Could somebody link me to some reading on the basic mechanical design of these

General CNC CNC Routers
, ,
1

Could somebody link me to some reading on the basic mechanical design of these machines? I’ve built 2 3d printers and I think I know how to do all the electronics stuff just fine, but all I know about CNC mill mechanics is “It should be tougher than printers.”

Basic questions right now are: What are the advantages of the different axis arrangements? e.g. stationary tool with the part moving in XYZ, stationary part with XYZ on tool, XY on the part with Z on the tool, etc. Which would you go with for a DIY?

and

Roughly how much force am I going to want my machine to be capable of on each axis? Assuming the absolute hardest material I want to mill is mild steel? What if all I’m going for is aluminum for now?

Lastly, can anybody recommend a reasonable spindle in the $75-$200 range? I don’t want to go so cheap that I’ll just replace it in a month, but I don’t want to go overboard either.

Thanks

2

Alright so from a brief survey of posts in here it looks like stationary part with XYZ all on the tool is the most popular.

I’ve been looking at mcmaster for linear motion parts, and I’m thinking about P/N 5966K26 + 5966K16 for ball screws. They can handle tons more load than I imagine I’ll be using, and they’re barely more expensive than leadscrews + anti-backlash nuts and come with ~30% better efficiency. Anybody know of any reasons not to use these? The only issue I’ve come across is that they need some form of preloading to prevent lash, but I’m sure I could come up with something to add a little friction to the system if it becomes an issue.

For actually constraining the axises, I have 12 LM12UU linear bearings I bought by mistake off ebay, so I’m wondering if 12mm precision rod is stiff enough for this? Using 2 rods per axis and with axises either 1 foot or 2 feet long?

Lastly, I still don’t really know anything about spindles.

Thanks

edit: Oh also, assuming I use those ball screws, my 2 options for motors that I’m looking at right now would provide either ~400 N of axial force or ~1100 N of axial force. Is 400 enough? It would save me boatloads of money if it were because I have stepper drivers to match those motors already, and the motors themselves are $15 instead of $50.

3

Alright I found some scholarly papers that I sort of understand, it looks like I’m going to want the ~1100 N option.

4

Diameter (mm) :

Length (mm) :

Bars :

Load (Kg) :

E (N/mm^2) :



Bars weight :

Bend (middle point) :

paste this code to a new html document and load it in your favorite browser, it will help you to find a good value for your rods.

5

in short 20Kg of load (you have to add every single component which is “above” the rods) generates 40microns of bend with 1 foot rods and 360microns with 2 foots rods, as you can see there is much difference.
If you have to mill PCBs you can’t go for the 2 foots bars but with wood it should be ok.

6

Diameter (mm) :

Length (mm) :

Bars :

Load (Kg) :

E (N/mm^2) :



Bars weight (Kg):

Bend (um) :

new version

7

@Nick_Parker Not sure what the prices are like on MacMaster but I’ve been sourcing Linear motion hardware direct from China via Aliexpress, I have bought 16mm lead screws Linear rails/carriages and 12mm precision rods, all of which have been good quality and much cheaper than the alternative US suppliers I’ve looked at. http://www.aliexpress.com/wholesale?SearchText=ball+screw&catId=0&initiative_id=AS_20130528172248

8

Ooh yeah, those do look cheaper. I’ll poke around some.

9

Holy hell those rods are cheap. 4 350mmx12 for 4 bucks.

10

Ok thanks for your help folks. I’m looking at SBR20 rails now and SFU1605 ball screws. I still have to decide what to do for the actual frame, but it will probably just be some hefty aluminum extrusions or water jet cut steel plates. Depends mostly on price.

11

12mm is pretty small for a linear rail for a cnc router. Even fully supported, it’s on the small side. I would say 3/4" at least for the rails, and at least 5/8" ball screws. Nook Industries is a good place to get ball screws from.

I know alot of people frown on welding the frame together, but if your a good fabricator, you will be able to get it very square. This is also a very cheap way to get a strong frame.

12

Yeah so the current plan is to waterjet cut a steel frame, find a good way to grind the edges to perfect 90 degree angles (the water jet cuts a slight angle), fit it very well together, spot weld it, assemble the machine and make sure the axes are absolutely square, then finish the welds.

Or I may just go with Alu extrusions, depends on the prices.

I’ve also decided on SBR20 rails for sure now, SBR12 is much weaker and actually comes out to the same price because of aliexpress shipping cost trickery.

I’ll be tracking further progress at http://nickscnclog.blogspot.com

13

That sounds hard to do. Start with a welded tubular structure and design gussets as required.

14

Paul: I find cutting tubes with any accuracy to be a rather difficult proposition, so I would say your suggestion is much harder to do.

15

Yeah I’m currently on T-slot with maybe some waterjet cut supplementary parts to make it all fit together. T-slot is surprisingly cheap for how strong it is.