The Mrs aint gonna be happy about this little project!

[r0bsk1]

Ok so here's my latest incarnation. Gantry is 2000mm wide. Rails are 25mm.

I've stacked 3 60x60x5 box sections on top of each other and will be welded together. The edges have a 10mm radius on them so will need several passes to secure them.....at least thats how they did in on Discovery's Extreme Engineering on a bridge build. Do you think i'll need a backing plate?

Each 60mm section will have 2 supports to prevent the beam falling forward or back. I've used 20mm ali plate for now but may weld box section cut-offs direct to the beam to form a 'T' on the ends instead.

A spinning ballnut mount is bolted to the back of the Y axis bearings.




Ta
Rob

[Jonathan]

From the drawing it looks like the plan is to mount the rails on the top and bottom of the welded box section? If so you'll have problems with how much the steel will distort due to the welding. The rail spacing needs to be very accurate (guessing probably <0.05mm) for them to run smoothly, which is hard to get even before welding. Perhaps consider bolting them to a backing plate instead, since this allows adjustment to get the rails parallel. Also you will get much less deflection on this Y-axis if the ballscrew is placed between the rails. Currently when a force is applied parallel to Y, the Y-axis bearings can deflect in the same direction since there is a long horizontal distance (parallel to X) between the cutter and ballnut/rails.

If you decide to mount the rails on aluminium plate (as in post #15), then yes it is best to use tooling plate, but realistically the tolerances of the bearings are not that good so you would be fine with normal aluminium plate, but definitely not flat bar. If you were using profile rails then tooling plate would be the best option.

[r0bsk1]

The rails are bolted to the box section, even in post #15. Earlier designs for a 1000mm gantry had them mounted on 90x45 ali extrusion which was bolted to a back plate.
I was hoping that tacking the sections together first and then do gradual passes would eliminate any heat distortion but if it's out of spec for the bearing before i even start, well that puts the mockers on that idea. Any idea how 90x45 ali extrusion would cope over 2m?

I would have liked to have the ballscrew mount on the Z axis upright so I would have to remove the middle box section and bolt to a 20mm back plate but then then is no room for the ballscrew mount unless the width of Y is 200mm+ to accomodate the mount as i feel it best to surround the Y bearings. Y ballscrew is currently 100mm off centre from the Y bearings.

[Jonathan]

but if it's out of spec for the bearing before i even start, well that puts the mockers on that idea. Any idea how 90x45 ali extrusion would cope over 2m?

I should have been clearer, it might be out too much even before welding, but you can get round that. Instead of welding three pieces together couldn't you start with one or two bigger sections, then bolt them to a back plate? If you have a fastener every 100mm or so you can use them to adjust and remove the bend enough to mount the rails. Two pieces of 90x45 aluminum extrusion to mount the rails on would be fine, or you could use aluminium box section.

I would have liked to have the ballscrew mount on the Z axis upright so I would have to remove the middle box section and bolt to a 20mm back plate but then then is no room for the ballscrew mount unless the width of Y is 200mm+ to accomodate the mount as i feel it best to surround the Y bearings. Y ballscrew is currently 100mm off centre from the Y bearings.

Spacing the Y-rails out 200mm+ to accommodate the ballnut mount would be good as in addition to making the axis much more rigid due to the ballnut being in the optimal position, it will also make make it stronger because the bearings are further apart. However you don't want the rails too far apart without increasing the spacing of the bearing blocks since it risks them skewing/racking. The ballnut mount could be made a fair bit shorter by moving the mounting holes and having only one available ratio without changing the belt.

[r0bsk1]

You set up index plates on the top and bottom of the box section which not only would help deal with small distortion of welding But also give you a rebated straight edge with which to create a master index edge for setting up the rails with. This also has the advantage of being able to do dis-assembly and certain types of maintenance much much easier. First you have to decide which edge is you master edge and work from that in the actual building. Though bolting works much better in most cases. This is why King bridges that are bolted together from many plates have been know to last over 80 yrs of active use.

Michael

That would be very tricky to do as the box section is 60mm and the rail base plates are 65mm.......unless i've completely misunderstood. I'd have to mount somthing to the side and offset by 2.5 mm. Tricky but doable.....

I should have been clearer, it might be out too much even before welding, but you can get round that. Instead of welding three pieces together couldn't you start with one or two bigger sections, then bolt them to a back plate? If you have a fastener every 100mm or so you can use them to adjust and remove the bend enough to mount the rails. Two pieces of 90x45 aluminum extrusion to mount the rails on would be fine, or you could use aluminium box section.

There's no way im going to be able to bend the box section. I'd expect the ali back plate to bend before the steel. However they do "look" straight.

Spacing the Y-rails out 200mm+ to accommodate the ballnut mount would be good as in addition to making the axis much more rigid due to the ballnut being in the optimal position, it will also make make it stronger because the bearings are further apart. However you don't want the rails too far apart without increasing the spacing of the bearing blocks since it risks them skewing/racking. The ballnut mount could be made a fair bit shorter by moving the mounting holes and having only one available ratio without changing the belt.

Right so that sounds like the ballnut mount is upright mounted against the Z upright and runs between the Y box section beams. I had i laying on it side poking out the back. The rail base plates would be approx 330mm apart if the mount is 200mm. What could you get the ballnut mount down to?

[r0bsk1]

Here's a quick mock up the ballscrew mount running between the Y rails. All uprights are approx 500mm




Back plate is 2190x330x20mm.......quoted.......wait for it........£300 +vat for ecocast!!!
5038 is £205

Back to the drawing board i guess

[Jonathan]

That would be very tricky to do as the box section is 60mm and the rail base plates are 65mm

The SBR25 rail base is 55mm wide unless I'm missing something?

There's no way im going to be able to bend the box section. I'd expect the ali back plate to bend before the steel.

We're only talking small fractions of a milimeter which doesn't require that much force. Both will bend a little bit...

What could you get the ballnut mount down to?

About 160mm if you're happy being limited to only using 1:1 or 1:2 ratio without changing the belt. Don't quote me on that though as I need to draw it first!

Back plate is 2190x330x20mm.......quoted.......wait for it........£300 +vat for ecocast!!!
5038 is £205

It would be fine with 1/2" thick plate, but that doesn't really change the conclusion as it's still expensive.

[Ricardoco]

From the drawing it looks like the plan is to mount the rails on the top and bottom of the welded box section? If so you'll have problems with how much the steel will distort due to the welding. The rail spacing needs to be very accurate (guessing probably <0.05mm) for them to run smoothly, which is hard to get even before welding. Perhaps consider bolting them to a backing plate instead, since this allows adjustment to get the rails parallel. Also you will get much less deflection on this Y-axis if the ballscrew is placed between the rails. Currently when a force is applied parallel to Y, the Y-axis bearings can deflect in the same direction since there is a long horizontal distance (parallel to X) between the cutter and ballnut/rails.

If you decide to mount the rails on aluminium plate (as in post #15), then yes it is best to use tooling plate, but realistically the tolerances of the bearings are not that good so you would be fine with normal aluminium plate, but definitely not flat bar. If you were using profile rails then tooling plate would be the best option.

With reference to the section marked in bold, you simply cannot make such a sweeping statement, ive just welded my complete 7x4 and i used my dual dial guages one the gantry was clamped and ready to weld and there is no diference in the distorsion after welding or certainly no more than there was before the welding. If you weld it in the correct fashion and you have used a jig to hold the pieces, then it is fine. I will admit that if you use a 2mil rod on an old stick welder crank up the amps and weld 2mm box together you will have problems but a modern mig used to tack the items together before finishing the welding of it turns out fine, im not saying Its better than Ali, just that if care is taken during the welding of steel box no more surfacing and leveling will be required than before you welded it. Obviosly when mounting the rails onto the gantry you will have to do some remedial leveling work but its just not difficult, I will show you how ive done mine in the next week or so and it certaily ticks the 0.05 accuracy box you mention with ease


Rick

[Jonathan]

Even before welding, it's questionable whether steel box section is accurate enough to mount round rails on opposite faces, so no matter how good the welding is it's not a good plan. It certainly wouldn't work for profile rails without further machining or shimming.

it certaily ticks the 0.05 accuracy box you mention with ease

How do you know?

[Ricardoco]

Even before welding, it's questionable whether steel box section is accurate enough to mount round rails on opposite faces, so no matter how good the welding is it's not a good plan. It certainly wouldn't work for profile rails without further machining or shimming.



How do you know?

Because ive measured it.