[blau_schuh]

I'm building (what hopefully will be) a pretty solid and rigid gantry-type cnc machine out of 80/20.

The gantry will ride on overkill THK HSR35 rails.

I would like to experiment with mounting a ~35mm diameter ballscrew on one side of the gantry, rather center & under the table.

I can see the possibility of racking with some less stout (v bearings) setups, but don't see it occurring easily with a good quality linear bearing setup.

Any opinions on this?

Eventually going to be moving to a nut driven setup using a 75" long ~1.5" diameter screw mounted on the side as well.

[Al_The_Man]

Have you considered a fine mesh R & P set up? One method is to use dual racks and a shaft across the table to drive the opposite side. This prevents racking, in large routers, the accuracy from using B.S.'s is not generally required.
Al.

[blau_schuh]

Have you considered a fine mesh R & P set up? One method is to use dual racks and a shaft across the table to drive the opposite side. This prevents racking, in large routers, the accuracy from using B.S.'s is not generally required.
Al.

Hi Al,

Not considering rack and pinion as I already have a variety of ballscrews on hand and I wanted to play around with a nut driven big ballscrew. Although I may at some point (be forced to?) rethink the design.

Interesting idea though with the long shaft. I like the sound of that and may use it on a future plasma build that im collecting parts for.
Wondering out loud what kind of diameter is needed on the driven shaft on say a 5' span in order to not get the equivalent of screw whip during really fast rapids.

[Khalid]

90%chance is you will get racking issues...

[blau_schuh]

90%chance is you will get racking issues...

The rails and trucks in question...
Attachment 248994

As I understand it (explained pretty badly): Racking is where the far side of the gantry momentarily (and maybe even continually in the worst case) lags the near (driven) side of the gantry due to a near force causing slight rotation at the rail mounts (or deformation in the gantry frame itself). I can easily see how a less costly v block/roller bearing setup would easily be tweaked/rotated when during the initial force due to inherent slop in the design, and how this would be easily noticeable to the naked eye in the final product.

Is it safe to assume that high quality linear bearings & trucks have almost zero play and aren't going to rotate with a reasonable amount of force applied? I'd have to guess Yes. This leaves the only possibility for racking to occur in the actual gantry structure itself...where the gantry actually deforms momentarily. Not sure how to measure this without some high speed data gathering of position on both sides of the gantry, as the racking would have to be in the what, a couple hundredths?... thousandths? ... tenths of thousandths? i'd guess that it may only last for a fraction of a second as the far side of the gantry whips back and loses position but then whips forward and gets back to expected position.

[Al_The_Man]

Another problem by making both side tightly compliant, which you need to do if only one side is driven, is that binding etc can happen due to machine growth etc with temp change etc.
To prevent racking when only one side is driven usually means a wide side span to each side of the gantry sides, and this in itself is often undesirable.
Also a gantry used for machining/routing etc loading can occur on one side, as opposed to Plasma etc that does not have this problem so much.
Al.

[tjschweizer]

Look up a "moving knot". I am on my phone so I cant link anything, but google "moving knot cnc"

[rbjem]

I would not recommend a single side drive gantry.

As Al touched on I think you will have problems with racking (I sure did when I had a center driven design) unless you design/engineer it to work while driven on one side, in which case the design will be impractical.

[blau_schuh]

I would not recommend a single side drive gantry.

As Al touched on I think you will have problems with racking (I sure did when I had a center driven design) unless you design/engineer it to work while driven on one side, in which case the design will be impractical.

I didn't have any perceptible issues with my previous small machine (~3.5' x 2' cutting envelope). Center driven ballscrew, 4' HSR35 linear rails (smaller trucks than current rails). 8020 3030 frame that was crudely tossed together.

Do you have a link to your center driven design? or any pics you can share?

[rbjem]

Sorry I don't have access to any pics right now (sitting at the airport starting some vacation time). However, my gantry span is ~38". And now that I think about it, I was using cncrouterparts skate bearing system at the time, but am now using thk20 equivalents.

I think Al has it right though that you would have to over build so much that it far cheaper and easier just to do a dual drive setup.

Eta: I have also reinforced the $hite out of the gantry and frame since the cncrouterpart days.

[blau_schuh]

Sorry I don't have access to any pics right now (sitting at the airport starting some vacation time). However, my gantry span is ~38". And now that I think about it, I was using cncrouterparts skate bearing system at the time, but am now using thk20 equivalents.

I think Al has it right though that you would have to over build so much that it far cheaper and easier just to do a dual drive setup.

Eta: I have also reinforced the $hite out of the gantry and frame since the cncrouterpart days.

I wanted to avoid buying more screws oh well.. I have some pretty nice high quality star screws sitting in the basement I wanted to try. I may try it just for kicks since the 8020 allows for such easy mounting.

The final plan was to mount a large diameter (45mm) ballscrew with a 10mm pitch to the side of the gantry and do a nut driven design and get "scary" fast rapids . Thought it might be fun to do even if it isn't too practical. I have two of these large diameter screws and wanted to eventually sell one of them. Having both of those suckers on the machine would be pretty insane though!

[rowbare]

You could mount a single ballscrew under the table and drive both sides with a tie plate. That would better balance the thrust. Also you can reduce the tendency to rack by placing your bearing blocks further apart.

If you design in the ability to add a second screw or convert to R&P then all you lose if it doesn't work to your satisfaction is time...

bob

[Khalid]

You just calculate if one end moves 1mm off what will be the displacement at the other end... Racking will always occur believe me..

[Meduza]

I think you might get away with driving it at just one side if you space the blocks on the driving side far enough apart, only a single block at the other side,and design the gantry with a very strong triangular frame.

[cgallery]

I think you might get away with driving it at just one side if you space the blocks on the driving side far enough apart, only a single block at the other side,and design the gantry with a very strong triangular frame.

That idea seems promising. But my only experience is with a tiny table-top K2, but I'm just saying that seems like a clever idea.

[louieatienza]

35mm is not overkill if you mount it on t-slot extrusion. If they are not fixed tight against a datum edge the tracks will flex and bend. I think your best bet is to mount the ballscrew centrally. Single side drive will shorten your workable gantry length, unless you build a cantilevered machine in cast iron like a Biesse router.