[Bruggles]

I am just wondereing where you think the best place for ballscrews is on the gantry. I am looking for accuracy and obviously I would like to keep cost down, but I don't want to cut corners if one way is significantly better than the others. I am planning on using thompson type linear bearings, the pillow block and round shaft type

one Ballscrew under the table in the center linear bearings on the top of the table

one ballscrew under the table in the center linear bearings on the side of the table

two ballscrews one on each side of the gantry connected by belt

two ballscrews, one on each side with individual motors

On both of the two ballscrew ideas the linear bearings would be raised, so that the y axis basically sits on the linear bearings, and the shafts would mount on top of the raised surface. On the single ballscrew ideas, the linear bearings would be on the table with a raised gantry

I plan on making a cnc router for routing plastic and aluminum, that is 48x48.

[eman5oh]

I like putting the ballscrew and the bearings under the table. This gives them better protection against dust and other crap that can get into them.

[Bruggles]

where under the table do you mount the bearings, just on each side, as if they were on the top, but instead on the bottom?

[eman5oh]

where under the table do you mount the bearings, just on each side, as if they were on the top, but instead on the bottom?

Just under the table on the out side. The particlar placement will depend on the type of linear bearings used and how you want the foerces to act on them. I am useing round shaft with thomson style bearing, these have the same rating vert and horz mount, but if you use a THK style rail the load ratings will differ.

[Bruggles]

Has any body found that with just the one ballscrew under the table that it sacrifices some accuracy? is there anybody that has done both? Just trying to solidify my design in my head before I start the drawings, so i don't end up doing like 50 revisions.

[ViperTX]

Bruggles....."Ah, no....isn't that the same way that all mills are designed."

[Bruggles]

True, mills do have just one ballscrew, but they don't have moving gantries, they generally have moving tables. I don't know how much difference that makes, but I know that a gantry is a lot more of a cantelever than a table, where its ballscrews and ways are allmost in the same plane, which is only a couple of inches below the top of the moving surface (the table), where gantries are often say 2-4 inches from the ways to th ballscrew, and the top of the gantry is around 12 inches from the bearings. I'm not trying to argue, but there are differences, I just wanted to see if anybody had experiences with both setups.

[WhiteTiger]

I prefer a planer type design, as mentioned by Bruggles. Moving table x axis, y and z on stationary overarm. Allows for much more rigidity and single screw x axis is very stable with the rails mounted solidly to the bed plate.



Tiger

[widgitmaster]

It seems logical to place the centrtline of the screw at the same center of motion of the slide bearings, if the screw is higher or lower, it creates a binding force. This would hold true with either a movable table or gantry design.

My 2 cents!
Eric

[Bruggles]

Widgetmaster, I was just arguing with myself to figure this thing out and you are completly correct.

The closer to the plane with the bearings both the top of the gantry and the ballscrew are the better and more ridged. And from basic statics principles it seems that it doesn't make much difference the location, as long as both lever arms(the distance from the bearings to either the gantry or the ballscrew) are minimized, so where ridgitity will be affected is if I can minimize one lever arm without changing the other, so the moment forces wouldn't be so great on the bearings supporting the gantry. I hope that made sense, and thanks for all of the replies you have helped me to figure out what system I am going to go with. As far as ridgitity is concerned I believe the two ballscrews would be more rigid, but not necessarily worth the cost, it seems that I could go with larger/higher quality bearings for cheaper and obtain similar rigitity. The one thing that would make two ballscrews worth it is if I had to span a large distance, and needed to have supports in the middle to eliminate table and bearing sag. For this current table I should be able to overcome those issues without difficulties. My final design will consist of bearings on the side of the table, so that the round thompson linear bearings will have the same strength for holding the gantry and for withstanding the gantry lifting do to cutting forces etc. It will also yeild the benifit of dust protection. I also plan on using a spinning nut idea to try to maximize feed speeds and maintain accuracy. Thanks for the help, I will try to post some drawings etc in the diy router section when I am closing in on my final design. Once again thanks, you all have helped me in my decisions very much.

[tizken]

I would think that the force that the ball screw is pushing against is up at the cutter not the linear bearings so you would want the screw as close to the bottom of the table as possible.

[keithorr]

One screw centered under the table with linear rails mounted under the table.
A crossmember holds the recirculating ball nut and the caged ball guides. The uprights attach to the crossmember. The rail "hangs" under the table near the edge.

[JRoque]

Hello. I'm going to get flamed for this but here goes it. I have 1 screw on the side with 2 bearing blocks per side that sit 8" above the table. To reduce racking, I installed a .5" steel plate in the shape of an "I" across the X axis. This extends the bearing blocks a few inches outward to keep it fairly rigid. In fact, pushing on the far side of the X axis (furthest from the Y screw) twists the machine's aluminum frame before any racking can be measured. This setup has worked well so far cutting PCB, plastics, wood, aluminum and steel. If I ever see the need to improve rigidity, I'll add a second screw on the opposite side. Next up and crisscrossing the frame with scaffolding braces.

JR

[bushwakka]

I made a router with 700mm across the gantry (Y), and a single ballscrew underneath. I had 2 issues.
The cutting forces cause the gantry to skew, lots at Y=0 and Y=700, and none at Y=350,
A slight amount of play in the bearing (bearing to rail, and bearing insert to bearing housing) would see around 1mm of play (y=0 cutting x-, versus y=0 cutting x+).

The other issue is table support.
1. Its difficult to support the cutting surface of your router as the gantry underbridge limits what you can do under the table, particularly if you are trying to keep everything in close alignment.
2. You can only have legs at the corners, as the underbridge prevents you putting additional support under your long sides.

If I was redesigning mine, I would put a ballscrew either side. I don't think it is majorly important whether they are coupled with a belt or driven individually.
This allows you to following benefits
* cutting force skew is removed
* table support of long rails can be improved (more legs of bracing)
* table cutting surface can be stronger (more room underneath for support
* gantry can be lighter (no underbridge required)

Cheers
Bushwakka