[mcArch]

Spindle: 5.3hp w/ VFD from HomeShopCNC.com
Drivers: Gecko203V's
Mach3
X: 25mm profile rail from PbcLinear.com, 20pitch rack & pinion(20tooth pinion), geared 3:1 with 5mm HTD timing belt & pulleys, Pair Keling 2880 steppers
Y: 25mm profile rail from PbcLinear.com, 20pitch rack & pinion(20tooth pinion), geared 3:1 with 5mm HTD timing belt & pulleys, Keling 1840 stepper
Z: 30mm profile rail from PbcLinear.com, 5/8" ballscrewKeling 1200 stepper
A& C: 4:1 geared with 5mm HTD timing belt & pulley, Keling640 stepper.

CandCNC MP3000 Dthc with ISS2 spindle speed control for spindle. System will either use a switch or have to change plugs for plasma/router mode.

CAD model is AutoCad2007 solids model and is essentially the same as for the 96x48x22 build, but with a few upgrades/mods. The big change will be the design of the Y carriage to increase stiffness and allow better alignment of the Z-ballscrew.

The gantry tops out about 8 1/2' above the floor & the top of the Y carriage is at about 9 1/2'.

[multiplex]

subscribed!

[mcArch]

-Pacific Bearing 25mm Profile Rail is bolted in place temporarily and bearing carriages mounted to mock-up gantry verts. The CNC drill attachment to version 90037.01 had perfect alignment with the PBC rail holes - 60mm o.c.
-Brackets for the motor mount assemblies are drilled & plasma cut from a sheet of 7ga (3/16") plate.
-Motor mount cages are fabricated at base of gantry verts
-Verts are welded up squaring on the leveled and squared base.
-The top of the gantry measures about 8'6" above the floor.
-Trapezoidal box truss tacked in place

video of the drill attachment drilling the brackets on the plate: (other videos of the machine fab. too)
http://www.youtube.com/calvinoarchit.../0/jaYBjBNKQow

-Calvino

[PlasmaGuy]

Nice work!

I look forward to your progress!

Tom

[DanOSB]

What cam software are you planning on using for 5 axis?

nice project btw

[mcArch]

Thanks, let's see how it turns out . . .

I plan on using the CNCtoolKit by Rab Gordon:
http://www.cnc-toolkit.com/index.html with Gmax or 3dsMax
for my machine, but the owner of this machine will most likely use RhinoCam or MasterCam.

I'm working on getting them to at least teach the cncToolKit so that the technology and the ability to make things with the 5 axis technology might begin to be more prevalent.

[DanOSB]

only problem i can see if you are using torch height controller wtih 5 axis.. and cutting bevel edge or something.. it has to move the x or y as you move the torch up or down..

so some settings would have to be changed with mach.. or whatever controller software you are going to use.. it would be very impressive and cool to be able to do beveling 5 axis with torch height controller..

[mcArch]

I thought about that . . . yes, using the THC while doing a bevel cut would be tough, but typically a bevel cut would be on a piece of steel thicker than 1/4" and you would most likely not have much warpage & could get away with not using the thc in that instance. Maybe there's a way to rotate the coordinate system . . . although there would still be quite a problem getting the Z to translate geometry to a combination of motor directional vectors . . . maybe just stick with the first argument that the thickness and warpage would not require thc in that case. Cheers!
-Calvino

[mcArch]

Gantry is rolling. Working on the Y carriage, Z in final drawings/CAM.

http://www.youtube.com/calvinoarchit.../0/SIwlTJE8rc8

The drill bracket on my 4x10 is great! You can see the 3/16" plate on the table bed there with holes drilled for the Y brackets. The tubes being drilled in the photo are for the profile rail on the Z axis - the main structural verticals of the Y carriage.

That Profile Rail is time consuming no matter how you mount it. Holes at 60mm o.c. (2.3622in) over 146"X, 78"Y, & 80mm o.c. for 40" of Z rail. It's definitely worth the time & cost of the project has the budget for it. Glides nicely, plus the end & side seals work very well. One note, it comes with plugs for the counterbore holes, but I think that if one were to have the rail drilled & tapped from the bottom side, it would give you an even longer bearing life because even though there will be plugs in the holes, the plasma dust is so fine that it will collect at the joint between the plug and the steel rail and likely a little bit will get slung up into the bearings with each cycle.

Forging ahead . . .
Cheers!
-Calvino

[mcArch]

5 axis table build.
2010-08-31

Y carriage fabrication
Main rails are 2x3x.125" tubing. 30mm profile rail is bolted directly to inside face of rails. Top bearing plates are bolted to allow for fine tuning and alignment. Lower bearing plates welded. Cable Carrier brackets must be part of carriage design.

Gantry Modification
Issue: Shorter gantry will provide a stiffer, more rigid machine - better cut quality. A/C head, with 13" tall spindle requires additional height to clear underside of gantry when it rotates at full height. Option: re-design Z carriage and A/C head so that it operates out in front of gantry. Allows shortening height of gantry by 10", requires Z carriage to cantilever 10" - strength, deflection and rigidity of Z carriage design allows this. Issue #2: When A/C head is moved forward (away from gantry) the router/plasma head is off end of table at max. X coordinates. Solution: Cut gantry verticals, swap them side to side, thereby flipping the angled geometry of the vertical trusses and moving the A/C head to the X- direction enough for max. X coordinates. Operation: Cut tops of gantry vertical trusses just below horizontal truss, remove gantry horizontal truss & vert. tops, flip the lower vertical trusses to opposite sides, square, level, re-weld. Then complete a second plane of vertical, lateral truss section providing more than twice the rigidity of the lateral truss section due to decreased cantilever plus the additional vertical truss plane.
Complete Y carriage fabrication
Complete the welds, attach the profile rail to the gantry truss horizontals, install carriages, mount & level the Y carriage.

This is just the preliminary assembly of machine. It still have to dis-assemble it, clean it, and paint it, then final re-assembly, installation of electronics, setup, tuning, and finally, test cuts.
Forge Ahead.

[mcArch]

Z Carriage Fab
The trussed Z carriage is made of 3/16" plasma cut steel plates, 3/4" sch 40 pipe for the main verticals, and 1/2" CR round bar. The bearing carriages are bolted using oversized holes to allow for fine-tuning alignment. Shims can be used if needed for alignment as well. The linear slides are 30mm Pacific Bearing models. I could have used 25mm, but they didn't have them in stock & I figured the additional strength would not hurt seeing as how the Z has 10" between support bearings & then cantilevers like 30" - the larger size will help decrease deflection for me on this one. It slides nicely so far . . . & very stiff! Ball screw machining & mounting is next. I still need to find a better shaft coupler, I have a Lovejoy L-090 which is a cast steel spider coupling with a phenolic or vinyl insert, but I'm not crazy about it, It seems like it will develop some backlash . . . Still looking into it, but if anyone knows where I can order a rigid, flexible coupling, maybe in aluminum that is a clamp style instead of a set-screw, I'd appreciate the link & heads up.

Motor Bracket Fabrication
The Motor brackets for the X and Y axes are made from 3/16" plasma cut plates and consist of a plate that holds the motor (with a 24 tooth, 5mm Htd timing gear), has a center pivot shaft, and holds another shaft with a 72 tooth, 5mm Htd timing gear and the pinion gear (check out the cad model images & photos for a better explanation). The gearing is 3:1 to achieve approx. 400ipm travel with a motor speed of 400rpm. The timing belts are HTD 5mm, 15mm wide for the X and L series (3/8" pitch) , 1/2" wide for the Y axis. The shafts for the pinion gear and for the bracket pivot point are milled from 1/2" CR (cold rolled) round bar and use 13tpi for the nuts, use 1/2x1.125x3/8" sealed bearings, and are flatted where the pulleys/pinion gear mount with set screws. I made the shaft flats on my lathe by mounting my drill to the toolholder and inserting a 1/2" carbide end mill, then lock the lathe chuck & pass the mill down the length taking off about .002" ea. pass (I need to CNC the lathe next).

My rack & pinion are just from Mcmaster.com and are 20deg., 20 pitch, the rack is in 6' lengths and is 1/2"x1/2" with holes for #10 machine screws at 4" o.c. and the pinion is a 20tooth pinion with a 1" pitch diameter (giving it PI inches of travel per revolution - hence the 3:1 gearing to get back to approx. 1inch/rev.)

X Motor timing pulleys. These pulleys are 24tooth HTD, 5mm and are to be mounted on the NEMA42 stepper motors which have a 3/4" shaft with a 1/4" key. The pulleys came with 3/8" bore and had to be bored to 3/4" and had to have a keyway broached into them. The shop press is used to broach the keyway and the bore is enlarged on the lathe.

[mcArch]

2010-09-10 Log 05
Complete X motor drive assemblies:
The NEMA42, 2880oz-in. motors are fit onto mount plates and gear rack is mocked up in place to ensure alignment and correct pressure from the compression springs. The motor mount brackets had to be mirrored (swap right for left & left for right) because of the gantry mod. The pinion gear is not centered in the 24" base of the gantry and the rack on the W8 stops short of the zero end of the W8's. The compression spring location had to be re-located to the opposite side of the pivot brackets.

Complete & mock-up Y motor drive assembly.
Mock-up Y stepper on its mount plate assembly, final location of compression adjustment spring. The pinion gear, the timing gears and belt are all located and aligned to ensure correct fit before paint and final assembly.

Z Carriage Ball Screw machining & installation. The ball screw is a 15mm, 0.2 pitch unit from kelinginc.net. It was ordered 41" long and had to be cut to length, machine the ends to fit in 1/2" bearings at the lower end, into a 1/2" bearing in the top end, and also approx. 1.5" of 1/2" dia. flatted shaft end to fit into the shaft coupling. The ball screw is heat treated & has been induction hardened to Rc 58 - 60 (very hard) C5 carbide cutting tools will not cut it. The solution is to use a diamond cutting wheel on a low speed (3800rpm) grinding tool, mounted to the lathe tool holder. The shaft flat is made in this same way so that the dimensions can be closely controlled. The machining appears to have turned out quite well, the Z carriage is assembled with the ball screw, the bearing plates top & bottom allow for fine tuning alignment. The vertical motion is tested with a cordless drill for any tight spots and is adjusted accordingly until motion is smooth throughout the 29" of travel.

[mcArch]

2010-09-10 Log 06
C axis pivot bearing final design & fabrication. The C axis rotates about the Z. The connection between the A/C head and the Z gantry is crucial because it must be a cantilevered bearing. Imagine the front wheel of a car . . . the spindle. Additionally, the shaft must be controlled by a stepper motor and geared to maintain enough accuracy. This solution utilized tapered bearings with a 1.875 ID (These are skf and will have a custom shaft turned from a solid piece o 2" CR (Cold Rolled) round bar. The cold rolled steel has a tighter dimensional tolerance because of the manufacture process. There is less deformation because of the low temperature at which it is formed. The CAD model shows the concept here. There is an upper & lower bearing which sit in races that get bolted with six 5/16-18 bolts to the upper & lower plates forming the base of the Z gantry. What is missing from the cad model are the sides around the uper & lower plates to which the bearing races are bolted. See the last four photos in this set for the cover/shear plates. This forms a steel box section which serves to enclose the bearings, gears, and belts, as well as provide a structural, box section which resolves the high rotational torque forces imparted on the structure by the cantilevered, A/C swivel head. With the Spindle measuring 13.5" alone, plus 2 more inches for clearance of electronic wiring, coolant lines, etc., plus another 2 inches for the structural connection to the main shaft in the tapered bearings, and you have about 20.5" of cantilever. This pivoting cantilever must then resolve the forces from the 5.3hp spindle plowing through aluminum at full depth (whatever that will be) and keep deflection - overall deflection that is - to about 0.005" (not feeling too confident about that number right now). We'll see what it actually comes in at shortly.
2009-09-13 - Caxis pivot shaft/assembly The short shaft is turned down to the tapered bearing I.D. and a 1/2"-13 stud is cut into the top end to secure it to the upper bracket on the Z-gantry. The critical part to the assembly of this composite, especially since it is a welded assembly, is the clamping/positioning system to mock it up keeping the axes in line with each other, plumb, level, and in line with the C-pivot bearing assembly in the Z carriage. The drawing (5th drawing in this set) with the two yellow rods shows how this is accomplished. The 1" cross shaft (A-axis shaft) is made continuous through the router bracket initally and will have the center cut out once it is completed, and with a 1/2" hole bored in its center where the C axis will pass through. A 1/2" cold rolled steel rod is polished down so it slides through the A-cross-shaft at the centerline of the spindle axis and through a 1/2" hole that is CNC positioned and drilled in a plate left in the center of the the base/spindle attachment plate temporarily. This allows the 1/2" rod to be an alignment tool which fits up into the C-pivot bearing shaft which has a 1/2' hole bored into its underside for this purpose. The assembly can then be positioned on the table, clamped, and welded incrementally with short opposing welds to avoid as much of the inevitable movement caused by the heat of welding. This is a very crucial assembly because the warping will likely cause one of the axes to be slightly off-center. I'm looking for a tolerance here of about +/-.030". This should only affect part programs which may use the a/c axes in a 3d rounded surface & may make out of round by a little less than 1/32". Positioning tolerance and deflection should remain tighter than this. The machine is meant for R&D for building construction elements so the tolerance should be acceptable. I would like to see it tighter, but cannot expect too much more because of the process limitations. The next one may have to be milled from solid stock or cast & milled. (gets very pricey very quickly).

[Dustin407]

Hey McArch,

Great Progress!!!!!! I have reviewed all your comments and pictures. It looks like everything is coming along very nicely. I understand the 5 axis machine and how it works but its a completely different beast from a 3 axis table. I wish you luck on your journey. BTW I see you have some 3-D autocad skills...nice!!!!

[mcArch]

Milling the flats on the C axis main shaft:

[nomedia="http://www.youtube.com/watch?v=h3CBsdKQF74"]YouTube - C-axis shaft flat milling.[/nomedia]

Manual motion test A & C axes:
[nomedia="http://www.youtube.com/watch?v=2TvkRLvM84A"]YouTube - A-C axis head manual mock-up[/nomedia]

Spindle initial mounting. The actual spindle is longer than the provided drawings and so the coolant fittings and the elec. plug will interfere with the top of the C bracket. Possibly 90deg fittings for the coolant and the wire connector will be available and solve the issue. The spindle has its center of gravity at the A axis shaft so the motor is not moving the mass of the spindle, only the inertia affects acceleration time. Inserting a spacer under the mount plate of the spindle would solve the interference issue, but would put the center of gravity below the A shaft & increase the torque required of the A motor.
-Mike

[mcArch]

Thanks Dustin, it's definitely a journey . . . life is!

Yea, I've been on Autocad since '96 when I used Autocad12LT to build the additions to Tropicana Field (here in Tampa) in 3d. I did them all with 3d-faces or extruded lines, but it's a round building, like 700+ft dia & like 5 stories tall & I had to draw the elevations & wasn't about to do orthographic projection for all of that, so I just learned 3dAcad . . . I used LT for most of it, but some stuff I had to use another guy's computer in the office I was in at the time to draw/build the stuff that was beyond the LT capability.

It's come a long way since then.

:cheers:

[NEATman]

Impressive Buile mcArch!

You had mentioned in an earlier post that you were looking for a better coupling.

Take a look at these manufacturers:
http://www.ruland.com/products_services.asp
http://heli-cal.com/cm/Products/Flex...ings/Home.html

I've used both brands before on commercial products, and they are very good quality. In your situation, you should stay away from Bellows style, as they tend to be epoxied at either end.

Keep up the great work!

Keith

[mcArch]

Thanks Keith, I'm going to go with a clamp style, double disc coupling from Rick at HomeshopCNC.

[mcArch]

Moving right along . . . but I think I've got an issue with the A * C axes & torque/machining force.

With 8" to the collet from the A axis pivot (same max. for C axis), plus a max. of 4" tool stick-out (say a long 1/2" end mill) that gives me a lever arm of 12" right? So the way it is now, with a 640oz-in stepper geared 4:1 (60:15), I get (640)*60/15/12/16 = only 10.66# of available force at the tool tip to machine with? I think I did this right.?

This is why Bartuss used a 50Nm stepper (~7000oz-in) for the A & C on his big 5 axis build, I see. So, I need a much larger motor & a different gearing solution. Only thing is I'm limited on space for the motor, only about 3.75"max. of motor height with a NEMA34. Homeshop has an RS1290 that's 3.75" height

So if I re-do this using a worm gear with 20:1 & use the RS1290 (1257 max avail with 7.8amps max from gecko203V wide open) I can get 104# avail for machining force. This may work ok for the 5.3hp spindle.? The other issue is that it will take like 3 sec. to turn the axis 1 rev. At 90deg, this works out to . . . wow, 188ipm on a 3"dia circle. That's not bad actually, the smaller the circle, the less the ipm though.

That works for the A axis, but the C I cannot get a worm gear on the axis very easily (the bracing on the Z carriage is sort-of in the way. If I go with the RS2550 (2486 max avail) and change to 72:14 (5.14:1) 5mmHTD, I can only get 53.3# avail max. when the A axis is at 90deg. This will not be good for the 5.3hp spindle, right?? I suppose I need to find a way to get that 20:1 gearing on the C, or maybe 10:1 to keep my step pulse lower.

What to do . . . what to do . . .

[jsheerin]

In designing my 5 axis router (not built yet, so take with a grain of salt), I was looking at 6N*m continuous torque servos for my A and C axes with 100:1 gear boxes. This would give me around 300lbf continuous cutting force at full acceleration with a 12" pivot distance to tool tip, just fyi. I was targeting 4.5 seconds for a full rev but accelerating from zero to that speed in 1 degree of rotation. That would give me a speed at the tip of 1500ipm (if it was just pivoting). Also remember your torque will fall off as speed increases with a stepper.