What's new
What's new

Machining thicker carbon fiber ring shapes


Sep 30, 2006
SE PA, Philly
I am asking about the best (cheapest) way to make carbon fiber ring shapes with non-standard cross-section, of about 1.5-4 inches in diameter.

So I am thinking of getting back into model rocketry after many decades. The big boy edition is called High-Power Rocketry. As in 98mm diameter engines generating 500lb of thrust or more. At least one amateur built a craft that made it to space (above the Karman line, about 100,000 meters). So significant stress on parts like fins and stuff. To be clear, my first foray back into the mix will be a 3 lb rocket with only about 20-40lb of thrust which will reach about 300 feet. Once tested I can use more powerful motors and will get to several thousand feet with the same rocket. If you're interested, there's a pretty sharp young gal, Xyla Foxlin, an engineer who does a great job showing how she made her latest rocket (it hit mach 2.2 and 23,000 feet). It's here.

One way that rocket builders address the large stresses is to use high strength materials like fiberglass and carbon fiber. Sometimes a composite: you start with a paper tube, then lay CF sleeves over it with epoxy. Also, where there are joints, everything is filleted with epoxy, sometimes with additives like cut carbon fibers. They fillet EVERYTHING, sometimes laying on CF cloth patches and gluing to the rocket tube to reinforce fillets. I was thinking about an alternate way to strengthen rockets and have a few ideas. They involve making pretty thick disks and then machining away a lot of stuff. Could be a PITA. One idea is to "pre-make" a fillet by casting. So for a 99mm id tube, with a CF bulkhead of 3mm CF, one might make a 99mm OD ring in 3/8 inch plate, cut it to (for example) 3/8 length, and then turn an internal concavity. This would epoxy into the rocket tube against the bulkhead. Diagram below. Some questions:

1) The material costs alone for cutting each ring out of sheet for (for example) a 4" diameter rocket would be 25-60 bucks. Are there any commerical products that have the size (e.g. 4 inch od) and the bulk (3/8 square cross section) that could be repurposed with machining? Or would these be as expensive as making from sheet? Or maybe (since a single engine for these things ranges from 40 to several 1000 bucks) I should just shut up and make my expensive rings?
2) Are premade fillet rings a thing?
3) Another idea would require a 99mm od tube about a 3/4 inch long and 5/16 in wall thickness. This would be machined considerably into a fairly unusual shape. Any ideas on how to score something like this?

One problem with CF is that its so darned strong that stuff isn't made in thick sections, and for my purposes I need thick sections (for bigger gluing surface area). Another is the cost. So I may have to try this in fiberglass. Heavier, about 40% less strength, and also a PITA to machine. But possibly affordable.

I guess one other option would be to machine the rings out of (for example) titanium.


I may be misinterpreting what you are saying, but increasing your entire cross section just to gain glue surface area does not sound logical.
Does a machined fillet into a reinforced fiber material behave the same way as one where the fiber follows the contour? My limited experience making CFRP parts would say no, but I've never delved into it from a highly technical perspective.
Both replies above are interesting, thanks guys. Some of the tubes people use are cardboard. By itself sometimes or often with a composite wrap. But if we have (for example) a carbon fiber tube, the substrate is stronger. That is, it doesn't matter what the psi shear stress rating of the glue is, if the paper tube is what will fail. So if one uses cardboard tubes, one must have a larger glue area. The other thing that came to mind was that the "fillet" shape is probably unnecessary. For a CF fitting, the cross section can be an "L" shape. Which suggests that if I can get a CF tube that nests nicely I could cut a section and turn a bulkhead from sheet and glue the two together, which should be fine for compression.

Keep in mind that these rockets - at least the one I am dreaming about making - can experience 40G's or so.

That's another thing that I was thinking about - compression vs tension. Some super high-end epoxies have compressive strength of about 24000 psi, but with tensile strength 12000psi (way better than standard JBWeld, but not near steel, Ti, or CF. The point is, the fillets shown are all compression. I would still have to turn a bulkhead, and probably would cut a small rabbet to allow easy assembly. Then I could put a small amount of CF clothe round the fitting and apply a epoxy fitting. This would make final assembly a LOT cleaner. Glue in a couple of CF assemblies rather than trying to form fillets deep in the tube, using a popsicle stick.

Does it have to be CF? No. But other materials do not have the strength to weight ratio that CF does. And if one has spent $600 bucks on a 6" CF body tube then making the other parts of CF seems to make sense. The CF tube weighs 2.6 kg, while a fiberglass tube weighs 3.44 kg. There is a "Super Cardboard" that is stronger than normal Kraft paper tubes, and that weighs 1.5 kg but is much weaker. Anyway, I think that a small ring of 4mm CF, plus a small length of nexting CF tube, may be the cheapest way to do the compression joints.

Still, if someone makes CF fillet rings, that would be cool.
The strength of CF is in the fibers, not the epoxy. CF has very low strength in the direction the fibers aren't running as the epoxy is carrying all the load in that direction. I don't know that I really understand what you are trying to do, but machining something from a big block of CF is not a good idea, it will have no strength in the direction perindicular to the weave and you won't be abel to tailor the layup to your needs. Why not just layup the CF into basically the shape you want to begin with?

As an aside, I see you are in Philly? Where around here can you launch high power rockets? I thought all the amateur launch sites were out in the desert.
Jaguar, Your point is a good one. As I said above, a lot of the high power rockets are made with the motor tube and body tube "filleted" together, often with pure epoxy (!) but sometimes with an additive (colloidal silica or calcium metasilicate, which are, roughly West systems 406 and 404 fillers) or with carbon fiber or glass staple. These are not the super strong carbon fiber material that people think. Another way of strengthening is to use CF fabric, epoxied... well, pretty much everywhere. To the body tube, to the fins, in the fillets for the fins. This has the strength of CF fabric but its not prestressed (sometimes one stretches a sleeve over the tube in strong tension, but not always), and so the material isn't even good for low tensile stresses - the CF straightens while the epoxy stretches and you get poor properties. I think that you're right. As I mentioned above, my search now is for a tube that just fits inside at CF body tube (e.g 99mm). And I'll bite the bullet and buy some 1/4 or 3/16 or 4mm plate and turn a disk and add a circumferential rabbet. I can then epoxy it and glue in a little fillet outside of the rocket. Then I can install a known-good joint cleanly and reliably. BTW, there is another idea I have that's not shown. I think that for that I'll just have to glue up two pieces of 1/4 CF plate. Then I have to machine it a bit. The plate has fibers stretched across different directions and I'm hopeing it will be strong.

For my initial tests, I'll use a model rocket (medium power, unregulated) engine for about 300 feet of altitude (for a 6 foot tall, 4 inch diameter rocket). I can launch it near my house. For the Level 1 Certification test, I'll use a more powerful "H" motor and expect to see about 1000' of altitude. Maybe more, depending upon the engine chosen.

One aspect of parachute recovery is that if the rocket goes up a mile, and descends at 2 mph, while the wind is 15 mph, the thing can end up 7 miles away. So here in the east, "low and slow" is good. In CA and AZ, where there are deserts and dry lakebeds extending 100s of miles, you get folks trying for 20 or 30 thousand feet. There is also "dual deploy' in which a small drogue chute is ejected first and the rocket descends really fast. Near the ground, sensors initiate a second propellant charge that blows another, bigger chute out. This keeps you near the launchpad. Probably the place I will do my test is in Maryland.
I don’t have a lot of experience of carbon fibre but did see rings being made for some military app once, they basically wound fibre into a groove in a form while applying resin, nothing fancy, it was a lathe with a vfd and coil of fibres, seemed to work.
The form looked like pu water pipe or gas pipe.
I would guess stronger than carving sheet?
Any reason you don't just use Aluminum? For tubes and flat plates CF is great, but for fittings Al is likely just as good if not better.
boslab, I was thinking of doing just what you were suggesting. Glue the fins to the motor (inner tube) but have a "landing" on the forward part of each fin that is of a diameter less than the body tube. Then just wrap epoxied CF tow around the three fins on that "landing". This is probably the optimal situation: the CF fibers would be in tension which is exactly what I want.

jaguar, I was also thinking about aluminum. As my imagination is cheap (implementing my imagination can be expensive) I was thinking of the same "landing" idea as above, but using three aluminum flat struts, with two screws on each vertex. Too complex. I think that an Alum ring, in the shape of a large fender washer, supporting the inside of a section of CF tube, would be ideal. I'm going to play around with this and make a drawing I think. I mean, we're already into composite materials, why not make a CF aluminum composite? This allows me to use Al's compressive strength (not a CF strong point!).

Then again, I'm an engineer. I may be overthinking this...
Why not look at reinforced thermoset plastic? We call it Tufnol in the UK and there is a company of that name that produces several grades and shapes including tube.
The right grade and weave should do the job, it is quite abrasive to machine and best worked with diamond tooling
If it were my project, I would first look into machined 7075 rings. They could be filleted without too much trouble, and then lightened with pockets. With your estimated 30-40# of thrust, I think you could machine away most of the web between the OD and ID fillets and still be way strong enough. If/when you increase the thrust, just decrease the size of the pockets.


Study how composite boats are built, foam core sandwich and tabbing.
Use the carbon’s strength to your advantage and mold parts that move the load path away from your joints.

Also to note, normal west systems resin is too brittle for secondary bonds. You’ll want to use a rubber toughened epoxy.
How hot does that motor tube get?

Instead of supporting it with two rings why not cut a full length tube from H80 (or denser) Divinycell pvc foam? Or make two rings say 1" thick? Would be stronger, lighter and cheaper.

Another solution would be to use a thin (1/16"?) G10 or similar ring at the upper end of the motor tube, turn the thing upside down and fill the void with two part urethane foam. Let it overflow out the top (bottom of the rocket) and trim.

Either way you would have to cut a slot for your fins (mill?) which would make them stronger too.

I've been vacuum bagging composites and working with high density foam for 30+ years for my R/C boat racing hobby.
Last edited:
Just use aluminum. If you're worried about weight and gluing area, face groove the aluminum.


You're only talking 40 pounds of force spread across two rings. Your glue area is something like a square inch with a 1/8" thick bulkhead, so the shear stress on the seam is something like 20psi. That's nothing. Increasing the thickness to 3/8 would triple your glue area, bu there's not much practical difference between 6psi and 20psi on an epoxy glue joint.

You'd want fillets anyway. The filler converts the stress from being in shear to being in tension or compression. FIllets are cheap, fillets are easy. FIllets don't have stress concentrations, which matters in a high vibration environment.

I know the temptation, but get some experience with doing things the usual way before you decide to go out on a limb redesigning stuff. High power model rocketry is a pretty well established field with very easily definable forces. Try what works first.
I was highly into this stuff decades ago as a kid & young teen. Was always jealous of the adults that could purchase the high-powered stuff & travel to the desert!

I think you're missing half the reason for fillets in this method of construction: Yes if everything is smooth, stress risers are minimized; but even more important: The act of laying in the epoxy / glue fillet ensures really good "wetting" of the joint as the glue gets smooshed into place. Machining a fillet and sliding it into the tube is going to shear off your glue line -- getting that exactly right is more work / more difficult than the wet fillet.