[Zach_G]

Well, after a couple years of putzing around, I've finally came up with the parts and design for a CNC mill I'd like to build. I've contemplated everything from moving bridge gantries to hexapod layouts, but I think this is a good merger of all my design goals:

• Extensive aluminum milling and light steel work
• Large work envelope, at least 10" Y travel
• Compact and lightweight (less than 350lbs) because I move around alot
• Affordably built, with aid of school machine shop ($2k budget)

I guess I should include some features of my design:

• Mini mill spindle driven by treadmill motor ala JFettig's conversion
• Precision ground ballscrews
• Triple stack nema 34's on X and Z, 23 on Y driven by Centent drives
• All steel construction, 3/4" plate and a 7x4" rect column with 1/2" wall thickness
• Travel: 16"X , 10"Y, 13"Z
• Estimated 290lbs
• Bolted construction so I don't have to deal with stress relieving
• Massive 35mm rails on Z, 15mm on others
• Bellows for way protection

The immediate reaction I expect is why the heck is the Y axis attached to the Z? Well, originally while messing around with C frame designs, I was having a hard time keeping everything compact and light yet maintain ridgidity. That 10" of Y travel really put the spindle way out there, so why not have it move along its own support structure llike a half built gantry? With the components I've purchased, this really works out well I think and enhances each of my design goals.

One of the things I worry about most is maintaining ridgidity while milling metals so I took the opportunity to do some FEA in CosmosWorks. The following pictures show the resulting displacement using a 200lb force (purple arrows, randomly selected a force that seemed big) on the spindle 6" above the table and restrained at the interface to the mill base. The spindle tip displacement comes out to about .05mm.


One of the nice things is that this displacement represents a worse case scenario with the Y fully extended. As you can see most of the deflection is in the twist of the Z column, but as the Y moves in there is a smaller moment arm for the spindle force to act upon which drastically decreases deflection. Plus as the Z goes down, the column length decreases as well. I did a comparison FEA with a more standard C frame design utilizing the same components under the same conditions and settings which resulted in about .08mm deflection.


So there you have it, my rather unorthodox design beats out a similarly built C frame mill in ridgidity while being lighter and more compact. This assumes 200lbs is a significant force in milling. The majority of parts have already been purchased, I'd list the cost but I'm too afraid to total it up! Hopefully I'll start building this thing in within the next 2 weeks. Still need to get a small ballscrew for Y axis. I'd also like to find a real milling table instead of using this 3/4 plate. Seems a bit flimsy especially after milling T slots but at least it's ground flat on both sides already.

[handlewanker]

Hi, I don't think your design is so unorthodox when you look at past and present mill designs.
Take your design a step further and lengthen the Y axis slide, add a new Z axis column and you have on of the worlds best jig borer designs, the Swiss Societe Genovese, every jig and toolmaker's dream machine.
The Genovese had a built in rotary table, at least the one I'm familiar with, but then there are probably different models and combinations.
As I see it, there is no gain in doing away with the second column, except a bit of weight loss, but the rigidity gained with the twin design is tenfold, and so we are back to the flat bed router concept.
I reckon if you included a rotary table in your design then it would open up a whole world of possibilities.
As far as weight is concerned, forget it, when you cut down on structure to save weight, it will flex at the worst possible moment.
Ian.

[handlewanker]

Hi Zach, just saw your comment at the end of your post about the 3/4" plate being a bit flimsy after milling tee slots. You gotta be a masochist to mill tee slots in steel for that length.
As you are cutting the edge of technology by fabricating the set-up, why not fabricate a tee slot table as well?
All you need to do is use a 3/4" plate as the base, less if you want to save weight, and not machined yet, and add steel strips to form the bottom and top of the "tee".
The centre part of the tee is drilled with 1/2" holes down its length, pitched about 2" apart and staggered off centre to avoid warping. This is then welded through the holes to the 3/4" plate.
The top of the tee is also drilled and welded in a like manner, and welded so that welds in the now "top of the table" are proud of the surface to allow machining and grinding back to give a flat surface.
If the tee slot material is carefully lined up and afterwards licked out with an end mill, then it is to all intents and purposes a solid table.
The only trouble with using steel for working surfaces as opposed to cast iron is that an indentation on steel displaces material and throws up a burr, whereas cast iron gets an indentation that goes beneath the surface, unless it is a heavy dent that will raise material.
Even if you decided to go for a cast iron table the tee slots would still have to be machined out, whereas the fabricated job is just a light clean up to make the slots parallel.
Milling tee slots for any length in steel will definately seperate the men from the boys.
Ian.

[svenakela]

You're design is pretty close to the Correa A-series I've been working with, but they got the rails on the side. They are sort of automatically protetected from chips then and the "boom" can moved instead of the head. Those are reeeeeally nice machines.
I found a picture an old A10.

[digits]

Well, after a couple years of putzing around, I've finally came up with the parts and design for a CNC mill I'd like to build. I've contemplated everything from moving bridge gantries to hexapod layouts, but I think this is a good merger of all my design goals:

• Extensive aluminum milling and light steel work
• Large work envelope, at least 10" Y travel
• Compact and lightweight (less than 350lbs) because I move around alot
• Affordably built, with aid of school machine shop ($2k budget)

I guess I should include some features of my design:

• Mini mill spindle driven by treadmill motor ala JFettig's conversion
• Precision ground ballscrews
• Triple stack nema 34's on X and Z, 23 on Y driven by Centent drives
• All steel construction, 3/4" plate and a 7x4" rect column with 1/2" wall thickness
• Travel: 16"X , 10"Y, 13"Z
• Estimated 290lbs
• Bolted construction so I don't have to deal with stress relieving
• Massive 35mm rails on Z, 15mm on others
• Bellows for way protection

The immediate reaction I expect is why the heck is the Y axis attached to the Z? Well, originally while messing around with C frame designs, I was having a hard time keeping everything compact and light yet maintain ridgidity. That 10" of Y travel really put the spindle way out there, so why not have it move along its own support structure llike a half built gantry? With the components I've purchased, this really works out well I think and enhances each of my design goals.

One of the things I worry about most is maintaining ridgidity while milling metals so I took the opportunity to do some FEA in CosmosWorks. The following pictures show the resulting displacement using a 200lb force (purple arrows, randomly selected a force that seemed big) on the spindle 6" above the table and restrained at the interface to the mill base. The spindle tip displacement comes out to about .05mm.


One of the nice things is that this displacement represents a worse case scenario with the Y fully extended. As you can see most of the deflection is in the twist of the Z column, but as the Y moves in there is a smaller moment arm for the spindle force to act upon which drastically decreases deflection. Plus as the Z goes down, the column length decreases as well. I did a comparison FEA with a more standard C frame design utilizing the same components under the same conditions and settings which resulted in about .08mm deflection.


So there you have it, my rather unorthodox design beats out a similarly built C frame mill in ridgidity while being lighter and more compact. This assumes 200lbs is a significant force in milling. The majority of parts have already been purchased, I'd list the cost but I'm too afraid to total it up! Hopefully I'll start building this thing in within the next 2 weeks. Still need to get a small ballscrew for Y axis. I'd also like to find a real milling table instead of using this 3/4 plate. Seems a bit flimsy especially after milling T slots but at least it's ground flat on both sides already.

I like it - have you considered having the Y-axis rails in the horizontal plane, with the spindle sliding between them? Wouldn't that put the spindle axis right in the centre of the Z-column and so help prevent column twisting?

[digits]

Hi, I don't think your design is so unorthodox when you look at past and present mill designs.
Take your design a step further and lengthen the Y axis slide, add a new Z axis column and you have on of the worlds best jig borer designs, the Swiss Societe Genovese, every jig and toolmaker's dream machine.
The Genovese had a built in rotary table, at least the one I'm familiar with, but then there are probably different models and combinations.
As I see it, there is no gain in doing away with the second column, except a bit of weight loss, but the rigidity gained with the twin design is tenfold, and so we are back to the flat bed router concept.
I reckon if you included a rotary table in your design then it would open up a whole world of possibilities.
As far as weight is concerned, forget it, when you cut down on structure to save weight, it will flex at the worst possible moment.
Ian.

I don't suppose you have any pics of such a beast?

I have been thinking about a dual column design myself - I can't decide whether the whole Y-axis bar would move up and down for the Z-axis, or whether a 'quill' type Z-axis on a fixed Y-frame would be better.

One problem I see with the twin, fixed column design is that it makes loading and unloading the table rather hard if it's in a coolant enclosure, and I suspect tool changing is probably a pain when you've got a column between you and the spindle.

[Zach_G]

Thanks for the comparison to other machines ya'll, nice to see it's a tried and true design. On the other hand I guess I can't claim any sort of ingenuity points, darn.

Ian, good point about the double Z columns, even attaching a small undriven secondary column would drastically decrease deflection due to torsion in the main column. It was a hard decision but I think I'll stick with my current open architecture to allow easier access to the table. If needed the Z column can be "beefed" with more steel or filling with some sort of epoxy-cement like has been discussed in other threads. Also, thanks for the reality check on milling T-slots in steel. Right now I'm looking at buying a replacement table for Grizzly's G3358 Mill/Drill: http://www.grizzly.com/products/G3358. Also sent requests to Harbor Freight to compare prices, probably hear from them Monday. If it doesn't turn out I'll definitely go the fabricated route. Do you think bolting the T's from the bottom could work in place of welding?

Digits, I speculate that the gains from moving the spindle in between the Z column would be offset by the fact that the column would need to be set back further from the table to accomodate the spindle. Still, thanks for the idea!

[RotarySMP]

The Swede had good success with a Palmgren table:

http://www.5bears.com/cnc09.htm

[handlewanker]

Hi digits, I don't have any pics of the Genovese, but if you went on the 'net and entered "Societe Genovese" I'm sure something would come up.
There was one on Ebay USA a month or two back, so they are around.
As far as the moving Y axis beam is concerned, you would need twin screws coupled together to drive the beam up and down, a fair bit of weight compared to just moving a long Z axis quill on a fixed Y axis beam, also much simpler.
Loading the work table only occurs when the table is moved out from under the beam and is out in the open.
Ian.

[handlewanker]

Hi Zach, if you decided to go with the single Z axis column, you could always cross brace the front of the column as it's done on a horizontal mill to increase the rigidity of the overarm when doing heavy milling with a slab milling cutter.
As far as bolting the Tee slot table together, I think it would work but introduces the problem of solidity due to the pull of the tee nuts when bolting jobs down.
You really want a homogenous mass and an unbroken surface to the table top, hence welding.
I would rather, instead of screwing, use rivets through the lot with countersinks on top and bottom and make sure the top of the rivets were sticking up when clinched so as to allow clean up afterwards.
Using square tubing for structures makes for a fairly rigid and light frame.
To further increase the rigidity of a square tube, something in the order of ten fold, 1/2" diam holes are drilled across opposit corners of the tube and spaced about 2" apart down the tube for a 4" square tube.
1/2" diam steel rods are pushed through the holes and welded across the corners through the holes.
The effect is similar to a cardboard box with no ends.
It has no rigidity and will flex sideways with very little efort, but if you brace the opposite corners with a series of through rods it becomes a rigid structure for very little extra weight.
The principle of the through rod is like the third side of a triangle which opposes the spread of the other two sides by being in tension, same as a roof beam.
The tensile strength of steel is about 27 tons per square inch and this opposes the corner elongation of square tubing when use criss crossed in this manner.
It makes extra work, but beats filling the tube with concrete or epoxy mix, and is heaps lighter.
Ian.

[Zach_G]

Ok, so I contacted Grizzly and they replied the next day with a quote of about $193 for their mill/drill cast iron table. Not bad considering its size. Harbor Freight on the other hand quoted $90 for the same table including shipping, however I literaly had to beat the information out of them over a series of several emails and phone calls. If this project wasn't so cost driven I would have gone with Grizzly as their customer service is impeccable, however I can't justify the cost difference just for an easy sale, heck most of this project is based on how hard I can scrounge. The other unfortunate news is unless they can pull a table off a returned mill, it will likely take 6-8 weeks to ship overseas from the manufacturer.

Ian, you're just full of helpful ideas. I revised my CAD model in CosmosWorks and ran the simulation. Although I didn't see an order of magnitude difference, it did cut deflection due to twist of the column by half. Perhaps the benefits are more prominent under higher loads or maybe the FEA can't accurately model the physical process.

It sure would be interesting to see how this stacks up against the ridgidity of other hobby bench mills. If anyone has a CAD model of a machine I could send it through CosmosWorks. Besides, I've got nothing better to do while waiting for that mill table sos I can finish the design.

[handlewanker]

Hi, I wonder if there is a home industry just waiting to blossom producing TEE slotted tables for DIY mills etc.
All you'd need is someone with a yen to melt cast iron and if you get your pattern right, that's another business opportunity launched.
Ian.

[Zach_G]

Long time no post. I recieved a rather heavy box from Harbor Freight, turned out to be my milling table I ordered about 6 months ago! It arrived just in time for me to be heading home for the summer so no work for another 3 months or so... I opened it up and of course its bone dry and rusting in patches from the long ocean voyage. Guess I can't complain though for $90 from China to my door. It's the 8x28 table and weighs a good 75 lbs.

I was playing with the CAD some more and came up with a gantry style machine that you see on the left, along with the previous design on the right:

Despite the added ridgidity, there are some things I don't like about it:

• Aspect ratio of the Z axis bearings is nearly 3.3:1. Not wanting to endure the added cost and complexity of a ballscrew drive on both columns, I worry about racking with just one screw.
• Way and screw covers on the Z axis wouldn't be fun, and the screw is a bit unprotected.
• It's a tad bit heavier
• Work piece/vise has to fit between columns

Overall I feel the C frame is cleaner and more attractive to look at, but that's just my artistic side speaking. Also for popular interest, I included a picture of the Societe Genevoise that was mentioned earlier (It might not actually be a Societe, but it's nearly identical)

[digits]

Thanks for posting the pic of that Societe Genoveise design - despite searching I never did find a picture - which makes it even more ironic that that's the design of the machine I am part way through building! I daren't really ask if that machine was well regarded!

As for your two designs - I would agree that the bridge would be better, if more expensive, with twin screws on the Z-axis. I don't see that the ways are any more difficult than usual to protect from chips though.

With the C-shaped mill design, won't having your spindle axis outside of the ways, and away from the column tend to make it want to tilt the whole Y-axis out of the X-Y plane? And won't having the Z-screw on the other side of that rail closest the spindle simply make that rail a pivot point? Fair enough, you have a second rail to constrain the motion, but to my untrained eye, it appears that you have a design that creates an unecessary torque, and then uses strength to fight it - couldn't you for example move the whole Y-assembly between the two Z-rails?

[sdopp]

Milling the slots in a plate for your table automatically gives you stress in your table. ( I speak from experience)

[handlewanker]

Hi Sdopp, quite right, anytime you interfere with the natural grain structure of metal it'll want to move.
I was pondering this problem of the Tee slotted table design last year, and thought about the "continuous metal" method of construction or rather manufacture.
As far as I know you can get cross sectional shapes in cast iron similar to allumunium extrusions, but not so slim in the cross section.
I was thinking of a lathe bed design, that had the raised vees and cross section with the grain structure going longways.
You could do the same for a tee slot table I suppose, and just clean up the "extrusion" with not much metal removal.
Looking at Zach G's second design with the gantry type construction and rising bridge, I would definately go for the twin ball screw on the Z axis to raise the bridge evenly, even if it added a bit more to the cost.
Counter balancing should be considered as very necessary.
That is if the spindle were fixed and the bridge did all the depth control.
The gains would be enormous, the cost is only once.
I can't see the merit of a long Z axis spindle design as you will still have to raise or lower the bridge to get the spindle close to the work without the spindle being stuck out and deflecting under load.
It would be better to incorporate the design aspects of a turret mill whereby the spindle moves a relativey short distance and the knee rough positions the table, but in this case the bridge, with it's one screw, would bring the spindle within range of the workpiece, and the spindle does the actual cutting depth adjustment.
At the end of the day you pay a price for rigidity or the lack of it.
Take a mill/drill, with the round column.
If anyone has tried to put one of these to real work they'll have to put up with the deflection under load.
If we're talking about wood, plastic and alluminium, then that's a different story.
Most people start off with this in mind and get a bit ambitious to try steel.
You only want to build a machine once, and rigidity is the first and formost factor to be considered.
With the C frame construction, consider the sections of a Bridgeport mill that are required to mill without deflection.
Ian.

[Newby2]

With no offence to anyone, and the information contained in the posts, has anyone given any thought as to how the old fashioned radial drill press is designed? What is trying to be designed has already been done decades ago. Just my 2 cents worth.
Steve