Hey guys,

I'm getting ready to design and build my second cnc router.

The first one ended up being to weak and flimsy.

For those of you that are new, here is the first machine...

The 36" bars on the x axis are 3/4", not nearly strong enough. There is way too much deflection... The 24" bars on the y axis are 1/2". Definitely too wimpy. Most of the deflection comes from the Y axis.

http://www.mikenchell.com/images/cnc_no1.jpg

So here's the important part. I'm designing a new machine, and I really can't count on the first machine being too accurate. So I want to design a machine that can be built with the least amount of cnc parts, and still have a robust design.

I'll attach a photo of the one that I am starting to design now. Yea, don't mind those slides, I still have to chop them to the right length. Anyways, the silver looking things are 80/20 extruded aluminum. I am using 1x2" extrusions in the design. The red parts are the only parts that I need to cut with my first machine. Probably plywood or MDF. Oh that gold colored thing in the middle? I'll probably buy a z axis from a guy on ebay!!!

The green slides are igus w slides. Based on messages I saw in this forum, it was suggested that I use a single slide on each x axis rather than a double slide on each x axis. But I let my buddy at work look at my design, and he says it needs a double slide on each side. What do you guys think?

http://www.mikenchell.com/images/cnc_no2.jpg

Mike...

Couple of points;

The Igus site has a lot of stuff to calculate loading, spacing and alignment. I asked similar questions of Igus tech and got these answers amongst others;

Whatever rail you choose you will need much wider spacing on the bearings on each slide. These rails are not as forgiving as ball slides when it comes to racking miss-alignment.

Single rails should do the X fine from a purely load capacity perspective. Single rails mean you have to mount each bearing individually with the accuracy required.

Double rails would make the cost higher and require much more precise alignment. between rails. A benefit of a double rail each side is that the double rail comes with a mounting plate with the bearings already mounted on the carrage so taking care of some of the precision and probably giving a more ridgid gantry.

These (igus W) bearings have significantly different load capacities in different load directions. Might want to consider that, for example the Y axis might be better with the rails on different planes.

Andrew

Couple of points;

Whatever rail you choose you will need much wider spacing on the bearings on each slide. These rails are not as forgiving as ball slides when it comes to racking miss-alignment.


I'm showing 6" in that photo. What width did igus suggest? 8"? 10"? 12"?

Single rails should do the X fine from a purely load capacity perspective. Single rails mean you have to mount each bearing individually with the accuracy required.

Double rails would make the cost higher and require much more precise alignment. between rails. A benefit of a double rail each side is that the double rail comes with a mounting plate with the bearings already mounted on the carrage so taking care of some of the precision and probably giving a more ridgid gantry.

My buddy at work was suggesting two single rails on each side. One mounted on the lower extrusion as shown, and one mounted on the upper extrusion. I don't think I have the ability to do something that accurate on my wishy washy cnc router.

My buddy also thought it would rack in a sideways direction. In other words, if you stood in front of the x axis, and grabbed the top of the z axis and pulled it toward you... I don't see how, if the load capacity is there.

These (igus W) bearings have significantly different load capacities in different load directions. Might want to consider that, for example the Y axis might be better with the rails on different planes.

So for the y axis you're suggesting a w rail on the top part of the top extrusion, and on the front part of the lower extrusion? What would that gain me?

Mike...

If those orange parts shown for the y-axis uprights are MDF you definitely want to stiffen them by increasing the section depth to the outside. The 'old machine' suffers from the same problem.

If those orange parts shown for the y-axis uprights are MDF you definitely want to stiffen them by increasing the section depth to the outside. The 'old machine' suffers from the same problem.

Please explain "increase section depth" to me...

Thanks,

Mike...

What he's saying is if you push on the side of the gantry uprights, the 3/4" MDF will flex and the gantry will move from side to side. By at least doubling them up, it will be stiffer. I'm making mine out of 2 pieces of 5/8 MDF, with a 1/2" Hardwood frame in between them. I also epoxied a 3/4" x 3/4" Aluminum tube running from top to bottom, in a rabbit down the center. You want them to resist flexing in all directions.

Not sure how tall the uprights are, but I'd use 2 bearings on the bottom rail, about 12" from outside to outside. A single bearing will probably have binding or sticking issues.

I'm showing 6" in that photo. What width did igus suggest? 8"? 10"? 12"?

I suggest you look at the online tools Drylin Expert System. You can work out the dimensions online using the tools there. It varies based on several factors. Basically the distance between bearings on a rail is determined by the leverage (load) produced by the distance between two rails ( racking) and the distance between the bearings and the applied load ( cutting bit and weights and lead screw). The farther between rails or the farther between the cutting bit and any axis rail or the farther between rails and leadscrew increases the loads on the bearings and so increases the bearing spacing. The Igus bearings need to be spaced wider than ball slides because of the different load capacities but the same calculations should be doen for bal slides - they're oftn not because the capacity of the ballslide is usually so large in several directions. From your diagram I'd guess you'd need at least 8-9" on the X slide. More if possible. Depends on the size of router you're putting on there and where the leadscrew is going to go.


My buddy at work was suggesting two single rails on each side. One mounted on the lower extrusion as shown, and one mounted on the upper extrusion. I don't think I have the ability to do something that accurate on my wishy washy cnc router.

My buddy also thought it would rack in a sideways direction. In other words, if you stood in front of the x axis, and grabbed the top of the z axis and pulled it toward you... I don't see how, if the load capacity is there.

I agree with the idea behind the first suggestion because its spreading the load and the top rail would be closer in on plane to the cutting forces. Problem is it would require really close tolerances to install and avoid binding from misalignment. However two seperate rails are no needed since one rail can support the load.

For the second suggestion Ger has it nailed - you need to brace the gantry to avoid movement in twist. trapizoid and flex. Best way is to brace it in several planes. The Bearings are not structural and the bearings cannot be considered to carry any structural load.



So for the y axis you're suggesting a w rail on the top part of the top extrusion, and on the front part of the lower extrusion? What would that gain me?

Maybe nothing, but if you look at the loads applied while cutting at some point on the table the loads on all the axis are not simple or in just one direction. With bearings that are stronger in one direction than the other you might need to consider how you align the bearings. The Drylin W is roughly half as strong on one direction as the other. Aligning the Y bearings one horizontal and one vertical would mean that in the Y and Z directions the supported loads are the same at 1.5x the max capacity of one rail. If the rails are in the same plane then the load supported is 2x in one direction and 1x in the other.

Again the real reliable info is on the Igus web site an dthe design tools for Drylin W are very useful when planning this stuff.

Andrew

Have you thought of putting a little extra into the existing router to "beef" it up? For example pull the 3/4" x rails and use them on the Y, and putting larger rails on the X. Or adding bearings on the bottom of the gantry running against a supported drill rod, or a flat piece of crs. There are probably a thousand other methods. While its an after thought, it at least makes this machine much more workable.

Phil

What is the cost of these compared to the thomson's or a thk style linear guide? I imagine they are less, but not by much. What I'm getting at is maybe since you are investing heavily in this machine, maybe aluminum would be a better choice for your framing system. Two 80/20 1530 sections and 10 square feet of 1/2" precision plate would not cost much, probably less than 125.00, and eliminate your flexing problems entirely with the same rails and a good design. Then youd have a machine more capabl that would last forever and take much more abuse, or use. Just something to think about.

I think you are about to repeat some of your issues with the previous machine. So I'd rethink the whole formulation.

If you had stiffness problems before I don't see this frame helping alot. But then again I never understood the idea of placing guide rods under table. These should be close to the top of the table. Likewise I don't see alot of y axis support either.

Thanks
Dave

Your original machine looks well built and only has problems with strength due to the small y-axis rods. I built a JGRO and I used 1.25" seamless shaft (very rigid) with skate bearings on my y-axis. You could use the 0.75" rails presently on your y-axis to strengthen your x-axis. If you improve the original machine it could accurately complete many projects for you.

Jason

What he's saying is if you push on the side of the gantry uprights, the 3/4" MDF will flex and the gantry will move from side to side. By at least doubling them up, it will be stiffer. I'm making mine out of 2 pieces of 5/8 MDF, with a 1/2" Hardwood frame in between them. I also epoxied a 3/4" x 3/4" Aluminum tube running from top to bottom, in a rabbit down the center. You want them to resist flexing in all directions.



Stiffness increases with the cube of the depth and directly with the elastic modulous of the material. Adding a web is better than doubling up. Aluminum web is better than wood web for the same cross section. The joints between all layers of material must be well glued to each other. Otherwise shear forces will reduce the max stiffness possible with any given composite panel.

Dude that is a superb machine and it looks absolutely excellent, I still have yet to start mine as gaining info is critical in getting it right but you're basic design is what I would build EXACTLY!! Why not try and change the existing machine with what you feel needs to and when you know its correct then build youre other based on your improved design, oh and if you want you can send me that machine he,he,he,he. Kammo1

Stiffness increases with the cube of the depth and directly with the elastic modulous of the material. Adding a web is better than doubling up. Aluminum web is better than wood web for the same cross section. The joints between all layers of material must be well glued to each other. Otherwise shear forces will reduce the max stiffness possible with any given composite panel.


My aluminum tube is between the 2 pieces of MDF. :)

My aluminum tube is between the 2 pieces of MDF. :)

Double web wide flange I-beam.

your machine follows how mine was made alot i dont have that much flex the big difference i noticed was the width of our gantries mine is set wider at the base so the weight is spred further and causes less flex i do lose a little of my run from it just a thought on a cheap way out till ya can go bigger and better

Double web wide flange I-beam.


That's it, though it's not rectangular. Much wider at the bottom 12-13" and narrow at the top ~3-4"

Thanks for all the replies...

Here is my second try at a new machine...

http://www.mikenchell.com/images/cnctry2.jpg

You'll note I flipped the bottom extrusion down, and set the w slide on top of it.

I also used a laminated beam on the y axis instead of 2 extrusions. And I put one of the w rails on the top, and the other remained on the lower front.

I doubled the thickness of the gantry sides, and also made the base wider. The base is now 9" wide. I'm going to have to make the table a little longer, as I'd like to have 36" of travel in the x direction.

The z axis is going to be a machined aluminum axis that I buy from ebay, similar to this one...
http://cgi.ebay.com/Z-Axis-CNC-Router-Slide-H-D-6-5-trav-w-Stepper_W0QQitemZ7532231770QQcategoryZ42938QQrdZ1QQcmdZViewItem#ebayphotohosting

The machine now has a lot more wood in it, and it's much heavier too. But I don't figure I'll be moving it much once it's set up.

So did I miss anything?

Mike...

I'd consider making the uprights similar to the way I did, using thinner skins, or, drill some large holes to lighten them up a bit. I'd make the Y axis a torsion box. Again, it's lighter and stronger than 2 pieces of MDF. And make sure the bottom 80/20 rails are bolted to a sturdy table that won't allow them to flex.

Do you plan on driving the gantry from a single screw in the center under the table?

I'd consider making the uprights similar to the way I did, using thinner skins, or, drill some large holes to lighten them up a bit.


I can do that!


I'd make the Y axis a torsion box. Again, it's lighter and stronger than 2 pieces of MDF.


Actually, I was thinking about 3 pieces of plywood laminated together. If I took the inside layer and cut a bunch of holes in it, would that be the same thing as a torsion box?


Do you plan on driving the gantry from a single screw in the center under the table?

Yes, I was thinking about a 1/4-20 threaded rod, like on the first machine.

Mike...

Actually, I was thinking about 3 pieces of plywood laminated together. If I took the inside layer and cut a bunch of holes in it, would that be the same thing as a torsion box?
Mike...

Use something like a 1" thick center, and drill maybe 1-1/2" holes @ 2" OC. Then use maybe 1/2" skins. I'd make the uprights the same way. Try it on one of the smaller parts first, and see how it works.

I wouldn't waste time drilling holes in the inner layer. The only reason for the air space in a torsion box is weight and material savings, it would be slightly more rigid with a solid core. Cutting holes in the inner layer would not save material, and weight usually helps to reduce vibrations.

Mike, here's my 2 cents.

1. gantry stiffness. You would gain more stiffness in the gantry if you could increase the height of the Y axis mounting plate. I don't mean make it taller overall, but draw a side section of the gantry. Figure out where the bottom tip of your router bit will be when the Z axis is fully retracted. Extend your mounting plate down to this dimension. You might gain a couple of inches here and that will dramatically increase the strength of the gantry.

2. IGUS slides. I see nothing but trouble with the way that you are trying to use the slides. I agree 100% with your friend who said that you need 2 bearing blocks per slide. Another issue, I talked to IGUS about some slides, not these particular ones though. I've attached a picture of their recommendations.

If possible Mount the bottom Y rail up the other way. Best load capacity and least binding risk is with the load through the bearing with greatest contact area. The mounting flange directions are the direction of least capacity. Again the Igus site shows this.

Try and brace the joins in more than one plane. Increasing the x section does little is the joint is weak.

If you're going to spend that much on Igus rails then consider something more for the leadscrew. 1/4-20 is slow and too whippy for a leadscrew imho. Vibration of undue forces due to ship will cause other problems later, incuding increasing the probability of binding. Cheap Enco 1/2"-10 Acme is only 4.59 for 3'

I know this is repretitive but once you've got the dimensions and loads put them through the Igus tools and see what they say for spacing and binding before you spend the money.

Andrew

Okay, I flipped over the bottom rail, and added the 1/2"-10 Acme screw. Can you still use that cutting board stuff to make anti-backlash nuts at that size?

I put a VRML file here, if anyone wants to view it in Cortona.
http://www.mikenchell.com/cnc/2nd_try_cnc.wrl

Cortona VRML Client 4.2 is a free plugin for I.E. which you can get here:
http://www.parallelgraphics.com/products/cortona/

If you'd rather import the solid into your Cad/CAM package and play with it, or redesign it for me :banana: ,
the parasolid is here:
http://www.mikenchell.com/cnc/2nd_try_cnc.x_t

and the acis solid is here:
http://www.mikenchell.com/cnc/2nd_try_cnc.sat

Mike...

yeppers you sure can ..... and it works fine you just need to make a tap from your acme rod you get

Okay, I flipped over the bottom rail, and added the 1/2"-10 Acme screw. Can you still use that cutting board stuff to make anti-backlash nuts at that size?

If you're using an actual cutting board, it's probably too thin. Get a thicker piece of Delrin from Ebay or Mcmaster-Carr for $10

Mike, what I'd recommend when you start building this router, is to try to allow for things to be easily changed if they don't work to your expectations. Ideally, you don't want anything to be able to flex at all. If you find they do, make it as easy as possible to be able to upgrade the machine rather than have to build another one.

"make it as easy as possible to be able to upgrade the machine rather than have to build another one"

gotta agree with that one my machine is a couple years old now but still runs fine when i was laying mine out i did try to keep it "open" for changes
the main change i have done is removed the rails for the x axis and replaced with igus w my rails were joined together when i got them and no band saw my z axis was done with a dolphin slide so it bolted right to the igus slides need to get some new pics up

Agreed! Everything will be bolted together with socket head cap screws this time, instead of drywall screws. I should be able to swap out most major components if something has too much flex.

The first machine would need a major rebuilding if I wanted to salvage it. Instead of 3/4" x rod, I would need 1" or 1 1/4". The Y axis would need 3/4" instead of 1/2". and a lot of the MDF needs to be doubled up. That's why I opted to start with a clean sheet of paper.

Question: you are using a lot of plywood on your machine instead of mdf or aluminum. What's your logic behind that? Just wondering...

Mike...

Mike, what I'd recommend when you start building this router, is to try to allow for things to be easily changed if they don't work to your expectations. Ideally, you don't want anything to be able to flex at all. If you find they do, make it as easy as possible to be able to upgrade the machine rather than have to build another one.

Do you have a before and after photo of your x axis?

Mike...

gotta agree with that one my machine is a couple years old now but still runs fine when i was laying mine out i did try to keep it "open" for changes
the main change i have done is removed the rails for the x axis and replaced with igus w my rails were joined together when i got them and no band saw my z axis was done with a dolphin slide so it bolted right to the igus slides need to get some new pics up

I'm using Baltic Birch. It's lighter and stronger (and costs twice as much) as MDF. A good place to get it is Pontiac Plywood. About 2 miles south of Great Lakes Crossing.

I've been there before!!! They've got the Baltic Birch in 4x4 and 5x5 sheets.

So it's lighter and stronger. Are you concerned at all about the effects of humidity, or contraction & expansion?

Mike...

I'm using Baltic Birch. It's lighter and stronger (and costs twice as much) as MDF. A good place to get it is Pontiac Plywood. About 2 miles south of Great Lakes Crossing.

I've had some parts I cut 2-3 years ago sitting in my garage, and they stayed perfectly flat. I've had MDF warp overnight. I'd paint it with a good primer and exterior latex, or the always popular hammer paint. If using the hammer paint, get the quart can and roll it on. Use 2 coats, too. Lowes sells the quarts, Home Depot has spray cans only.

Are you sure they have 4x4? Last time I was there they only had 5x5. I bought mine at Public Lumber, 7 mile and I-75. They'll cut it down for you. The 5x5 wouldn't fit in my truck. :(

A thought on the existing machine. Your rod end support isn't stiff enough. Put a second set of rod supports a few inches in from the existing ones. Two supports at both ends would be significantly stiffer and so reduce the defelction. It would reduce the travel of the aix though..

Andrew

I still dont understand why you dont just use aluminum for the entire construction. 80/20 will cut and tap the two main supports, and the rest can be made of 1/2 precision plate that you can cut with a regular wood table saw (with the right blade) or even a cheap jig saw. Otherwise build it just like you would with MDF, except forget about extra support, you wont need it.

I still dont understand why you dont just use aluminum for the entire construction. 80/20 will cut and tap the two main supports, and the rest can be made of 1/2 precision plate that you can cut with a regular wood table saw (with the right blade) or even a cheap jig saw. Otherwise build it just like you would with MDF, except forget about extra support, you wont need it.

Probably because aluminum will cost 4-5x as much. :)

just do a search my photos from when i first built it are on here need to get some pics of the change ...... as far as the mdf i used it for my gantry and my cut table i just painted the gantry and for the cut area i used plastidip its a rubber for dipping your tools in less slippery and made the wood air tight also on my cut area i routed channels that i put neoprene in in the center a hole is for a vac that holds the work down (no clamps)

Probably because aluminum will cost 4-5x as much. :)


And because I already have lots of good plywood sitting in the garage!!! ;)

Mike...

Theres nothing wrong with wood or MDF, but aluminum of course is far superior. If you have the lumber and dont have cash, well, thats a no brainer. But if your spending a bunch of cash anyway, and your experiencing problems with flexing, then it starts to make less sense sticking with the wood, especially when mixing it with other structural aluminum components. My feelings are go with one or the other, a low cost JGRO style MDF frame, or an all aluminum frame that costs more and doesent compromise. I was just trying to get you to consider the all aluminum approach, and wanted to assure you that it's something you could do with out too much added effort or cost.

Ive got a MDF table on my router and I'm replacing it because of the sagging, warping and flexing. I spent alot of time trying to strengthen underneath it, and I've concluded its not worth the hassle and frustration trying to make it sit flat and true, so I'll replace it soon with a 30.00 piece of 1/4 in. aluminum plate.

Just my two cents, the fun of this is doing it the way you want to, so dont mind me, I'm just sharing my experiences with you on MDF. Plywood may end up being completely fine, try it out, you could allways replace the wood panels with aluminum later if they dont hold up or start flexing.

(The above comments are the author's, and do not represent those of CNCZone or its management.)

I don't know where you are getting your plate, but the online price I have seen for 1/4" plate is $22 per square foot in a 3' by 4' sheet ($267 for the sheet).

rweatherly, Try this place for online metal purchasing. They don't always have the exact size but the price is pretty good.
http://www.lowcostmetals.com/index.asp?PageAction=VIEWCATS&Category=103

They do have good prices -- that's about $10 per square foot. Still MDF, is less than $.50 a square foot. I have just finished a JGRO, but have not yet gotten to the point of checking deflection. The most cost effective way of increasing the rigidity is probably a combination of MDF with steel angle -- depending on which part you are working on. (Steel has 1/3 the deflection for the same cross-section and is a lot less expensive)

Richard

I get mine from Industrial Metal Supply in Irvine, CA. and I pay scrap ( around $2.00 a pound, but they let me weigh it myself and theyre not too picky) for anything on the remnant floor. Buying metal online is just way too expensive, it just doesent make sense. If you find a local supplier of plate you can get a good price on thier remnants. Sometimes if I buy a bunch of wierd shaped stuff they will knock 50% off, or give me steel price. Its like anything else, if you look around you can get it reasonable, but online is definately not practical compared to shopping around on the right side of town. You also have to buy what they have cut, not what size you want, its much, much less.

(The above oppinions are the authors and do not represent the views of CNCZone or its management.)

i forgot (it was late ) i do have alum. angle running across under my mdf for cutting so no sag just cut a small notch and it lays between the alum ends and supports it