Best Aluminum Tube Shape Design for No Sag

You need to provide a LOT more information. Drawings would be helpful. 

I've taken photos of our existing and others tubes...how does one post these to site?

When you post there is a little green camera at the top. Press it and do the normal thing. Make sure they are not huge in size. You can then bring them to a scale of your desire.

Your opening post was a bit lack luster but a picture will tell a thousand words. See how you go, Ky.

Yep, a lack luster performance on my side. I have no idea how that box got there How come I can't see a box with a red cross in it? I hope I haven't confused the OP.

Here ya go, a red cross with lots of boxes. Hope the tubing will last, Ky.

I found out my Google Chrome browser doesn't fully support the Reply box toolbar with Camera icon so I had to switch to Internet Explorer. I'll post 4-5 pics later showing the different tubes. Thanks to all for your replies.

Look to nature grasshopper...
Del

Hey, it's a roller, so it has to be round. For maximum moment of inertia (and resistance to bending) place as much material as possible to the outside. Lo, a pipe or tube. If 2.50 inches is not an off-the-shelf item, 2-inch Sch 80 aluminum pipe would be close, and would meet the minimum i.d. requirement.

You need to provide better specifications than that.

Minimal sag is much too subjective a target to hit.

Generally, the more material in the cross section the better that material will resist deflection.

Also, aluminum comes in many different alloys and T6061 can be hardened for additional strength. Stainless steel might be another alternative.

Regardless, a material engineer can quickly tell you what you need to know if you better define your specifications.

Let me see if I understand the scenario. You have a round tube that can have elements in the ID running the length.

At first I would say you only need two elements across from each other to add stiffness in the one direction that carries the load. But since the tube can turn to different orientations the array of elements inside the tube need to be at least three, four or more so in all orientations you still have proper support.

Since this is likely an extruded aluminum tube you will need to play with diameter and stiffening elements combinations that your process allows. Some deflection will need to be tolerated. The configuration obviously needs to stay in the elastic deformation region. Make sure to include a significant factor of safety so when the roller is bumped it does not fold up.

You have two options as I see it. See if you can kludge a likely combination of this up in your shop and see how much it deflects with load. Is the 15 lbs distributed or point loaded somewhere on the tube. Don't forget to consider the weight of the tube itself.

The second and best option is this: Get an experienced Mechanical Engineer to look at this beam stress calculation. They will need to look at multiple configurations of the tube cross section that your process could manufacture. Based on the proposed cross section they will determine the section modulus and then the stress and resulting deflection of the tube.

That might sound a little complicated but to someone that regularly "stress guesses" it is a routine problem.

I would like to add one more thought.

I believe the tube needs to be able to support the load in any rotation orientation. If this is true a round tube with no internal stiffeners may be best since it is does not need the internal elements added or extruded into the tube. Also as suggested you may be able to use a tube that is allready a standard product.

This would make it very easy to experiment with to see how much a standard tube deflects. Also any Statics and Strength of Materials book has tables that gives you the section modulus of round elements of various diameters and wall thicknesses. This simplifies the calculations.

Just to be wacky, you could get a way with a thinner wall tube by filling it with a rigid foam (Polyeurathane?)
If you need the internal chanels you could put thin wall plastic tubes in before foam filling.
Of course standard off the shelf tube will be cheapest, but not so much fun.
Del

Blindman,

In order to minimize deflection, you need to maximize moment of inertia. If the tube has an OD fixed at 2.5", moment of inertia increases with increasing wall thickness. Fluting the exterior shape will decrease the moment of inertia, so a cylindrical shape is best.

I dont have much experience as a materials engineer, but if allowed, wouldn't grooves along the cylindrical shape help?

I would look into pre-stressing the tube with a small bar down the middle, like pre-stressed concrete bridge beams. Maybe the stress would make the tube more resistant to bending. Just have a nut on the ends you can tighten up to test the theory.

Extrusion some thing like this. The internal supports could be a plastic insert.

Likely the horizontal internal support is unnecessary.

Perhaps you could even eliminate the diagonal supports as well

As I understood it, these are roller blinds where the roller
(extrusion) rotates, wrapping (and unwrapping) the material.
That is, each internal support would be required. (on rotation)

While it may be possible to support the tube with only 2,
(at 90 degrees) with the intended spans of > 12 feet, then
I think they would likely all be needed.

The supports do not have to be of any great thickness, 0.5 -1mm
would be probably sufficient; and more effective (at 12 ft) than
any circumferential indents (or ridges.)

jt.

Just a shot in the dark.... (lift the blind please.)

In KY's picture, I see a peaked roof, two sides. If this roof is not going to be rolled up, how about just pre-bending the tubes upward, so when the weight is on them, they come down level. A common trick in all kinds of construction.

The OP calls for a minimum i.d. of 1.85 inches. Perhaps this is to accommodate some internal mechanism. If so, the various internal ribs or rigid foam will not work. Back to basics: a simple tube/pipe gives the best moment of inertia.

Tornado, I fully agree; a simple tube/pipe gives the best moment of inertia.

Extrusion wise, some more factors:

OP is better off with a simple tool. Multi segments move about, changing wall and flow. The die gains complexity, everything through to the pull tooling is more costly and "T" becomes inconsistent as does straightness with wall drift.

Assuming there is a race or groove, detail to engage the blind, he should duplicate this opposite, or if high torque is present, replicate the surface and volume effect, to balance the flow.

Inherently balanced flow is important if you want a consistent product.

Also, if extrusion is a new experience, consider specifying "visible face", as this means it will be handled and inspected as an architectural order, rather than a structural order. "straightness", "striations" from pickups, gouges and forklift impressions then don't go to be anodized over - as it is fully refundable by the extruder.

Kyzine

Playing with the extrusion section I wondered which would resist
deflection (sag) the most? Not being anyway knowledgeable on this
I first thought of:

however, why remove the cross members at all? e.g. (compared to my original drawing) The inner circle would act as a brace to the cross members.

Interesting this. Without experimentation, does anyone have calculations on
the different deflection (sag) properties? (It may be helpful to the OP.)

jt.

The second drawing reminds me of something I have seen, somewhere, just can't remember what. But I think it worked.

ps.. I meant to add (in the prev. post) to save material, the internal support
section need not be necessarily continuous. It may be possible (depending
on the external tube thickness) to have the sections only as "spacers" every
few cms (e.g. 20 cms) spaced evenly down the tube.

jt.

I think you need to find out about extruding - particularly tooling.

This may seem a bit abrasive - but totally ignoring all real data, experience and knowledge, previously posted is equally, if not more so.

Kyzine

I apologise if my ignorance of extrusions has been annoying.

The op asked and I offered, my thoughts based upon materials
I have actually seen and used. (including his existing sections.)

With no qualifications for extrusion, please accept my limited input
was solely based upon what I have seen, used, and even sold in
aluminium, and that no extrusion expertise was intended.

Qualified extruders should be able to provide a definitive answer.

jt.

Perhaps the roller tube in the old paper rolls for printing calculators? From an extrusion point of view, yes, the internal mandrels might drift, but aluminum is never straight as extruded anyway - aka dissimilar thickness walls don't necessarily affect the overall straightness once finished to dimension - so if you could ORIENT the final tube, fewer larger "I" beams to reduce deflection in one direction might be the ticket. Radiating spokes with or without a central hub insures that the part is strong(er) in any position - aka the "roller" comment.

One other thought if it is a roller - be prepared to replace what ever you come up more often than a static shape anyway - ANY deflection while rolling will end up being torsion applied to one or more points along the tube and the part will eventually fatigue there -

Just our two cents -

Skippy and Buzz

http://skippyandbuzz.blogspot.com/

Thanks all for your comments so far. Here are more details. The tube we are designing is for motorized/automated roller shades. These shades operate by rolling fabric on or off the tube. The tube requires a tubular motor inserted in one end and a gudgeon/idler inserted in the other opposite end. The ends are suspended by brackets and the motor is then switched to roll down for closed position or up for open position.

Tubes must be hollow to fit internal components but can have channels with the following details. Our motors are 46.5mm OD so we either insert the motor into a tube with ID to be min 47mm (see pic T-round) or we use a tube with an ID larger than 47mm with adapters to hold both motor and gudgeon in place (see pics L-round and L-fluted).

Our existing tube is style T-round OD=63mm ID=47mm wall thick=2mm weight=1.26kg/m Alloy=6061T5

I'm thinking to switch to a new design similar to L-round or L-fluted. I am leaning towards the simplest round design with little channels and no exterior fluting. I think this will give me the strongest design as well as keep material costs lower. I welcome your comments.

style- T-round

style L-round

style L-fluted

If all the od's of the tubes are the same, I favor the top one, it has the strongest looking inside. Simple test, put a 15lb weight around the center of each and see how much it deflects.

yes Mike....all OD and wall thickness are same.

I see the lower two include screw races at 120⁰, for end cap attachment?

I'm not sure why you have the extra fins adjacent to all the races, but the bottom race extras could serve to balance the die - (in surface friction and volume of the engagement race/slot) - if larger.

The fluting (in 3) I would think just adds tool cost. You'd need a "T" deep joggle to add any significant moment of inertia to the tube wall. And again 6 is not balanced for friction against the major detail, just against 3 sets of internal fins.

You look a bit light in overall "T" on 1 and 3, and heavy in 3. And too thick at fin "U" base (in 2 & 3).

If they are the same "T" - then what is "T"? 2.5 mm?

Kyzine

p.s. COSMOS is handy for doing simulation.