Filling Aluminium Tube Gantry

I'm starting to design a new router. I want it to be pretty heavy duty so I started doing some simulation in Fusion 360 on the moving gantry beams to start.

Comparing chunky aluminium profile extrusion, steel rectangle tube and aluminium tube with a granite core.. so far the stiffest result for the least weight is the aluminium rectangle tube with granite core. So I'm quite pleased with that.

Attachment 460604
Attachment 460606
Attachment 460608

I'm now looking in to the practical aspect of filling the tube and the best mixture etc.

Can someone please help me understand the differences between epoxy granite and polymer cement such as Polycrete - Polycrete is a Granite Strength Resin Repair Mortar for Concrete or Asphalt - Internal or External | Vuba

Are they the same thing? Polycrete has a resin and a hardener and then a 'powder' and I'm not sure what that powder is; granite powder?

If I went with Polycrete I presume I still need to add an aggregate such as AlOx or quartz pebbles?

If I'm best going with epoxy is something like Xencast X6 suitable? Xencast P6 Toughened Polyurethane Resin - Easy Composites

I'm also wondering about attaching rails. Do you think the 6.35mm wall of the tube is enough alone, possibly with over-length bolts that I can cast the fill around (and remove them?!), or do I need to add some kind of internal inserts? In the same vain of thought do I need to add something like bolts through the tube walls to aid adhesion of the fill, or is it going to stick well enough to the inside the the tube if I simply wash it down with some acid first to remove oxide?

Advice appreciated.

Also, I intend to cast the fill around a threaded rod that passes through the center of each beam and be tensioned during the cure. Then it will keep the fill under compression as well as give additional fixing for the gantry sides.

I do have a vacuum pump so I would de-gas the mixture before pouring and possibly also possibly pull a vacuum on the tube once I cap the ends.

Polymer concrete often refers to polyester resin. Polyester resin shrinks considerably during curing and generally is not a good choice.

The threaded rod down the middle is not needed for epoxy granite.

You would be better:
Thicker wall tube (less local vibrations and you need something to fix into)
Epoxy granite lining the inside of the tube
Hollow middle - you can use a piece of PVC / similar hollow tubing - material close to the neutral axis does not contribute to stiffness much.
Ends capped.


Epoxy does not stick to aluminium well. I'm not sure how you will do with this. Generally the best adhesion is done by sanding the alu with a course grit with epoxy on the surface. This roughs up the surface (mechanical bond) and removes the oxidised surface allowing the epoxy to bond (epoxy has to be on there to prevent oxygen contact and oxidation). You can then add more epoxy / stuff which will bond to the epoxy.

Thanks for your reply! Also thanks for the idea of leaving the centre hollow to save weight.

Can you expand at all on why epoxy granite would not benefit from being pre-loaded in compression?

I'm not sure how easily I could sand down the inside of a long tube 1.5m. However I've read that when using adhesive to fix aluminium it is common to etch with acid first (then rinse with DI water) for similar reasons. I wonder how long it takes for oxide to build up to a problematic level, would I get 30muinites to pour the fill?

This was the thickest wall ali tube I could find - 3" x 2" x 1/4" wall (76.2mm x 76.2mm 6.35mm wall). I could add a plate inside that is bolted through from the outside. Then cast the resin around that. Or just back the bolts with inserts and cast around it.

If I grease the bolt am I likely to get them out again?

P.S. I was only using the built in material library for 'granite' in fusion. If someone can give me reasonable data to model typical E/G mix that would be very helpful.
Attachment 460622

Hi Sash - The value of 55GPa for granite is low for monolithic granite but high for general epoxy granite (EG) that uses sand or basalt as its aggregate. Unless polycrete can supply its modulus I would not use it. The PU resin is 1/3 the stiffness of an epoxy resin and it can shrink so its not a good choice. Making filled tubes is a tough learning curve. Better off using thick tubes of steel or aluminium. The general reason for filling is to improve dampness. This can be achieved by using a dense foaming PU resin vs EG for fill. EG is very expensive once you do the costing especially to get to 55GPa stiffness as you have modelled. In reality you will get 30-35GPa this will change your results. Keep at it. Peter

re bolts - if you use mould wax, vaseline, or something like these - yes the bolts will release. Plus to etch aluminium its better to use caustic soda vs acids. Both need the metal to be rinsed very well after this as the reaction can continue well after a simple rinse resulting in it corroding after time... The surface will reoxidise in about 4 hours so be quick....

Hi Pete,

Thanks for the figures and advice.

I reduced the simulations Youngs Modulus to 35GPa and Tensile to 20Mpa based on a paper I found. It still comes out pretty good.

I also did a new steel simulation with dual 80mm tubes like the aluminium ones with e/g fill. I can only find 5mm or 8 mm wall thickness and these give either a lighter and weaker beam or a heavier and stiffer beam.

Considering my expected motor choice I'm aiming for a total gantry weight inc. spindle of around 60Kg so I'd probably need to opt. for the steel tube with 5mm wall thickness rather than 8mm.

So I think it leaves me choosing my enemies;

1) The Ali beam with E/G fill will have nice flat surfaces and also superior damping. However it costs more, needs me to mix and de-gas the filler as well as prep the tubing and bolts.
2) The steel tubing will provide a wall I can directly bolt in to, does not need any filling and is cheaper. However the surface will need to be face milled for the rails or levelled with epoxy and it has zero damping.

I'll have a think on it.

I kinda like the idea of trying out the E/G fill, but maybe I'm just making work for myself. The router frame would be steel tube anyway, possibly with sand for mass damping (so I can let it out if I need to move it).

Hi Sash - Degassing the EG is not worth it. Depending on how you mix the aggregate you will introduce heaps of air. When you degas you will only degas a small amount of the air near the free surface. The deep air can't get out due to the hydrostatic pressure and the torturous path to the top. I've played a bit with this and concluded its not worth it. Use a very thin epoxy for river tables or infusion and it will nearly look after itself... Machining the surfaces flat and true is the best path. People have mixed results with epoxy levelling and I still think its a poor bed for your rail. If your going to do it do it well. Peter

More thoughts - since you are familiar or comfortable with epoxy consider laminated aluminium. Its stiff and damp easy to work with. Say you wanted 10mm thick you can epoxy 3x3mm or 5x2mm togther and its really damp. Since aluminium is generally sold by weight buying "layers" is not usually much dearer.. So you can get your 1/4" wall tube and adhere a 1/4" skin to it to give you stiffness and thread thickness... The main issue with gluing aluminium is the moisture barrier on its surface. if you waft a heat gun or propane flame over it this will flash the water away and epoxy will stick really well.... Peter

You are right Pete, a double layer ali construction is a nice option! I've used constrained layer damping before and it is actually more effective at low frequencies than any materials intrinsic damping. I had thought to avoid it as I dont want to need to cut huge long pieces, but actually I can use pre extruded flat bar.

I had considered a box tube within a box tube with a filler 8nbetween but the gap is always quite large, like multiple millimeters at best. Simply fixing plate to the outside of an existing box section is a nice solution is increase stiffness and maintain a thin and effective damping layer with high compressive strength.

The same could be done in steel of course, but I'm not entirely confident I can face the steel myself given its length is more then my current machines maximum axis.

I'll do a few sims and figure out what is most workable in practice.

P.s. Do you mean degassing thenfilled mold is not effective, or even degassing the mix in a bucket before pouring doesn't work? I wouldn't have expected much to escape a near perfect vacuum sustained say for 30minutes, but if there are gaps remaining in the mix then air will fill them again once vacuum is released. Might work to have a bu ket that can degass and then vibrate (give itna few slams on the floor!?) before letting air back in?

Hi Sash - No point degassing then mixing in the aggregate you just put the air back in. If you degas the mix in a bucket you will only degas the top 10mm or so. The rest is under hydrostatic pressure and in contact with rocks etc. To degas such a mix you need to be able to move the mixture in the vacuum so you have "new" material coming into contact with the free surface... That's how commercial degassers work, they have a stirrer that keeps the material moving from bottom of container to top. Since you have a vac pump you can infuse the mix. Dry pack the mould, wrap in vac bag and place an inlet pipe in the middle. Then pull down to nice low vacuum and it will suck in the resin no worries until filled. Done a few test pieces like this and works well. Peter

Nice!

Attachment 460668

Hi Sash - That's 60N/um that's quite stiff but you have to add in the rest of the structure yet. Do you have a target static stiffness in mind? Re preload in compression... Concrete is very poor in tension so concrete beams and panels are preloaded on the compression side to "force" the structure more to the compression side to keep the concrete from going into tension once loaded. Generally epoxy is about 10x stronger then concrete so this tensile strength issue is not a problem. Plus machines being so stiff have very low stresses in them. If you have to preload the structure your probably way way off where it should be. Unless you get into tensegrity structures then they have to be in tension....Peter

Quote:

---

Originally Posted by peteeng

Hi Sash - That's 60N/um that's quite stiff but you have to add in the rest of the structure yet. Do you have a target static stiffness in mind? Re preload in compression... Concrete is very poor in tension so concrete beams and panels are preloaded on the compression side to "force" the structure more to the compression side to keep the concrete from going into tension once loaded. Generally epoxy is about 10x stronger then concrete so this tensile strength issue is not a problem. Plus machines being so stiff have very low stresses in them. If you have to preload the structure your probably way way off where it should be. Unless you get into tensegrity structures then they have to be in tension....Peter

---

OOpps - Should read prestressed (compressed) on the TENSION side! typing too fast

Prestress Concrete - Front Desk Architects

Looking at data for epoxy though, it does have better compressive strength than tension. So is it not still a benefit to bias in compression?

Or does it simply not matter within the elastic range?

Simon

Hi Sash - technically all materials have better compressive strength then tensile strength. If you dig into this you will find that there is no such thing as "compressive strength" materials fail by tension or shear. A compressive failure is an elastic instability problem or its how the test is done that produces the result. Strength does not matter in a CNC machine as the materials will never get stressed to near their failure limits. You need to understand the materials modulus better and if its tensile stiffness is different to its compressive stiffness. This also is described by its internal hysteresis (or damping) epoxy is plastic and is viscoelastic so its stiffness is dependent on its rate of change of volume. This is why it works as a damper in a constrained layer. It absorbs a lot of energy vs if that layer where a metal or unconstrained.. Gets complex but KISS.. Peter

another factor with pre-stressing is that you can make the structure asymmetric so it will deflect in one direction different to the opposing direction. This is usually not a good idea in a machine tool.... in a bridge where the load is always "down" this is a not a problem, and yes the machine will keep within the elastic limits of the materials by a large margin.

Compressive Strength Testing of Plastics

the comp strength for plastics is denoted at a certain strain such as 1% compression, 5% or 10%. So the material has not "failed" its just that it will continue to squish until flat then it will get apparently stronger! So if you can get hold of the load/deflection graph of the test this will give you a better idea of when the material started to squish vs a one off "comp strength" figure.... In your FE if you look at strain values vs stress this will teach you a bit about how far the material has actually stretched. There is no such thing as stress. Stress is a mathematical construct that makes calculating strain easier without using modulus and areas. Things in reality get stretched and compressed (strain) stress is a shorthand and sometimes misleading way to look at things...

Thanks Peter, that makes more sense for me now.

BTW my name is Simon ;)

I've started an over-all build diary and questions log . I'd really appreciate your input there if you have any points to add!

https://www.cnczone.com/forums/diy-c...17940-cnc.html

How about a square aluminum tube with a flat 1/8 or 1/4 cold rolled steel flat stock epoxied to the face?