[JeremyFisher]

Hey everyone,

I've put together a CAD model of my desktop router which I think is sturdy and now I’m hoping that the community can point out any design flaws I might have overlooked.

In short, I want it to be built like a rock (by desktop router standards).

I intend on using this machine for wood (MDF, Hardwood and Merbau) and occasional PCB milling, but, I want to have the option to easily cut aluminium (for model helicopter parts).
So easy ally cutting with a respectable surface finish and 0.1mm (0.004”) accuracy is my design criteria.

The current design is mostly held together with bolts and screws. The spindle will be held by an ally clamp.

Some specs:
Working Area: 396mm x 311mm x 100mm (15.6" x 12.2" x 3.9”)
Work Surface: 15mm (0.6") ally plate
Gantry Clearance: 92mm (3.6")
Spindle: 500W 33,000 RPM variable speed (5000-30,000 RPM) router
Actuation: NEMA 23 287 Oz-In Stepper, C7 spec ballscrews with 5mm lead
Material: All steel, except for ally z-axis assembly and ally work surface
RHS Uprights: 75x50x4mm (3" x 2" x 0.16") All functional surfaces milled for squareness/trueness
X-Axis: 2 x 16mm fully supported rail
Y-Axis: 2 x 20mm & 1 x 16mm unsupported rod
Z-Axis: 2 x 16mm unsupported rod
Distance between y-axis 20mm rods: 120mm (4.7")
Distance between x-axis bearings outer faces: 175mm (6.9")
Distance between z-axis bearings outer faces: 120mm (4.7")
Total Gantry mass at x-axis: 21.3kg (47lb)

Notes:
I know that for maximum stiffness I should use supported rails all-round, but I've already purchased the 16mm and 20mm rails and the associated bearings, so now I'm stuck with them.
The teal bars in figure 2s are the leadscrews.
The y-axis bearing/rod housings are a funky shape to cut down mass.
The brown block is just a block of merbau timber, to prevent sagging of the work surface and maybe give some damping too.
The z-axis motor will be mounted by all four bolt holes, not two as shown in the pictures.
The y-axis and z-axis lead screws will be protected by bellows or ally sheet.


So now I have three questions for the community:
What would you change or improve on the design to beef it up or simplify it?
Would my method of linking the x-axis lead screw to the gantry uprights create a weak spot prone to deflection at that point?
Considering that the y-axis rods will be firmly clamped in the bearing/rod housings, would I have trouble with alignment of the gantry? It's only a matter of making sure that the housings are machined identically, yes?


Thanks for any input!

[Phife]

Heres my opinions,

X axis screw to gantry is weak, you dont want to have such a long lever on your x nut to the moving gantry, it will want to flex and bend. I suggest redesigning this.

Whats up with the M shaped Y supports? a simple square or rectangle will be much more rigid and alot easier/cheaper to make. Trust me, the little amount of weight saved by making them like that is not worth the effort.

Also dont use a clamping end support on those Y ends, use the proper end supports that come with the rod, and use a setscrew on the Y ends that are M shaped. this will make your Y much more rigid.

You will have a tough time machining aluminum with that setup, dont expect a wonderful result. Those unsupported rods will flex.

Also that blue block, Im assuming its some sort of Square steel tube? how will you attached the open end to your X bearings? and how will you cut it perfectly square?

You have wide spacing for your X blocks but only a small square contact area to that Blue tube, you really should make use of the wide bearing spacing by either using a wider blue tube or adding an angled support to the blue tube.

[tskguy]

Hey,

Phife is right, those unsupported rods for the Y axis will flex big time if you try and mill aluminum. Sell them on ebay and use the money to get some linear rail. I also would consider a fixed gantry design with a movin table. For the size machine you are building you will end up with a much stiffer and more accurate setup.
I would also connect the y axis together with something mor substantial.

Eric

[cyclestart]

I also would consider a fixed gantry design with a movin table. For the size machine you are building you will end up with a much stiffer and more accurate setup.

I agree,. fixed gantry machines make good sense in that size. Less complicated and potentially more rigid and less expensive.

This site is still under construction but has some good advice:
Homemade CNC Router The Builder's Guide (FREE!)
(if the site owner is reading this, please keep up the good work)

Don't want to rain on your parade Jeremy but you did ask for opinions. No shortage of those here LOL.

[RomanLini]

Hey everyone,

I've put together a CAD model of my desktop router which I think is sturdy and now I’m hoping that the community can point out any design flaws I might have overlooked.

In short, I want it to be built like a rock (by desktop router standards).

...

Thanks for any input!

Nice CAD work! To critique this design I'd have to echo other people's comments because they are right of course.

In that size machine you will get huge rigidity gains from going to fixed gantry and moving table. I'll give my input having built almost exactly the machine you described (by specs) but done as a moving table build.

My design uses a "L shaped gantry plate" bolted directly to the base plate but at 90'. The L shape plate cannot flex at all trapezoidally giving massive rigidity from one cheap part. I see you are in Australia, you can ring one of the big aluminium suppliers to order the 2 main plates, they will even cut to size and cut the L shape for you on their table saw.

Once you have the gantry and base plate rock solid that can't move relative to each other, it's pretty easy to make the 3 axes pretty solid as you have flat rigid plates to bolt the X and Y directly to, and Z is the only axis mounted on another axis so you just need to keep the geometry good so you don't amplify slop.

I used plastic linear bearings as they are "snug" and it's more like a milling machine bedways; the precision of no slop but best at slower speeds. On a 400x300mm machine speed is not a problem!

You can see my little machine here; http://www.cnczone.com/forums/cnc_wo...ll_router.html and you are welcome to copy any ideas ie the "L shaped fixed gantry plate".

[JeremyFisher]

I've read on this forum time and time again that unsupported rod will deflect, so I did an experiment with the 20mm rods I bought.
I hung 12.25kg (27lb) off one of them in a simply supported setup and using a dial gauge, I found it deflected 0.1mm (0.004") at the centre. I didn't account for deflection due to their own weight, but at 400mm I think it would be negligible. I concluded that fully fixing each end, and having three of them would yield a rigid enough structure.

Also don't use a clamping end support on those Y ends... use a setscrew on the Y ends that are M shaped. this will make your Y much more rigid.
...
Also that blue block, ... how will you attached the open end to your X bearings? and how will you cut it perfectly square?
...
you really should make use of the wide bearing spacing by either using a wider blue tube or adding an angled support to the blue tube.

Why do you suggest that using a setscrew would make it stiffer? I was thinking that clamping around the rods with the entire housing would make for a more rigid structure. That's how the commercially available support blocks are designed.

The blue RHS was to be machined on all of its functional surfaces, thus making for a perfectly square end.
That aside though, the three bolt holes to which the m-shaped plates are attached would be over sized to accommodate for adjusting alignment of the gantry and compensating for any deformation due to welding. Would that be a sensible idea?

I agree about the angled supports. I might add some webbing there if I stick to that design.

Nice CAD work!
...
In that size machine you will get huge rigidity gains from going to fixed gantry and moving table.
...
My design uses a "L shaped gantry plate" bolted directly to the base plate but at 90'. The L shape plate cannot flex at all trapezoidally giving massive rigidity from one cheap part.

Thanks, I figured I might as well plan as best I can by having a good model and iron out any quirks on screen rather than in the workshop. I've seen your machine and good work on that. I was in fact planning on making a copy of your tool height setter as one of my first projects for its ease and usefulness.

When you say 'L shaped gantry plate', do you mean forming an L shape by bolting together flat sheet yes? Your gantry is a bit complicated looking (sorry) but from what I can tell, it's a plate going across the y-axis to which the y-axis rails are bolted, and that plate is attached to the table surface by bolting it to another perpendicular sheet?

The reason I have avoided the moving-table design was for it's large footprint. I didn't want to end up with a large machine with a comparatively tiny work area. The current design is already inefficient with space, so I didn't consider the fixed gantry as an option.

But if 4 out of 4 people suggest fixed gantry, I guess it's back to the drawing board for me. I'm still not sure though on whether I will use rods or supported rails on the gantry though.

Thanks everyone.

[RomanLini]

The "L shaped gantry plate" on my machine is one piece of flat 10mm plate, cut into an L shape. Because of that shape it can't flex trapezoidally like a square flexing into a diamond, which is possible on most other gantry types. Then it has just box section bolted on the back and uprights etc.

But the simplicity is one flat base plate, and one L shaped gantry plate, then all the trimmings.

I like the way you measured the flex of the rods, and you are right they will be pretty rigid if clamped at each end within the structure. But even at 0.004" flex you might get some bad cutting vibration because the cutting tool puts big pulses of force on the rails which act like guitar strings...

And thanks yeah that little Z axis tool setter is the best hours work I ever did on the machine. It's never missed a beat and always detects to within 0.01mm, I can even measure the difference in tool growth when the router shaft and tool get warm, it expands about 0.03mm longer.