Hello,
From an older topic I know that the internal radii seems to vary between 3.6 and 3.8mm.  The outer radii varies between 4.8 and 5.2mm.  That has been measured by hand. Is this information still true today (any process improvement?). 
What about the tolerance of the 20 mm of the side of the tube ? Is it +/- 0.1 mm ? In my application the tube shall slide inside the groove of roller which looks like this one <https://www.aliexpress.com/item/10-56-31mm-belt-groove-H-type-injection-molding-machine-safety-door-roller-square-groove-D10/32816622292.html>. The groove profile shall fit the tube side profile as close as possible. 

See you talking about the pultruded box section tubes? If you are thinking about some sort of structural mechanism I would advice against using them in that manner with roller support. Unidirectional tubes are really easily damaged when it comes to force transmission points. For very low loads it might work well.

We have not sourced any different specification box sections, so the existing tolerances have remained the same.

From a structural perspective, okemont is right in that you need to take care in terms of loading up these in a structure due to the construction  of the tube and the limiting factors that can effect pultruded  section.


Warren Penalver
Easy Composites / Carbon Mods - Technical Support Assistant

Thank you for your answer. 

Here is attached the force diagram applied along a 800 mm long square tube. 
F1 is a compression stress of 220N pic load (nominal value is at 100N). The four rollers are placed at forces reaction F2 and F3 which generate low bending stress force at ~2N. As the tube is sliding within the 4 rollers, the force transmission points over the tube change continiously. 
 
Do you still think that the Carbon Fibre Square Box Section 20mm is not suited to my application ?

 

First of all, your diagram isn't  in balance, as the leftward force isn't equal to the rightward
besides that, if the diagram is accurate, this might work, as 220N isn't that great force for a 20mm tube. But usually there is some sort of radial forces, especially when the axial force is dynamic (you are speaking of a force peak).
The problem is, that pultruded tubes have no fibres in a direction to take the sheer forces under radial load. As long as you are shure that there is no load perpendicular to your tube axis, this might work indeed.

I think that I will try to make it work. 
You have not answered to one of my initial questions. What about the tolerance of the 20 mm of the side of the tube ? pultrusion process is a kind of molding process ? It should be accurate at least +/- 0.1 mm ?    

Good Afternoon,

Process improvements have occurred in the industry but there are still some difficulties that will cause variance in tube measurements. A good example is that cured epoxy continues to harden and cure several days after the product has been baked and removed from its tool, causing the part to shrink slightly. Certain manufacturing methods and tube types are better than others (see our available product here, but there are others available in the marketplace: https://dragonplate.com/ecart/categories.asp?cID=304). For example, the common roll wrapped tubing involves wrapping sheets of material around a mandrel and compacting it during the cure cycle. The ID is closely controlled as this is the tooled surface, but any imperfections in the raw material thickness, film compaction, or resin content has a major impact on the final OD. The OD naturally also has ridges from the compaction process. It is because of this tolerance stack up and ridges on the outside that the OD has tolerances as wide as +/- 0.02”. Square tubes are a special challenge! To keep the sides flat, one would use call plates (otherwise the outer shape gets very round). Squeezing the sides flat tends to cause the thickness in the corners to increase, resulting in imperfect wall thickness and slightly bulging corners. Most of us in the industry have ways of mitigating the worst of this, some of which are proprietary, but it is extremely common.

For your application, I would recommend a roll wrapped tube. I would like to say pultruded tubing, but I don’t know of any manufacturer making square tubes that large. Most manufacturers in the US use imperial units and measure by the ID so you’ve got a challenge ahead. At DragonPlate, we have a 0.75” ID square tube mandrel and two wraps of twill material should get you close to a 20” OD. Something we could try if you’d be interested, but I’d look around the market place and see if you can find any stock items first!

We hope this helps!!

I would like to give it a try to the pultruded box section tubes even if it is not the stongest. I checked the twisting moment and made some crush tests on one sample of hobbyking Carbon Fiber Square Tube 10 X 10 X 600 mm which seems to be pultruded. It seems to be strong enough for my appplication.

Can you please measure the distribution of the side length 20 mm on several samples from your stock ? I need to draw a suited profile of the rollers for 3D printer   

Thank you in advance for your help