Carbon Fiber tubes and Tube Nuts

The other day, we switched out some thick, heavy aluminum tubing for some carbon fiber tubes (cheap chinesium tubes from ebay)

For the aluminum tubes, we used Tube Nuts McMaster-Carr

They work excellently in aluminum. Super happy with the fit.

In Carbon Fiber, however, it was very difficult to press in the nut, even though it was at the looser end of the range. When we got them pressed in, they seemed to almost cut into the carbon fiber wall a bit. It looks to have scraped away some of the layers.

After they’re pressed in, they seem to work fine but I felt like I was going to break the tube, even though it was carbon fiber.

With that being said, is there a better way to attach a tube to a plane other than a tube connecting nut? I know in driveshafts they glue aluminum ends on then use those ends. I think that’s a bit impractical, unless 3d printed.

Any advice or ideas?

(Insert composites safety & chemical safety shpiel here, tl;dr don’t breath the dust or fumes and the splinters are really annoying so wear gloves)

Yeah, CF is hella directional (anisotropic) and like many composites it behaves differently under distributed vs point loads as well (I forget what this is called).

Drawing of a moment load reacted through a tube nut vs reacted through a plug:


The same moment load at the tube nut results in high point forces where the two steel rims contact the tube, vs a distributed load along the epoxy bond. This is higher magnitude force is also pointed out on an axis the tube isn’t designed to be strong in.

Radial point loads like the tips of the tube nut will tear out individual fibers, and weaken the overall structure significantly. If the tube shatters, it will shatter where you have a tube nut. That being said, your loads may be low enough that it doesn’t matter, particularly with decent structure design.

In particular, avoid bending loads that leave a single tube nut as the sole moment reaction, and then you’ll avoid the scenario I drew above. These are drawn as “capacities” rather than as reaction loads:


If you really want to use tube nuts, you could try putting a 1/8" of extra epoxy around the interior of the tube to give the tube nut more material to bite in to before it gets to the CF.

(If you need to put torque through the tube, like as an intake tube, you need to glue, whether interior or exterior… tube nuts don’t do good torque transfer)


In industry the standard connection method to carbon fiber is bonding with epoxy, often scotchweld 2216. This is generally enough. Some customers will require you to use “chicken fasteners” in addition, even if they weaken the joint. Chicken fasteners are mechanical fasteners used in addition to the joint bonding because some people are too “chicken” to trust bonding. This is primarily because a proper bonded joint is very process dependent. The mating surfaces must be carefully prepared and cleaned, and there are often specific curing temperatures. If any of those steps are incorrect, the joint will be way weaker, and there is no way to visually detect a bad bond until it fails.

So, working with carbon parts is not for the faint of heart. They can be a lighter, stiffer solution, but it often takes very careful design and fabrication to take advantage of the superior stiffness to weight properties of the material. If you do it wrong, you’ll be heavier, weaker or both.


What you hit on here is why on bicycles with carbon steerer tubes they use expansion plugs with a large surface area to grip instead of star nuts like they used on old aluminum steerers.

Look at these for an industry example of what you are trying to do.

Edit: I forgot to mention the part of the bike I’m talking about is preloading the headset bearings via compression through the steerer tube, which is a high load application similar to what you are doing.

1 Like

Yes, I’m very careful with carbon fiber. Not only is it bad for you, it’s annoyingly bad, which is worse. It makes you itch if you don’t wear enough gloves and you get the fibers on you. (Don’t ask how I know)

Anyways, I assumed glue was the best way to interface with CF. In this instance, it’s only holding the sides on the intake together, so no real hard loads on it

How do other teams use it / what do they use for the inserts? I could see mainly structural use out of it.

I know @Brandon_Holley and the 125 crew have made heavy use of round cf tube epoxied into printed holes.

I definitely would not recommend tube nuts inside CF round tube.

For our 2019 elevator we printed slugs on our Markforged that we then used epozy to install.

Depending on the joint we had some sideways holes going through the slug that we had to drill clearance holes in the side of the CF tube for. In these cases it was best to wrap the tube in blue painters tape before drilling, and def make sure you take your time and drill straight or you will fracture the tube.

In other joints we had a nut pocket on the slug that we populated with a nylock nut before installing the slug, so then we could thread into the end of the slug with a screw to mount the tube.

We weren’t using the screw-to-expand threaded inserts we have grown fond of now back in 2019, if we had then we likely would have used those instead of a nut pocket for a nylock nut. Heat-set inserts would work as well (we just prefer the screw-to-expand because there aren’t any tools required to install them, and they seem to hold just as strongly as the heat-set ones.

Overall its definitely not as easy as mounting aluminum tube, so if you really don’t need the weight savings I wouldn’t bother. CF tube definitely won’t handle the side loads as well as aluminum will so if the tubes could possibly be hit by other robots I would really think twice. The large amount we used on our 2019 robot was 100% driven by weight savings, which was only because of the triple climb stuff we had on there. If we hadn’t attempted that then we would have kept everything aluminum.


Quick answer here: A 3D printed plug and the super high tech solution of high melt hot glue. The two keys are: gullets in the plug for the glue to hold on to, large enough plugs to tackle the aforementioned moment loads for your application. you may need to experiment with plug/tube clearance for your particular glue, temp settings, and intended use case.

Hot glue (high melt, we’re talking the good stuff) has the added benefit of not being permanent vs epoxy, however the overhead of getting a proper glue gun may dissuade some teams from trying it.

Hot glue, 3d printed parts and VHB comprise a surprising portion of 2767 machines…

stares at 2019 end effector… sigh


This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.