Ok, I'm starting to CAD a design for a woodworking (and maybe light aluminum) router. Work envelope is going to be around 48" x 36" x 7".

When I build stuff, I tend to over-engineer things. So, with this project, I want to make sure I don't do that too much, and spend a ton of money I didn't have to.

So, I looked up specs on THK linear rails/carriages. Even the smallest (15mm rails, 34mm carriages) have a vertical load capacity over 1,000 lbs. Horizontal capacity is over 580 lbs. I'd be using two carriages per rail, so even more distributed load capacity, right?

But the problem is...I *know* that these small rails are going to "look" like they are not strong enough. Should I just trust the specs and go for the cheaper, smaller rails? I mean, there is no way my router is going to see forces anywhere near the rated spec on these rails/carriages.

Same question about aluminum plate. My goal is to build the entire gantry (and z-axis) out of 0.500" stock - and try to use one screw size as much as possible for assembly. I've seen pictures of some of the routers folks here have built, and man some of them have aluminum parts that are 2" or 3" thick! I know some people get materials cheap or free as scrap, so maybe that's the reason.

On some parts (like the stanchion that connects the ball screw to the z-axis plate, and the z-axis sliding plate itself) I know I need thicker than 0.500" aluminum -- but am I okay using it on the rest of my design, as long as I pay attention to span distances? I can't imagine that there'd be much flex in a 0.500" plate of aluminum across a 12" span, right?

Thanks for any guidance you can give me about this topic.

-Chris

Chris,

You are quite right about the rails. When I was designing my machine I brought in the THK rep and he said that everyone over-estimates their needs by some considerable margin. Because of cost restraints he recommended 25mm GSR rails for my X axis which is 2200mm long. GSRs are like half a normal rail and therefore take less load. I am sure the 15mm rails will be more than enough for your needs.

Can't help on the aluminium - sorry.

Mike

But I can! (Help on the aluminium, that is :))

I just built this. (It's meant to be the start of a drill jig that hopefully will allow me to drill straight and parallel rows of holes for my CNC machine!)

The aluminium bar is 10x80 mm^2 (2/5" thick). It is 300 mm long (about 12"). If I press down on the end it flexes several mm! Most of it, perhaps 2/3, is because of the fibre board that the THK rail is mounted to, but it is not difficult to actually see the aluminium plate flex. I was really suprised about this since the plate feels so d*rn unflexible when not mounted ;)

Build with triangles; don't rely on the cross section of the material to give you the stability.

Arvid

Mike, thanks for the confirmation on the rails.

Hi Arvid,
Thanks for the real-world aluminum test :)

Just curious: I was thinking in terms of a 12" supported span of 0.500" aluminum - not one end secured and the other left free. So to be sure, if you lay your 12" piece of aluminum plate across two support members, and then press down in the middle of the plate, do you see any significant flex?

The design I'm thinking about uses an integrated aluminum "I-beam" as the span structure for the y-axis gantry. I threw a couple of triangles in, too, for luck :cool:

I'm attaching a couple of pics. In the second, you'll see the purple "I-beam". As you can see, the gantry is offset along the x-axis. I chose this design because I was striving to keep the z-axis vertical force centered within the footprint of the x-axis gantry. So, the z-axis is sort of "tucked in" to the gantry, and lined up nicely with that leading pair of x-axis linear carriages.

This also minimises the x-axis rail that is wasted due to the footprint of the gantry, at the same time allowing the rail carriages to be spread apart as much as possible.

Do you see anything about this box that says I can't get away with using 0.50" (13mm) 6061 al? The box is 51" (130cm) long, and 7" (18cm) tall.

Nice looking design.
What kind of belt are you planning on using to drive the two screws in your X axis?
I am considering this approach but am concerned about belts stretching over time and introducing position errors.
You might want to look at : http://www.scandinvent.com/CloserLook.html and the way they protect their gantry drive mechanism. Could be useful.
Keep the pics coming...looks first class.

You shouldn't have any problems with belts stretching. Go to http://www.sdp-si.com for all the information on belts you should need.

Not to sound jealous or anything, but you must own your own bank or something. I like you design, looks very much like a scaled down commercial rig. IMO 0.50 aly plate should be fine. I will be very interested in watching this thread, and build unfold.

Woodsnarfer, nice design, V nice cad. What software did you use?

Chris,
BeamBoy is a quick and simple freeware program that will calculate the deflection of beams supported in various ways. I just used it to calculate the deflection of an aluminum bar 1/2" thick, 2" wide, and 12" long. With simple supports at each end, and a 10 lb load in the middle, the bar will deflect .00173 inches. You can download the program here. http://www.geocities.com/richgetze/

It too me longer a lot longer to compose this reply, than it took to set up BeamBoy and perform the calculations. It defaults to using steel as the material, but for aluminum, all you have to do is change the Modulus of Elasticity under "Beam Properties" from 30 million psi to 10 million psi. If you try it and have problems figuring it out, let me know and I will walk you thru a quick example and you will be off and running. It sure beats guessing about material sizes.

Jeff

Hey, thanks everyone for the comments and pointers to useful websites. I traced through this design last night, imagining every torque on the thing, and I'm convinced that I'm okay with 0.50" aluminum. I'm not an engineer, nor a machinist, so I'm still not 100% comfortable...but close enough. If someone points out "hey, you've got a problem right there", that would be helpful, but if not I'll stick with what I have.

Some answers to questions:

1. I used Rhino 3.0 and the Flamingo 1.1 raytracer to do the model. I just started using Rhino about a week ago. McNeel has a great set of online tutorials (I painstaking went through several of them), and a discussion forum where lots of really smart people answer your questions (sort of like here! :) ) So, I'm really glad I went with Rhino. I am flat out amazed at the power of this software. Go to www.rhino3d.com and check out the photo gallery...I was shaking my head in wonder looking at some of that stuff.

2. Don't hold your breath watching this thread...I tend to spend a lot of time designing & thinking before I start building. I'm really not going to be happy with this design until every component is fully specified, and every screw hole and screw is placed. I want to be able to go from CAD to a set of plans where I build the thing the way I want it the first time, with as little wasted material (and components) as possible. Before the design is finalized, I do need to purchase the rails, screws, motors, pulleys and belts so I can dial them exactly into Rhino. That's weeks of hunting for stuff alone.

3. I plan on having each motor mount have some adjustment for both installing the belts and tightening them (if they ever need it). I'll also guard the big dual-drive belt, and either guard or move the small ones inside the gantry. For the large belt I was thinking a guard made of clear lucite might be cool.

4. The budget for this is as close to $3000 as humanly possible. This is going to mean frequent visits to ebay and my local scrapyard.

-Chris

I am considering this approach but am concerned about belts stretching over time

A friend told me that I should only buy belts with a fibreglass or metal core. Fibreglass does not stretch.

Cheers

Andy

New; 15mm to 35mm Profile Rails and flange style carriages

Visit http://www.marchantdice.com/linear/PROFILE_RAILS/profile_rails.pdf

Mike, thanks for the confirmation on the rails.

Do you see anything about this box that says I can't get away with using 0.50" (13mm) 6061 al? The box is 51" (130cm) long, and 7" (18cm) tall.

Nice looking design. I have Rhino too. What a great program! I ran your aluminum I-beam through "Beam 2D" using 51" between supports with a 7" web and 7" flanges all .50" thick. E= 10e+6. Placing a vertical load of 100 lbs at the center of the span the deflection was .0003". There's probably more play in the rail bearings than that.

Chris

Strength is the least of your worries, I'd worry more about stiffness. If its stiff enough to be accurate, it is almost certainly going to be strong enough.
BeamBoy looks like the right idea, or find a list of formulae from somewhere and haul out that rusty ole' calculator.
To give you a few indicators: centrally deflected stiffness goes down with the cube of the span (ie 2x as wide, 8x as sloppy). It goes up with the square of material height (twice as high: 4 times as stiff), and proportional to material width (twice as wide, twice as stiff).
As someone here said, use triangles a lot, this will increase the effective height of the material.