Designing a 4 x 4 steel frame router Inputs needed

[marchudson]

Recently started the initial design process for a new 48" x 48" router. This will be mostly used for hardwoods with one or two aluminum cuts thrown in here and there.
My initial plans are to use Hiwin 20mm rails for the x and y axis and 15mm rails for the z axis. For drive mechanisms, I'm thinking R&P using cncrouterparts reduction drives for the X and Y.
the Z will be a 5mm ball screw of some sort. For a spindle I plan on using a 2.2k air cooled VFD setup. Motors will be either NEMA 23 or 34 steppers.

So here goes the first question. I'm torn between using 2x2x0.25 or 2x4x0.25 to construct the gantry.
The 2x2 pictured below weighs in at 117lbs and the 2x4 is 163lbs and I haven't added the left and right uprights with steppers yet.
Should I stick with the 2x2 or 2x4 design?

I'm not sure if the 2x2 design will be stiff enough but I'm confident that it will be light enough for the steppers to provide adequate acceleration.

If I go with the 2x4 design, will the steppers (23 or 34) be able to provide reasonable acceleration?

Thanks for your time.




Similar Threads:

• cnc router frame, steel or alu?
• My Steel Frame Router
• New Machine Build- My steel frame router
• New Machine Build- Steel Frame 2' x 4' Router
• My 2nd Steel Frame Router

[wizard]

When designing your machine you need to take into account the distance your gantry will span. if you are looking at a 48" cutting area that most likely means an unsupported span of more than 50". I honestly don't believe 2 x anything would be good enough here. The is a general statement off the top of my head, the reality is you don't have enough information here for much more to be said. A Key parameter is the Z axis height as the Z axis can apply considerable torque to the beam. Keeping the same general layout of your beam I wouldn't consider anything less than 3" square tubing and that would probably be a bit weak.

That being said, all things considered designing for a single square beam is far easier. Look into the stickies threads where this gets discussed at length but a single square tube beam of suitable size offers a lot of bang for the buck.

As for motors, don't even think about them until you have a mechanical design that is satisfactory for you. Purchase steppers suitable for the final mechanical design, don't go the other way around because that will leave you with a mechanical design that doesn't meet your needs. So this means you need to define what you needs are, especially when it comes to non wood based materials. Once you leave the world of wood stiffness becomes far more important in a routers design.

[pippin88]

I used a 200x200mm (8x8") tube, 1600mm long for my 4' gantry.

Works well and is easy

[marchudson]

Thanks for the input guys. Looks like I was way off. Back to the drawing board.

[wizard]

Thanks for the input guys. Looks like I was way off. Back to the drawing board.

Work with an idea of what you want for Z axis clearance and how much you want to do in aluminum. To work in aluminum with the same sort of machining you might do in wood, you need a machine at least 10 times stiffer than what would be workable for wood. How well a machine will do aluminum is partly related to what your expectations are. There is a big difference say between engraving and running a 3/4" end mill 3/4" deep.

[marchudson]

Redesign number 1
Looked through a lot a builds on here and really liked the one by Fluxion. So I'll post it here to see what everyone thinks.
Base is made from 2x2x0.25 3x2x0.25 and 4x6x0.25
Gantry is 8x8x0.25

Any mistakes or weak spots that you might see, please let me know. Thanks for looking

[louieatienza]

I'm not the biggest fan of the Z axis rail configuration. It should be apparent that the spindle is mounted on a large overhand, i.e. lever. Also the blearing block have to mounted tighter together as you have to have any appreciable travel; check the width of the bearing spacing in the carriage, which negates the wider spacing on the gantry. I'd like to see a longer carriage plate with the rails mounted to the plate, and blocks fixed on the Y plate. This would allow to move the lower bearing block as far down as possible, giving the best stiffness.

[pippin88]

Gantry stiffness will be increased by end plates for little extra work

[dmalicky]

Very good suggestions above. The last pics posted show a machine that's on its way to being a really good aluminum cutting machine, but it has some notable weaknesses (mostly the Y and Z car). For only a wood-cutting machine, many parts of it are overkill. Since it is fundamentally a great starting point for aluminum cutting, my comments below are aimed at making it into that.

The base is mostly very stout. Actually the main table cross-sections are much larger than needed for stiffness, if you want to save $, but weight is good in a base. The weakest part of the table is the bottom of the legs -- the diagonal braces end needlessly far from the floor -- drop that lowest perimeter frame as low as possible. The diagonal braces and vertical center support are currently welded to the middle of the 4x6 tubes: those welds may distort the tube, and a 4x6 doesn't need those middle supports (even a 2x4 steel tube is enough for a 4' span). Skip all those and just use 1 diagonal tube from corner to corner of each face: this will also make the frame a complete rock.

For the gantry, yes, bulkheads or diagonal bracing dramatically increase the stiffness. But if it is a steel 8x8x.25, for just wood and lighter alum, it's enough as is (100k lb/in for the gantry contribution w/o any bracing). Keep in mind as a steel gantry, it will be very heavy and need big motors and drives, or be slow {in acceleration}. The main gantry weakness is the feet bearing spacing: for this gantry height and weight, it will want to tip under X-dir accel/decel. So space those bearings further apart, to the ends of the feet.

For the Y-axis rails, move them as far apart as possible on the 8x8 tube (up and down, in the side view pic).

The main weaknesses in the whole machine are the Y-car and Z-car: these will be much more flexy than everything else (as is). As Louie mentioned, it's better to flip the Z rails/blocks, to get the lower Z blocks as low as possible and keep them there {EDIT: but there are tradeoffs with a lot of Z travel}. The Z blocks are also too close together in the vertical direction-- get longer Z rails so the Z blocks can be vertically spaced at least 6" apart. Also make them 20mm like the others (15mm has tiny bolts). Last, space the Z rails wider apart (front view). A square pattern is generally simple safe advice.

The single plate Y-car and Z-car will bend a lot due to cutting loads in the X direction. A "C" section for each is much much better (as seen in the top view). The best way I know is to nest the C sections, like the attached pic.

For high stiffness, the spindle has to be rigidly coupled to the Z-car. The attached pic shows that for a round spindle. For a spindle that only mounts on its back surface, get the 'legs' of the C-section pretty close to the spindle, and/or use a thick Z-plate.

With only 4' of travel, ballscrews are much better than R&P: stiffer, more accurate, more precise, less wear, and similar cost. The R&P mounted above the gantry is really the worst place for it for stiffness -- too far from the cutting forces (Y-direction). A ballscrew fits nicely in between the Y-rails (as low as possible is best). Screws with 10mm lead should be fast enough {10mm is a good choice for high acceleration in detail work; for big contour movement in wood, 20-25mm lead may be better}.

Last item, which is harder to appreciate, is joint quality. All the structural stiffness in the world won't matter if any 1 joint is not very close fitting and tight. And there are a lot of joints. I have learned this the hard way. Flex always exploits the weakest link. E.g., a tiny burr that separates a joint can ruin it. That will require careful machining of all mounting surfaces, which is needed anyway for a stiff machine using profile rail (or it will bind). Conversely, if the build process won't have that level of machining, the benefits of those beautiful cross-sections and triangulation won't be realized.

With all those changes, this should be a great machine.

[dmalicky]

Here are most of those suggestions, and a few more, in pics:

[pippin88]

Great post dmalicky.

[marchudson]

dmalicky. WOW what an awesome reply. Thanks for taking the time to look over the design. I'm going to look over it and see what I can incorporate. I love the idea of adding c channels to the spindle carrier. Never thought of that. Thanks Marc

[wizard]

A nasty case of the flu has me just getting back to this thread! As usual dmalicky has a great series of comments, there is only one that I have an issues with.

This involves where to put the lower rail for the frame legs. If you put it to low to the ground you have a potential safety issue which has nothing to do with the structural correctness. What one can do is jam your feet under the rail causing personal injury. Don't ask how I know this. If you are wearing steel toes you even run the risk of getting caught under the rail.

Admittedly I have big feet but I only point this out to help you avoid unpleasant experiences. I'd want to see at least 6" clearance between the ground and the lower rail. That should leave enough clearance to protect even an MBA players foot.

[dmalicky]

Glad it's helpful. Apologies to Fluxion if my suggestions seem too critical. Actually, I was motivated to improve it because the original design does so many fundamental things right and the mods are straightforward.

Good point, wizard, as usual. The voice of experience. Hope you feel better soon. Yeah, a stiff structure is not worth an injury. 6" clearance would be no problem if the legs were bigger, maybe 4x4 or so.

I'm working on some FEA runs including the Y and Z cars, to go with the gantry work from a few years ago (wow, time flies) -- that should give some data on sizing of the C cross-sections.

[marchudson]

David
Thanks for the help. I'm going to try and tackle one thing at a time with your help. I made a quick drawing of just the Z axis carriage with fixed blocks instead of the fixed rails that I had
drawn before. I will add the "C" section bracing that you mentioned but have left it out of this drawing just for clarity.

I didn't want to go to far with the changes in case I'm way off with the basics of how to keep the Z portion of this build stiff.

[dmalicky]

That's smart to start with the Z axis. (Keep in mind that iteration will be needed later, as other components are added.) I'm assuming this is the Y plate with the Z blocks and rails on one side, and the Y blocks on the other. Could you include a scale or size? I will try to help along the way, but can't promise... I've been on vacation so have had more time lately.

Suggestions:
- The rails are placed nice and wide, and the blocks are in a ~square -- that's good. (Depending on how the spindle fits on the Z plate, the rails might be a little too wide... don't know yet.) If you want more stiffness for cutting aluminum, increase the vertical spacing of the Z blocks an inch or 2. I usually say "square" as first advice to keep it simple -- more accurately, the relationship of aspect-ratio and stiffness is non-linear (for Y-loads at the cutter): for Z-motion, a little taller than square is much stiffer. For Y-motion, a little wider than square is much stiffer. But for wood it won't matter.
- IIRC the spacers under the blocks are not needed. For ball nut clearance, mount the ball nut flange just above the top of the plate (with a small relief cut) -- a 1605 assembly should fit. Assuming 20mm rails/blocks. Most blocks bolt from outside-in, so Y and Z blocks can overlap and mount to the same plate (and that is ideal for stiffness). Are the Y blocks already placed to fit the 8x8 gantry rails? If so, you can probably shrink the overall square of these Z blocks.

[marchudson]

David

Yes this is the Y plate with Z blocks and rails on one side, sorry for not labeling them in the beginning . I have the Y axis blocks spaced as far as possible on the 8 x 8 square tube gantry. Measuring the distance between them,
I spaced the Z axis block the same distance in order to form a square.

Didn't realize that I didn't need the spacers under the Z axis blocks. I just assumed that I did in order to facilitate mounting the block and also to have room for the ball screw.
Here is a close up of the Y and Z blocks mounted to the same plate and overlapping without a spacer. The spacing looks tight but doable. Has anyone done this before?
Thanks
Marc

[dmalicky]

Yes, that equal size squares strategy is a good one. Nice work on the hole layouts -- it looks great to me. Having the blocks back-to-back is ideal for stiffness: loads go straight through, no bending in the plate. As you add more components, you may need to shift the arrangement some, but it will be stiff as long as the blocks are close to each other.

[louieatienza]

David

Yes this is the Y plate with Z blocks and rails on one side, sorry for not labeling them in the beginning . I have the Y axis blocks spaced as far as possible on the 8 x 8 square tube gantry. Measuring the distance between them,
I spaced the Z axis block the same distance in order to form a square.

Didn't realize that I didn't need the spacers under the Z axis blocks. I just assumed that I did in order to facilitate mounting the block and also to have room for the ball screw.
Here is a close up of the Y and Z blocks mounted to the same plate and overlapping without a spacer. The spacing looks tight but doable. Has anyone done this before?
Thanks
Marc

Probably most column mills do so on the X-Y saddle. A little easier to do if you have flange-style bearing blocks....

[eat5hams]

Great thread, I'm also looking to emulate this design. Any updates on how it went?