Hello, all! I am starting to build a 32" x 60" CNC router table. I may be going about this in an unconventional way, because I am plunging in with construction without having every last detail engineered out to the Nth degree. Many of the "build logs" here could be perhaps called "completion logs", in that they show a finished result in their first entry. Not this one! That said, I do have a few givens and druthers established.

1) I am using an existing laminate-clad tabletop as my starting point. I got this massive panel for a song at Construction Junction (http://www.constructionjunction.org), a local building-materials recycling center. I've seen a ton of table designs, and even some "finished" tables, that had no work surface! Yes, I can understand that a CNC router table's work surface may suffer a crash now and again, but I'm willing to live with that for this go-round. This is actually my second project start. I got a good start on a Seventh Sojurn, (John Kleinbauer design,) but I concluded that it was going to be insufficiently rigid for my work. Plus it needs a gazillion parts! Anybody want to buy a "kit"?

2) The table will have a moving wrap-around gantry, using the 1530 series of aluminum extrusions by made by 80/20 Inc (http://www.8020.net). I am NOT going to build the whole table with this stuff, due to the expense involved. Only the gantry and the supports for the axis drive screws will be 80/20. Even that material will be obtained via "80/20 Garage Sale", their eBay outlet for overstocks and out-of-spec materials. I do have access to a metal-cutting bandsaw at work, which will allow me to gang-cut a pair of 33" sections for the gantry, hopefully ensuring uniform length where that is important.

3) I will be using the linear bearings offered by Ahren Johnson at CNCRouterParts (http://www.CNCRouterParts.com) . These things are dynamite! Ahren's system is based on using 1/4" thick metal strip stock as the rail. I'm using 1018 cold rolled steel, a pair of 3" wide strips for the X axis rails. Ahren's system needs a minimum 4" wide strip for the single rail used on the Y and Z axis, I may go wider here. Ahren's carriages and motor mounts should GREATLY simplify the build, it looks like an EXCELLENT value in the old "time versus money" tradeoff.

4) Stepper motors and Gecko G201 drives are already in hand. Control software will be the open-source EMC.

Here's a sketch of the way the tabletop, a 2x4 spacer, and the X axis rails are going together. Having the carriages "underslung" this way seems likely to keep most of the sawdust away, we will see. I have the front X rail already in place. When I drill the mounting holes for the rear X rail, I intend to elongate the holes so that I can adjust that rail for proper distance between. (In other words, I can "preload" that back rail once the gantry is assembled.)

MacMarty,
Thanks for the nice workds about the carriages -- I think they will work well for you. Starting with a pre-made table top might be a good idea, and will help you have a feel for how rigid the machine is as you build it. Good luck -- I'll be watching to see how it goes!

Ahren
www.cncrouterparts.com

Ah, it's the weekend, and time for more work on the Green Machine. NOT! Too sunny, and too much honey-do. Manual labor gives me time to think, and today it's about steppers and Acme-threaded rod.

Like probably everyone, I wish I could start out with powerful steppers and leadscrews. The reality is that I am building this thing on the cheap-as-appropriate mode. I bought Gecko drivers because they seem to be the gold standard. When this started out as a Keinbauer project, I bought three IH234012 steppers from CAMtronics, planning to run them in 8 wire parallel for maximum speed. These should give me 157 oz-inches of torque, not large when compared to many projects I see out there.

So here's the dilemma de jure: should I stick with these (paid-for) steppers, and set up with single-start acme rod because of the relatively low power? This will get me up and running, get some cash flow coming in, and then I can upgrade to more powerful motors and multi-start acme rod for X and Y axes down the road. Yes, I'll be buying stuff twice, which goes against my nature, but bootstrapping is how we are getting this done at all!

I'd sell the motors you have on Ebay and buy some 282 oz motors from http://www.kelinginc.net/SMotorstock.html At $40 each, you might not have to spend too much more than you can get. That's the cost of not doing enough research. :)

And then buy the multistart acme screws now, rather than buy again later. It'll be much cheaper in the long run and not too much more right now.

Although, since you have Geckos, even more powerful motors may be a better choice. This is really a hard question to answer, as realy only you can decide. I'm thinking that you may be able to use the multistart screws with the motors you have, but I'm not sure how well they would work right now.

Marty,
I agree with Ger -- go ahead and purchase larger motors and higher lead screws now. I have 300 oz-in motors on my machine, and sometimes wish I had something a little larger. With the Geckos, you've already spent most of what you need for a higher performance system. Also, the higher lead screws will let you get more speed with very little hit in force due to their higher efficiency (and the lower speed you can run your steppers at). Since these drive the type of nuts you need, I would say the screws are an even higher priority than the motors, especially since motors are easy to swap later.

Ahren
www.cncrouterparts.com

(Thanks for the advice on steppers, everyone. Gonna get some bigger ones.)

AND, at long last I have some progress, with photos to share. I've built up both of the x-axis rails, and the gantry to boot. Here's what I -think- I've learned. Note that in all three photos, the table is lying on its back.

The gantry is 80-20 1530 series, with triangular gussets at all four corners. Strong like bull! The first time I put it together, I put the gussets on the inside of all four corners of the gantry. Then I realized that this would leave the bottom horizontal member pretty far away from the table, requiring a thick spacer block to get the drive screw into position. Came back to it next day and realized that things would work out a lot better if I flipped that bottom member over, so that the gussets are on the outside of the rectangle. ("Bottom gussets outside")

You'll note that I've constructed my gantry so that the vertical rails mount to the end of the horizontal rails. My intention here was to design a solid rectangle with 90 degree corners. With that as a given, the next challenge was how to arrange for adjusting the fit between the rails and gantry. Gotta be perfectly snug, right?

The current solution was to mount the front X rail with 1/4" holes and 1/4" lag screws, nice and parallel to the edge of the prefab table, and screwed down tight. The back X rail uses the same lag screws, but I've oversized the holes so that the rail can be moved in the horizontal plane. The third photo shows a slack adjuster, (there's one at each end of the table,) which allows me to carefully push the rail out, to take out all slack. There's no room for the gantry to "yaw" as it rolls to and fro. I'm planning to upgrade to Ahren's 8" bearing blocks, to make this relationship even more rigid in the yaw axis.

With those in place, now I'm moving on to the Acme rods and bearing blocks. For some reason, the bearings don't slide smoothly along the rod: the bearing starts out smoothly enough, but gets really tight on the X screw, and then binds up at one point. Something's out of spec somewhere, where'd I put that micrometer? I have 2-start 1/2" Acme screw, and the VXB 1/2" radial bearings Ahren recommends. Any there any curative measures known to everyone but me?

MacMarty,
Great to see progress on your machine! The adjustment system you have for the rails is an excellent idea.
I'd like to help with your binding problem, but I'm not sure I understand exactly what the problem is you are seeing. Does the system slide smoothly without the ACME rod installed? Are you just seeing the screw bind? If so, this is most likely from a small misalignment that requries the screw to "flex" in the middle of the travel. Even 1/16" off can make everything bind, so you might need to readjust or even shim the 2x4's you have under your rails.
Or maybe you're just having trouble fitting the radial bearings on the ACME rod? Let me know -- I'd like to help.

Ahren
www.cncrouterparts.com

Hi, Ahren -

At the moment all I have is the radial bearing in the bearing block, nothing else, and the bearing hangs up after only six inches travel down the rod. It's very clear that there are fine tolerances at work here, no room for twist or slop. My ACME rod is black oxide finish, and perhaps there's a rough spot. I'm tempted to try a few gentle passes with some ultra-fine emery paper to take off any invisible nicks or bumps on the outer diameter. (McMaster shipped my two rods in a tube with no padding, maybe there's a ding or nick I can't see.)

I guess the main question is, why do you need to slide the bearings on further than 6"? Aren't the support blocks at either end of the rod?

That being said, if they just won't fit, there's nothing wrong with a little sand paper to knock them down a couple of 0.0001"s. Which is probably all it will take. However, I would reconsider your order of assembly before sanding the entire rod, unless there's some critical alignment that you've already set and are trying to maintain. Either way, good luck!

Ahren
www.cncrouterparts.com

Hm. Maybe I'm having a really dense day, too much time out in the sun this afternoon. Or else I'm missing something -really- fundamental. Let me wallow in ignorance for a moment. (Somebody will thank me someday. I hope.) Sorry in advance for the verbosity here also. (I'm an ex-project manager, and hence incurable.)

OK, the motor mount (and motor) is bolted on at one end of the "drive-line extrusion", yes? The shaft coupler at the motor end supports that end of the ACME rod. At the extreme other end of the ACME rod is another bearing, presumably supported in a block attached to the far end of the extrusion, and that bearing's really just there to support the far end of the spinning rod. Otherwise the far end of the ACME rod would flail around when the gantry is near the motor's end of the system.

Drive line extrusion is attached to the underside of the table. If we make the motor spin, all we have (so far) is a spinning threaded shaft. No attachment to gantry, and therefore no gantry motion.

Your bearing block, the VXB radial bearing, and bearing block cover are a three-part sandwich that is to be attached to the bottom beam of the gantry, right? The "thread follower" components (Dumpster nut and other associated bits) are on either side of the bearing block "sandwich", and something in that lot engages the ACME threads, and thereby serves to push against the bearing block sandwich. That complete assembly moves the gantry back and forth along the ACME rod when said rod rotates. Right?

My problem (perhaps poorly stated earlier) is that the VXB radial bearing in the center of the gantry bearing block "sandwich" will only slide 6 inches down the ACME rod before it jams onto the rod. At this instant I -only- have the 3-part bearing block sandwich installed: no Dumpster nut, which I thought was the part with threads that engage the ACME rod. (None of the thrust washers etc either.) In this condition, I would expect to be able to slide the gantry along its full travel, with the inner circumference of the VXB bearing sliding gracefully and precisely along the ACME rod. It's going to do that in normal operation anyway, yes?

Maybe if I had all of the driveline parts in hand I'd know what it is that I am overlooking...

Ah, now I see where the confusion is coming from. The bearing blocks do NOT move -- only the nut (well, the nut and whatever you use to attach the nut to your gantry, which is not one of my bearing blocks).

You need two bearing blocks per axis. One is just a radial bearing, typically mounted at the end opposite the motor, which only constrains the screw radially. The other is the "stack" of radial and thrust bearings -- this keeps the screw from moving radially and axially -- remember, there is an equal and opposite force on the screw to whatever thrust you are putting into the gantry. This stacked up bearing block absorbs the thrust load -- otherwise it would all be on the motor bearings, which is definitely not a good idea.

Hope that helps,

Ahren
www.cncrouterparts.com

OK, everyone all together now, "Ahhhhhhh! Now I understand!"

This is what I get from starting my journey with a Kleinbauer Seventh Sojurn. In his design, the reaction thrust is on the rod is taken care of by a skate bearing on the far (non-motor) end of the rod. His "nut" is just a piece of aluminum angle, tapped 1/4-20, and bolted to the bottom of the gantry equivalent, and his motor coupling is just a length of rubber tubing. OK for milling model airplane ribs, but not for what I'm now planning.

We needz a better pikshur, as my LOLcat would say. Ahren, I'm sorry to say that the nice CAD drawing on your Components List page has enough "slack" to lead me astray, but we've recovered just in time! I'll soon sketch one up, post it for your confirmation, and proceed with a clear mind. See, isn't it good to have a noob wander by and clarify assumptions? :D

Glad to see things are clarified, and sorry for any confusion the diagram caused. I may replace the drawing on that page soon. I'm hoping to release plans for a complete machine using my parts soon, so that might be the time to do it.

Anyway, hopefully this doesn't put you too far off track -- I think with your current design, you can probably just mount the other bearing block in a fixed position, and use your aluminum angle trick with the dumpster AB nut.

Good luck!

Ahren
www.cncrouterparts.com

Ahren

ahren,
I'd be interested in your plans when you get them finished. Any sneak peaks?

I've started to get some parts and I'm thinking about using your bearings.

Sieg,

Thanks for your interest. To keep from hijacking this thread, I'll post a sneak peak over in my product announcement thread here:

http://www.cnczone.com/forums/showthread.php?p=463438#post463438

Best regards,

Ahren
www.cncrouterparts.com

By now I'm sure you all know that GlacialWander's build (http://www.glacialwanderer.com/hobbyrobotics/?p=17) has been featured on www.HackaDay.com.

I posted a related question over there, and I'd like to pick the collective brain here also. It relates to the way the Z axis is mounted to the Y framework. (GlacialWanderer took some lumps because the Z axis beam is static: only his router moves up and down. Others pointed out that the axis frame -AND= router should go up and down.)

Here's my puzzle de jure. What about building the Z axis as an open-center box beam, wrapped around the top member of the Y axis? Seems to be that this could allow one to mount the router (or a belt-driven spindle) on the centerline under the Y beam, and thus reduce tipping forces when driving the Z down into the work. (I'm not an engineer, I sure there are better ways to express this.)

Thoughts?

Marty,
The biggest problem with Z axes is the cantilever on the Z, which is inherent to any design that doesn't raise the whole gantry (and while this is possible, it limits the workpiece height to the length of the shortest tool used). The lever arm from the bearings to the tool is typically the limiter in terms of stiffness. I'm not sure I'm envisioning your box beam idea right, but if all you do is move the Z closer to the gantry (or even to the gantry centerline), you haven't changed things too much in terms of lever arm -- you've just eliminated the short side of the triangle.

I think the lumps taken by Glacial were more to the point that his Z was hanging below the gantry, and that there was no cantilever on his Z. This means he has a taller gantry than he really needs, since he can't reach the worktable, and he isn't really even benefitting from the high Z since his axis is hanging down into the work area. I don't know that anyone was really suggesting that he move the whole gantry up and down.

I think they were suggesting that he mount the Z axis rollers fixed at the height of the gantry, and have the rail move up and down between them. This approach would take some re-engineering, but might be possible. I originally stated that you would need to rotate the nut to do this, but that might not be true. I'll have to think about this...

Ahren
www.cncrouterparts.com

The lever arm from the bearings to the tool is typically the limiter in terms of stiffness. I'm not sure I'm envisioning your box beam idea right, but if all you do is move the Z closer to the gantry (or even to the gantry centerline), you haven't changed things too much in terms of lever arm -- you've just eliminated the short side of the triangle.


Well, let me try to explain the idea better. (I may need to break out the graph paper again. :)) To me it's a given that the Z assembly (rail and 8020 beam) will move vertically. I'm thinking to do one of two things:

1) I have a tall gantry, I could add a second beam-and-rail across the middle of the gantry rectangle, and have the lower Y bearing ride the underside of that rail. In effect, this is to create a really WIDE Y-rail, and would in itself add tons of stiffness to counteract bending moments in Z.

2) By arranging things so that the centerline of the bit falls near the Y-plane centerline of the gantry, I end up with the line of thrust being downward in the plane of the gantry. An even amount and direction of thrust on all of the gantry-foot (x axis) carriages. No bending moment in either direction, (which I visualize as the router pitching up and down,) just vertical thrust. I think I get some of this beneficial effect as long as the bit falls within the rectangle described by the four bearing points of the X carriages. In Glacial's case, when the Z presses down into the work, (and he's indicates that his is an engraver operating on the face of the X plane, not a router where some cuts will be edge-face cuts,) there will be upward force on the right-side bearings of the x carriages, and downward force on the left-side bearings of the x carriages. Right? (Or is my head just full of sawdust and swarf?)

Marty,

1) Yes, a wider gantry will help with stiffness, but only in terms of the torsion of the gantry itself. There still has to be a part that hangs below the gantry that is cantilevered, and this will still be under a bending load from cutting forces. If you decide to move the "Z-assembly" you'll only see a stiffness benefit from a wider gantry if you spread the bearings further apart (perhaps by using an extended carriage).

2) "Vertical" force is only evident during plunges into material. It can be substantial, but is typically not what causes racking or deflection. Even an engraver will have cutting forces that push on the side of the cutter, and these are what can cause severe deflection if you have a long Z acting as a large lever arm (the distance from the center of the carriage to the tip of the tool).

In Glacial's case, it looks like this is not currently the case, but when he changes things around so that the Z plate hangs down and the router is cantilevered off of it, this distance will probably be longer than the distance the router is sticking out from the gantry. Still though, shorter levers are better if you can make them work for your design.

Ahren
www.cncrouterparts.com