[pippin88]

I've recently completed (sort of) an MDF build using ballscrews + supported linear rails from china. It works, which is not an inconsiderable achievement. However, it's not very square and rigidity is an issue. I feel it's best to just bite the bullet and start the frame again.

1100mm x 600mm x 300mm total length of rails
Aiming 950mm x 400mm x 150mm usable travel.

I'm planning to use 40x80 aluminium extrusion for the frame
Aluminium plate (~12mm) for gantry sides
I already have the 80x80 gantry bridge

Attached is an eDrawing (.easm) file of the design at present.





Things left to do:
Work out motor mounting for Y axis
Finalise frame - Will my frame be strong enough? Is it overkill? I note the FineLineAutomation FLA-100 has much fewer cross members in a slightly larger machine. The gaps between my cross members are 240mm
Decide whether it's worth moving from a piece of 40x80 with Z rails on side for Z axis to a plate with Z rails on front. This would probably buy me 50mm more usable Y axis but at the cost of rigidity in a number of planes.

Thoughts?

[Phife]

Are you sure that the Supported rail's mounting holes line up with your aluminum extrusion slots? I believe last time I checked the holes on the 20mm rails were 30mm apart.

Having the rails and extrusion compatible makes everything a lot easier to build.

[pippin88]

The mounting holes on the rails do not line up with the slots. I'm planning on drilling and tapping the extrusion. The mounting holes end up sitting nicely over the meaty part of the extrusion
This is what I have done on the Y bridge and it worked quite well. Bit of work though.

[JerryBurks]

The mounting holes on the rails do not line up with the slots. I'm planning on drilling and tapping the extrusion. The mounting holes end up sitting nicely over the meaty part of the extrusion
This is what I have done on the Y bridge and it worked quite well. Bit of work though.

That is what I did with M6 shallow head Allen bolts and it works well. 1/4" screws are already too fat and the thread will break through the extrusion wall. However, to get the rail precisely in position I machined a few clamping jigs to keep the rail precisely centered on the 1545 extrusions, then made pilot dimples with 6mm drill, drilled through with 5mm and tapped with M6 all with the rail in place. That made sure the position to be exact.

FWIW I think your gantry verticals are way too thin. You need some bracing (angle, U- or better tube/box) or it will flex sideways. If your gantry is not super rigid, the single centered lead screw will also make the whole thing twist/rack when machining at the table borders.

[Phife]

Any reason not to use extrustion that is compatible with the rails? like 60x60 or 60x120? Save you from having to drill and tap all those holes.

Also I agree, thicken up your gantry uprights or add some bracing.

[Shane Blacklord]

I have looked over your design an have a question how wide is th stock you are using for the sides of the gantry? it should be at least 1/2 inch. The bottom and top crossmembers should be larger. Too little and you will run into problems see Neo7's build on youtube

Regards
Shane

[RossMosh]

Any reason not to use extrustion that is compatible with the rails? like 60x60 or 60x120? Save you from having to drill and tap all those holes.

Also I agree, thicken up your gantry uprights or add some bracing.

This is the first thing I noticed as well. The 20mm rail has the hole spacing to work with metric extrusion.

[pippin88]

I'm planning 12mm alu gantry sides or thicker if I can get. I'll pre drill some holes for ribs to stiffen if required. I've seen plenty of builds with 1/2" alu sides with apparently no dramas.

A lot of design choices have to do with availability of extrusion in Australia.
From the best value of our local suppliers 60x60 is actually lighter per metre than 80x40 and costs more per kg of aluminium. Throw in the fact they don't make M5 spring nuts and using 30/60 series stuff didn't seem as attractive.
I haven't ruled it out though.

[pippin88]

Thinking heavily of using flat bar (80x25mm) for the long side pieces
Advantages:
1. Easier - drill straight through and screw into extrusion holes
2. Cheaper - cost per metre almost the same as 80x40 despite weighing 66% more. Throw in the fact I won't need fancy fasteners ($2.50x10 each side) to join the bits, just normal bolts.

I'm also adding a 3rd truck to each X (and probably Y axis)




The angle / setback of the side plates and Y gantry is just a rough guestimate at the moment.
Will be adding fins to gantry side to stiffen.

[Phife]

What is the advantage of adding a third truck to each axis? I cant see how its going to make anything better.. I would ditch that Idea IMO, more work, more parts to maintain, no benefit..

[PaulRowntree]

Thinking heavily of using flat bar (80x25mm) for the long side pieces
Advantages:
1. Easier - drill straight through and screw into extrusion holes
2. Cheaper - cost per metre almost the same as 80x40 despite weighing 66% more. Throw in the fact I won't need fancy fasteners ($2.50x10 each side) to join the bits, just normal bolts.

How does the stiffness wrt sideways deformation compare to the 80x40 pieces?

[Darylb]

How does the stiffness wrt sideways deformation compare to the 80x40 pieces?

80/20 1530 series extrusion bent about the weak axis is about twice as stiff as 80mmx25mm bar (0.483 in^4 vs 0.262 in^4)

EDIT: oops... i just realized you weren't comparing it to the 80/20 pieces many designs use for the uprights. To answer your question, the stiffness goes up with the height^3 therefore the stiffness is about 4.1 times greater therefore deflection will be reduced by a proportional amount.

[PaulRowntree]

80/20 1530 series extrusion bent about the weak axis is about twice as stiff as 80mmx25mm bar (0.483 in^4 vs 0.262 in^4)

EDIT: oops... i just realized you weren't comparing it to the 80/20 pieces many designs use for the uprights. To answer your question, the stiffness goes up with the height^3 therefore the stiffness is about 4.1 times greater therefore deflection will be reduced by a proportional amount.

Let's see how this works ... please correct me if I mix this up.
The 8020 site has Ix=0.48 in^ for the 3"x1.5" extrusion. A solid bar that is 25 mm high x 80 mm wide (so oriented to measure sideways deformation) would be 1" x 3.2". The moment for a solid rectangle is [(1)^3][(3.2)]/12=0.27 in^4. Is this right?
If I haven't gapped anywhere, the extrusion will be ~twice as stiff sideways wrt the solid bar. To calculate the actual deformation, you would have to input the geometry and the anticipated loads.
Cheers!

[Darylb]

Let's see how this works ... please correct me if I mix this up.
The 8020 site has Ix=0.48 in^ for the 3"x1.5" extrusion. A solid bar that is 25 mm high x 80 mm wide (so oriented to measure sideways deformation) would be 1" x 3.2". The moment for a solid rectangle is [(1)^3][(3.2)]/12=0.27 in^4. Is this right?
If I haven't gapped anywhere, the extrusion will be ~twice as stiff sideways wrt the solid bar. To calculate the actual deformation, you would have to input the geometry and the anticipated loads.
Cheers!

Yep, thats what I had. The formula for the moment of inertia of a rectangle is I=(1/12)*[Base Dimension]*[Height Dimension]^3

That 80/20 1530 is ~twice as stiff as the 80x25 bar and the 80x40 is ~twice as stiff as the 80/20 1530.

For stiffness:

80x40 > 80/20 1530 > 80x25

Knowing this perhaps OP might consider using a square section or something of the like to increased the section modulus he has to resist Y-axis deflection

[pippin88]

3rd truck helps distribute the load of the gantry.

I was thinking the cross 80x40s should provide plenty of resistance to y axis deflection.

My physics is far from amazing, but the long pieces and cross braces should form a torsion box in that plane. The biggest deformation problem will be downwards on the long (x aids), which shouldbe much the same with solid 80x25 vs 40x80 extrusion?

[Darylb]

3rd truck helps distribute the load of the gantry.

I was thinking the cross 80x40s should provide plenty of resistance to y axis deflection.

My physics is far from amazing, but the long pieces and cross braces should form a torsion box in that plane. The biggest deformation problem will be downwards on the long (x aids), which shouldbe much the same with solid 80x25 vs 40x80 extrusion?

40x80 extrusion bent in the strong axis (Like your x-axis is) has a moment of intertia of 1.81336 in ^4 versus 80x25 solid bar which has a moment of inertai of 2.5627 in^4. Though the solid bar is a bit stiffer, it wont be cheaper and in this case, I think it would be better to go with the extrusion as it is easy to mount stuff to it.

[pippin88]

Daryl, thanks for that calculation.
I can get a 4metre length of 80x25mm bar for $235. 40x80 extrusion costs $50.60 p/m in Aus + cutting fee, bringing it to almost the same price as bar. Then add the 10x connectors needed for the crossmembers, per side (=$25 per side) and it's significantly more expensive.
Only marginally easier to attach things to for my build (as supported rails need to drilled and tapped either way)

[pippin88]

After more pondering and looking around I've reached the conclusion that I really want a T-slot table.
The only reasonably priced extrusion in Aus I can find geared towards this is from Australis Engineering
I could fairly easily bolt that on top of the design above.
Total cost for frame above + T slot would be approx $600 without gantry sides
That got me back to thinking about using 30 series profile which has slots that match the hole spacing on my supported rails.
A frame made entirely out of 60x60 (running in the long axis, 1.2m long beams x 8 with no gaps between joined by ends plates +/- a strap or two underneath) comes to around $500 (again without gantry sides).

The 60x60 option would be pretty much bolt together.
Thoughts?

[Phife]

After more pondering and looking around I've reached the conclusion that I really want a T-slot table.
The only reasonably priced extrusion in Aus I can find geared towards this is from Australis Engineering
I could fairly easily bolt that on top of the design above.
Total cost for frame above + T slot would be approx $600 without gantry sides
That got me back to thinking about using 30 series profile which has slots that match the hole spacing on my supported rails.
A frame made entirely out of 60x60 (running in the long axis, 1.2m long beams x 8 with no gaps between joined by ends plates +/- a strap or two underneath) comes to around $500 (again without gantry sides).

The 60x60 option would be pretty much bolt together.
Thoughts?

To me its a no brainer, the time you will save by going to something that will just bolt together without any marking drilling and tapping will mean you will have a working machine much quicker will less room for mistakes. It will also allow you to make adjustments and add stuff on later.

[Benonymous]

Using two sizes of bar stock, you could make a C channel for your X axis sides like I have for my Y axis. It involves a lot of drilling and tapping but the results are pretty decent for not too much $$$. I used countersunk cap screws.