Build Thread- O.K., here is another DIY plan

[JerryBurks]

Right now I am using an off-the shelf CNC Shark Pro Plus. Actually a good starter but I am not happy with the rigidity of the design and the small size. So I started collecting stuff for a new machine and did a complete new design. I read some of the general topics here about screws, steppers and linear bearings but otherwise ignored proven designs (wouldn't that be boring, being an engineer....)

Here is the plan (I already have the materials listed below so I better make it work....):
- table size 48" x 36", made from 1.5" T-slot 80/20 rails (type 1545). Machine operating envelope: 36" x 33" x 8"
- dual Y-axis steppers, coupled with 8mm HTD timing belt. Total 4 steppers NEMA34 with 1600 inch-ounce each for XYZ
- Planet-CNC USB controller
- all supported rails, U-box-style gantry design, 12" spacing between guide blocks.
- ground 15mm ballscrews, 1100 mm long from Kuroda, better than 0.001mm backlash. With 15mm pitch a little more than I like but I hope the strong motors will compensate. Got a pretty good deal on these surplus items ($150 each) but not sure yet how to protect from dust.
- sub-structure and gantry made from 3/4" bamboo plywood (stronger than birch and still easy to machine with my current CNC machine). Not exactly cheap, though. The weight/stiffness ratio should be better than aluminum in thick sections and it can be easily glued/screwed etc..
- machine is intended mainly for wood, plastic and occasional light aluminum work


Well, please shoot holes in my plan before I start cutting materials....

Thanks, JB

[ger21]

Well, please shoot holes in my plan before I start cutting materials....

OK,

1) Does Planet-CNC allow you to slave the motors? I just looked at the manual, and didn't see it mentioned? Or do you have another plan?

Typically, when using 2 motors to move the gantry, there's no reason to tie them together with a belt. It just adds extra cost and complexity. It might even hurt performance, as there's a pretty good chance that the motors will always be fighting each other a little bit.

2) 1600 oz motors are probably 4 times more than you need. The most common choice for a machine that size is a Gecko G540 driving 380 oz steppers, running at 48V. There's a pretty good chance that the smaller motors may actually give you quite a bit more speed.

3) I'd rotate the rails the gantry is riding on 90°. Mounted sideways like that, all the weight will be carried by about 1/3 of the bearings.

4) Ditch the bamboo plywood, and go with Baltic birch. It's much stronger, more stable, and cheaper.

[JerryBurks]

Hey, thanks for the feedback! here my rebuttal......

1) Does Planet-CNC allow you to slave the motors? I just looked at the manual, and didn't see it mentioned? Or do you have another plan?
Typically, when using 2 motors to move the gantry, there's no reason to tie them together with a belt. It just adds extra cost and complexity. It might even hurt performance, as there's a pretty good chance that the motors will always be fighting each other a little bit.

Well that may be a weak point in my plan. I was going to drive both steppers in parallel and the timing belt was more intended to prevent going out of sync altogether. Under normal circumstance it would not be needed. But I wanted to prevent the motors from ripping the machine apart if one should fail.



2) 1600 oz motors are probably 4 times more than you need. The most common choice for a machine that size is a Gecko G540 driving 380 oz steppers, running at 48V. There's a pretty good chance that the smaller motors may actually give you quite a bit more speed.

Agreed, but I am using 15mm pitch screws since I got them cheap (vs. the usual 5mm pitch recommended) so that may be just right?

3) I'd rotate the rails the gantry is riding on 90°. Mounted sideways like that, all the weight will be carried by about 1/3 of the bearings.

Thought about that hard and changed the design already a few times. I have yet to see some directional load measurements under usual operating conditions. At the end I believe the actual gravitational weight of the gantry (maybe 50 pounds in my case?) can be handled by the rails in any orientation. The machining forces however can apply in any direction except "up" so I am not sure if it makes a big difference. Admittedly, I opted for the horizontal layout for simplicity reasons.

4) Ditch the bamboo plywood, and go with Baltic birch. It's much stronger, more stable, and cheaper.

Well I did some research and I dispute that. While the Baltic Birch is indeed a bit cheaper, the bamboo is supposed to have a 20%-30% higher bending modulus. Real good quality Baltic is not easy to come by, anyway.
Besides that the bamboo plywood looks really cool (you know like the cutting boards in the expensive kitchen shops)

[ger21]

Agreed, but I am using 15mm pitch screws since I got them cheap (vs. the usual 5mm pitch recommended) so that may be just right?

I'd never recommend 5mm pitch screws, unless you want a slow machine. For a router, the best balance of speed and precision comes from about 2 motor turns per inch of travel. You're close to that. I still think the 380oz motors are the correct choice. Most guys building 4x4 or larger routers are using rack and pinion with the same 380 oz motors, at 1 motor turn per inch, and are moving heavier machines at 1000+ ipm. I'd expect similar performance with the 380oz motors and your ballscrews.

Real good quality Baltic is not easy to come by, anyway.

Try finding really good quality bamboo plywood. The stuff I've seen is very inconsistent. Some flat, some warps like crazy. We had 300 sheets in our shop 2 years ago. I'd take Baltic birch every time.

[Phife]

Originally Posted by JerryBurks 3) I'd rotate the rails the gantry is riding on 90°. Mounted sideways like that, all the weight will be carried by about 1/3 of the bearings. Thought about that hard and changed the design already a few times. I have yet to see some directional load measurements under usual operating conditions. At the end I believe the actual gravitational weight of the gantry (maybe 50 pounds in my case?) can be handled by the rails in any orientation. The machining forces however can apply in any direction except "up" so I am not sure if it makes a big difference. Admittedly, I opted for the horizontal layout for simplicity reasons.

You will not be able to mount the rails like you have them anyways, the mounting holes for the SBR rails dont line up to the mounting slots on the 8020 in your config.

Also, on your Y axis, when the gantry is cutting and moving away from you in the X direction the entire Y/Z carriage will be pulled away from the supported rails, this is the weakest direction for these supported rails.

[louieatienza]

Originally Posted by JerryBurks Right now I am using an off-the shelf CNC Shark Pro Plus. Actually a good starter but I am not happy with the rigidity of the design and the small size. So I started collecting stuff for a new machine and did a complete new design. I read some of the general topics here about screws, steppers and linear bearings but otherwise ignored proven designs (wouldn't that be boring, being an engineer....) Here is the plan (I already have the materials listed below so I better make it work....): - table size 48" x 36", made from 1.5" T-slot 80/20 rails (type 1545). Machine operating envelope: 36" x 33" x 8" - dual Y-axis steppers, coupled with 8mm HTD timing belt. Total 4 steppers NEMA34 with 1600 inch-ounce each for XYZ - Planet-CNC USB controller - all supported rails, U-box-style gantry design, 12" spacing between guide blocks. - ground 15mm ballscrews, 1100 mm long from Kuroda, better than 0.001mm backlash. With 15mm pitch a little more than I like but I hope the strong motors will compensate. Got a pretty good deal on these surplus items ($150 each) but not sure yet how to protect from dust. - sub-structure and gantry made from 3/4" bamboo plywood (stronger than birch and still easy to machine with my current CNC machine). Not exactly cheap, though. The weight/stiffness ratio should be better than aluminum in thick sections and it can be easily glued/screwed etc.. - machine is intended mainly for wood, plastic and occasional light aluminum work Well, please shoot holes in my plan before I start cutting materials.... Thanks, JB

I actually have the same Kuroda ballscrews that I will be using for a build. I'm finding the hardest part is actually finding someone locally who is willing to machine the ends, properly. I would say that you should use a proper fixed end beaing block at the motor end along with a flexible coupling, so that you do not use the stepper's bearings to carry the thrust and axial loads of the screw.

I also plan to use a dual drive in the long axis, but am considering reconfiguring my design so that the screws and lm rails are under the table, which shoould add protection for the ballscrews (I am unaware if Kuroda makes dust shields for these as they have not returned my emails.) I also plan to use high-speed stepper motors with a 15:36 pulley reduction, which combined with the screw pitch, should give roughly 4:1 gear reduction, so I'll have some speed and resolution.

THe bamboo plywood would give your machine a cool eco-look! Heck, Asus is making laptops with bamboo shells....

[JerryBurks]

Originally Posted by ger21 I'd never recommend 5mm pitch screws, unless you want a slow machine. For a router, the best balance of speed and precision comes from about 2 motor turns per inch of travel. You're close to that. I still think the 380oz motors are the correct choice. Most guys building 4x4 or larger routers are using rack and pinion with the same 380 oz motors, at 1 motor turn per inch, and are moving heavier machines at 1000+ ipm. I'd expect similar performance with the 380oz motors and your ballscrews.

Hm, interesting. Looks like the motors that I bought are a bit overkill. But at least I can traverse then at light speed or maybe just reduce the drive current a bit.

Try finding really good quality bamboo plywood. The stuff I've seen is very inconsistent. Some flat, some warps like crazy. We had 300 sheets in our shop 2 years ago. I'd take Baltic birch every time.

I happen to have a wholesaler in the neighborhood and picked two really flat sheets. These are 3-ply with a single thick core and 2 thinner perpendicular layers. Besides stiffness I like them because is there is serious meat in the cross-sections to screw into, unlike regular plywood.

After all, this was to a degree a choice of just doing it differently and for the looks.

[JerryBurks]

Originally Posted by Phife You will not be able to mount the rails like you have them anyways, the mounting holes for the SBR rails dont line up to the mounting slots on the 8020 in your config.

I was not planning to use the 80/20 T-slot to bolt the rails down but rather drill 2 rows of holes in the aluminum and tap with M6. They are 30 mm apart while the profile is about 38mm

Also, on your Y axis, when the gantry is cutting and moving away from you in the X direction the entire Y/Z carriage will be pulled away from the supported rails, this is the weakest direction for these supported rails.

All I found about load directions was a data sheet from Thomson Linear. They are not the rails and blocks I have but similar. It said the load capability "sideways" and "up" to be 50% of the nominal maximum load.

Considering the worst case condition for the z-axis to run into the table while close to one gantry column this would yank a "normally" mounted rail upwards (in case of my motors/screws with approx. 1000 lbs). This is already not the preferred direction and mounting the rail sideways has the same load limit. In comparison to the motor forces, the weight is peanuts.

The worst case for an x-axis collision should even be better than mounting "normally" since one of the block sets is loaded in the preferential "down" direction while the other one can still contribute 50%.

If it turns out the loads are too high I still have the option of adding more ball guide blocks. At $15 each they are not too expensive.

[JerryBurks]

Originally Posted by louieatienza I actually have the same Kuroda ballscrews that I will be using for a build. I'm finding the hardest part is actually finding someone locally who is willing to machine the ends, properly. I would say that you should use a proper fixed end beaing block at the motor end along with a flexible coupling, so that you do not use the stepper's bearings to carry the thrust and axial loads of the screw.

Same problem here. I plan using the fixed bearings on the opposite end of the motor. I.e. drill out a 3/4" shoulder screw on a lathe, slip over the ballscrew end and weld carefully through a few cross-holes. The actual bearings are a pair of 3/4" conical roller trailer bearings. That should allow me to pre-load for zero play. On the other side the ballscrew will be connected to the motor with a custom made clamp coupling and the motor hung with some axial movability.

I also plan to use a dual drive in the long axis, but am considering reconfiguring my design so that the screws and lm rails are under the table, which should add protection for the ballscrews (I am unaware if Kuroda makes dust shields for these as they have not returned my emails.)

Kuroda has the "Lubseal" wipers that fit to the recesses of the ballnut. But my screws don't have them and I don't know if that can be retrofitted and where to get. I would like to add moving curtain style dust protection later but not sure yet how to do this practically.

[DonFrambach]

Originally Posted by JerryBurks I was not planning to use the 80/20 T-slot to bolt the rails down but rather drill 2 rows of holes in the aluminum and tap with M6.

Tapping into an aluminum extrusion may not be a very secure method to attach your rails!

[Phife]

Originally Posted by DonFrambach Tapping into an aluminum extrusion may not be a very secure method to attach your rails!

Exactly what I was gonna say, there is very little material there to tap, I would advise against that.

Also, I wouldnt rely on the motors bearings to support your shaft, and i wouldnt "hang" your motor or allow it to have play in any direction, this is a very bad idea. Allow your coupling to take up any mis-alignment issues, flex couplers are pretty cheap and work well..

And, i dont think you understood what i was saying about your Y/Z supported rods. I will draw up an illiustration later when I get home.

[ger21]

Looks like the motors that I bought...................

Did you buy one of those Wantai kits with the 24V power supply?