available for many. the reason why is it doesn't stick to the steel very well. and second, it does appear to creep.
after mounting the rails, wait a few weeks, take them off, then re-scrape them.

also i'd go with rails that have a bolt hole every 30mm instead of 60 if you can find them.

edit:that image was not supposed to be rotated.

Quote:

---

Originally Posted by gbrose85

Took a while to get back to this. The mill is intended for "hobby" use, primarily machining of aluminum. The design is somewhat constrained to components that I already have and the fact that I do not have direct access to any "professional" level machines.

The existing components are
(1) a steel plate 33x12x1"
(2) an XY table mounted to the steel plate - the table surface is a T-slot plate that is 18x6"
(3) two linear rails - THK SHS15 380mm, ea with two bearing blocks, wide type.

Based on the thread comments, and the need to simplify, my thoughts are to weld a 12x6x1" column support extension to the base plate. The column would be a 6x3x24" steel tube with a 1/2" wall thickness. A bottom plate welded to the column and bolted to the base plate will permit alignment adjustment.

I am presuming that the welds, properly done (shopped out), can be made to full thickness, will be as strong as the underlying material, and not unreasonably subject to fracture. I am also presuming that the materials will not warp too much due to welding.

The broad question is, is this design reasonable?
More specific questions include whether epoxy leveling -- under the rail supports and at the column/base mating surface -- is a reasonable alternative to machining flat surfaces.
Size and type of bolts? Most column designs that I have been able to see have used threaded holes rather than through holes and nuts. Is there a preference for threaded mounting holes?

---

Hi, based on your available materials and methods, if you worked to that design you won't go far wrong........it's not the ideal design and could be improved a lot more, but within your capacity to work with metals, this is probably as good as you'll get.

We could indulge in a long and tedious essay on how to alternative methods and how it "should" be made, but in the end it's you who have to do it, so go for it, learn by your mistakes and improve all along the way......nothing like experience to hone your skills.........I guarantee that your second and third model will be so far advanced it would almost be rocket science....LOL.
Ian.

@Eldon_Joh

I don't understand the flat plate on the bottom.
What you want is another closed box about the same size and shape as the Z column.

I went through several iterations looking for something simple and (I hope) sufficient. One of the first was close to your design, though retaining the large plate for the added mass. Then a constructed box and then just the welded on extension. Moved to the bolted 2 plate configuration thinking that the benefits of adjustability outweighed the added complexity. In any case, FEM seems to indicate that the simple welded on extension won't measurably deflect -- well under a mil with an applied 150lb static load.

Is there any reason that stiffeners or a box support could not be added later?

Is there a build thread for your Frankentag? What level of precision are you getting (deflections at the spindle nose when milling different materials)? What size linear rails are on the vertical axis? Did you look at the SC-15P epoxy by Precision Epoxy Products?

@Ian
Thanks for the comments. If there was one change to be made, what would make the most difference in terms of ultimate precision?

Thanks!

No I don't have a thread for it. deflection at the weldment the spindle is mounted to is not measureable once the two tenths of slop in the bearings is taken up. but the spindle is very sloppy because it is aluminum, 6203 bearings, etc.
The next thing for me to do is make a dovetail plate from steel and bolt that to the z axis. but I think I'm just going to build a new spindle. I really can't decide what bearings to use though.

Because the rails are about 7 inches apart on centers, (8 inch wide channel iron) and the bearing blocks are 10 inches apart and the spindle is 8 inches from the column, whatever slop is in the bearings is the amount of slop at the spindle, approximately.

The rails are LS20, light preload.

at the time I didn't have any knowledge sc-15p existed, i used mirror coat because my dad already had a 1.5 quart kit, and didn't need it.

i would concider it acceptable but.. i have read over at practical machinist that all epoxyies have creep problems. so, its useful for some machines, but if you're chasing a tenth then its not.

within a week i should have a 12 by 18 surface plate and i'm going to rescrape the epoxy surfaces for those rails, as well as the epoxy surface the Y axis bolts to.
so i should be able to tell you how much the epoxy has creeped in the last year once i pull the rails off.

I would take the large 1" base plate and turn it 90 degrees. Then mount both the column and dovetail both to that single plate. It would be stronger then having that small plate at the back with the column mounted to it and bring everything together in one structure.

Ben

@Eldon_Joh: Thanks for the added information. Very encouraging. Have you thought about changing to larger bearing balls in your blocks to increase the preload? I have read of others doing it but, of course, have no experience myself. Would be very grateful to hear how well your epoxy performed.

@Ben: Not sure I understand your suggestion. The planned extension is 12x6". If the large plate is rotated 90 degrees in the XY plane the only thing changed is the absence of the welded joint -- the large plate would extend back the 6" necessary to replace the extension. The large plate is the same 12" width as the extension. If the weld is as strong as the original material I do not see the difference. What am I missing?

First it will keep you from welding the small plate to the large one. Second it will simplify the base to one piece that everything attaches to. There is no benefit in the way you are doing it just extra work and more to go wrong. It will not make the table any more rigid so I am just suggesting it is not necessary. I guess I just don't see the point in doing it the way you have designed over just using the one plate.

Ben

Quote:

---

Originally Posted by gbrose85

@Eldon_Joh

I don't understand the flat plate on the bottom.
What you want is another closed box about the same size and shape as the Z column.

I went through several iterations looking for something simple and (I hope) sufficient. One of the first was close to your design, though retaining the large plate for the added mass. Then a constructed box and then just the welded on extension. Moved to the bolted 2 plate configuration thinking that the benefits of adjustability outweighed the added complexity. In any case, FEM seems to indicate that the simple welded on extension won't measurably deflect -- well under a mil with an applied 150lb static load.

Is there any reason that stiffeners or a box support could not be added later?

Is there a build thread for your Frankentag? What level of precision are you getting (deflections at the spindle nose when milling different materials)? What size linear rails are on the vertical axis? Did you look at the SC-15P epoxy by Precision Epoxy Products?

@Ian
Thanks for the comments. If there was one change to be made, what would make the most difference in terms of ultimate precision?

Thanks!

---

Hmmmm......tricky question.........but my two pennoth worth would have to be..... abandon the XY table set up as it's dove tails and they are not the ideal CNC moving medium.....keep the column as is and weld a base plate to the bottom for bolting.......make a bottom rectangular tube extension to form an L shaped base that the column bolts onto, welding a 20mm thick pad to the end face to bolt the column down on...........cut the old base plate to the same length as the new base tube extension and have it 50mm wider than the L shaped base tube to form the mill base with 4 holes for bolting down to a bench etc

That would give you a basic mill structure.....add two steel strips 20mm thick X 50mm wide to the top of the base tube to mount the Y axis rails on etc etc.

David Decaussin's UMC-10 mill is something like this configuration..

Basically it's two pieces.....a column and a base .......it's as simple as you can get.

The base steel tube needs to be wide enough to directly mount the linear rails on with steel strip pads.

As you're going from scratch, I'd make the column and base from the same steel tube with the thickest wall you can get.

It's difficult to give dimensions for the column and base etc, as it has to be proportionally derived from whatever mill work envelope you decide on,........ I'll attach a sketch of a typical fabricated all welded steel design I think is almost self explanatory........it's main components are 10mm steel plate, but steel tubing is also in the same category........in this design the column is bolted internally from the bottom without the normal exterior side flanges.

BTW, in the sketch the column can be a just a simple square steel tube with two 10mm thick angled side plates welded on for added resistance to flexing......this will give you added mass where it counts most.

The cost of materials compared to iron castings is minimal if you can do the welding yourself or a with a friend.
Ian.

Hi Gb - Did this machine get finished and any comments on its performance? Build photos? Peter