Scratch Carbon Fiber 18MM MD Build

[SkyDart]

Hi,

I'm new here and am just getting back into the hobby as an adult. I have several Estes builds and launches under my belt and am going for my first scratch-built rocket. While I want to go for my L1 I thought doing a LP build and getting some experience in building techniques and design would be beneficial. I have attached the open rocket sim for my design. It is an 18mm ID carbon fiber tube that I will form around a mandrel, nose cone that will be molded with a 3d printed mold and fins that will be done as a flat lay-up and cut to shape. Planning on doing a tip-to-tip layup on the fins for added strength and the parachute will be made from ripstop nylon and upholstery thread. The shock cord will be paracord. Everything will be built from scratch except for the ejection baffle that I just can't justify the added cost and design time for when I can buy it at Apogee Rockets for $5 + $10 shipping.

One question is my math for the airframe tube. Simple math says to get my wall thickness of 1.32mm with .22mm carbon fiber cloth I need 6 wraps. Building on an 18mm ID the circumference is 56.549mm and if I'm going to have 6 layers you multiply this by 6 to get 339.3 mm. However, each layer the diameter and thus the circumference gets bigger so if I cut the wrap to exactly 339.3 if will fall just shy of 6 wraps and have a slightly thinner spot. Is it worth doing the math layer by layer to find an exact length? I have a Cricut that I could use to cut accurate dimensions, but it would still be +/- .5 mm and then I still have error margins in my assembly process plus surely the epoxy adds a small bit of diameter (1/10ths of a mm ?) between the sheets that I don't know how to calculate. Is having a slightly thin or thick spot in the wall even an issue?

To build on that when doing the fin-to-fin layup I will add to the thickness of the tube. Should I use less fabric in the tube where the fin-to-fin will be so that the OD is consistent or just lay on top and have those areas slightly, ~.44mm, thicker?

I do have access to a laser cutter, 3 axis CNC mill and 3d printer that I will be using for this build.

Anyhow looking for feedback on the design and what issues I might run into with it before I purchase the mandrel and design the mold for the nose cone.

 

[bad_idea]

Welcome to TRF. Very interesting and challenging project. I'll be interested to watch this come together.

I've yet to lay up a complete tube, and I have yet to work with carbon cloth, but from working with fiberglass, I've found there is quite a lot of give and take in the fabric, to the point I wouldn't worry about dimensions down to the millimeter. From what I have seen of carbon cloths, the situation looks similar to me, but like I said, I have no direct experience there yet.

From tubes I've glassed with cloth (instead of sleeves which I've mostly used lately), I found trying to get a consistent thickness between the tube and tip-to-tip over the fins to be more trouble than it was worth (to me), but if I were to do it again, I'd wrap the tube (or mandrel in your case) with the number of layers I wanted under the tip-to-tip, install the fins and apply tip-to-tip, then wrap further layers around the tube ahead of the fins to match. It would still not be easy to get a good transition from the section before the find to the tip-to-tip section. There may be better ways of doing this.

Your tube questions may be best posed in the Techniques forum, where making tubes comes up more often than in the individual "power" fora like this one.

 

[cls]

Plan to make several tubes, try it different ways. Don't imagine you are going to roll one and get it perfect the first time.

Carbon can be difficult to wet out, to get the epoxy to saturate it.

And then, there's too much epoxy for optimum weight of the lay up. So most folks plan to squeeze out most of it. How do you plan to do that?

Also, outside wrapping and compression will make a nice smooth finish. You'd have to sand a ragged finish for a long time.

Last, my feeling is 18mm is too small for any fabric except fiberglass veil, like 0.8oz. I've never seen such light fabric in carbon fiber. if you just want to play with carbon, start with 54mm tubes.

 

[SolarYellow]

That wall thickness tube is going to be very heavy for an 18mm motor to lift and waaaay stronger than you need. I would not go much over 0.5mm thick, and that's only because that's as light as I can easily buy. It's still 0.020, and that's heavy duty for a cardboard tube in the size model rocket you're talking about. FWIW, I would just buy the tube. These are actually pretty nice:

https://www.aliexpress.us/item/2251832683673842.html

That will be essentially perfect, lighter, smoother, etc. than you can make. At that 19 ID/20 OD size, you can also use a Quest 20mm nose cone, if you want.

Doing composites on a very small scale like you're talking about and getting good results is actually more difficult than doing it on a somewhat larger size, IME. It's difficult to get the fibers to follow the tight radii you will need for it to look decent. I second the advice to start with something in the 2-inch range. A mandrel is more expensive, but maybe you can buy some tubing rather than solid.

You're obviously free to do what you want, but it seems to me you're making this so difficult you may never finish it, and far more expensive than it needs to be. If I was doing something like that rocket, I would do the tube linked above, a Quest nose cone, and make the fins from any of the many sources of flat carbon fiber plate on AX. Thickness of 0.5 mm should be plenty, especially if you plan to do tip to tip, but I'd probably just go 0.75mm or 1mm and epoxy them on with good fillets. There isn't a commercial motor made for which that won't be strong enough.

With the 19 ID/20 OD tube, you could also make a forward payload bay from Quest 20mm tubing with the Quest NC, and it will be big enough for an Eggfinder Mini. You need the cardboard to get the GPS signals in and the radio transmissions out.

 

[prfesser]

No No No No No No No No No

Personal Message me (the envelope, top right of your screen) with your mailing address and you will receive a 500mm long, 18mm i.d., 1mm wall carbon fiber tube I purchased online---that I'll never use. I just stuck a C6-5 in there and it's slightly loose, may require a few wraps of masking tape for a good fit. Cost with shipping is one 'thanks'.

 

[SkyDart]

Plan to make several tubes, try it different ways. Don't imagine you are going to roll one and get it perfect the first time.

Carbon can be difficult to wet out, to get the epoxy to saturate it.

And then, there's too much epoxy for optimum weight of the lay up. So most folks plan to squeeze out most of it. How do you plan to do that?

Also, outside wrapping and compression will make a nice smooth finish. You'd have to sand a ragged finish for a long time.

Last, my feeling is 18mm is too small for any fabric except fiberglass veil, like 0.8oz. I've never seen such light fabric in carbon fiber. if you just want to play with carbon, start with 54mm tubes.

Thanks for the input, a lot of it is I want to play with carbon fiber. I plan on squeezing the excess epoxy out with a mold release shrink tape. Not wanting to start with a 54mm tube as I don't have ranges that can support launches with that size engine that are frequently accessible. After receiving the carbon fiber cloth today I plan on moving forward with the design. I may make changes to the design; if I encounter challenges with the nose cone I may go with a 3d printed nosecone that will reduce performance but increase stability.

If I can master it at a small scale I can perfect it at larger scales. I currently have 6' of 12" carbon cloth that should provide me with multiple attempts at the body tube and fins

 

[SkyDart]

That wall thickness tube is going to be very heavy for an 18mm motor to lift and waaaay stronger than you need. I would not go much over 0.5mm thick, and that's only because that's as light as I can easily buy. It's still 0.020, and that's heavy duty for a cardboard tube in the size model rocket you're talking about. FWIW, I would just buy the tube. These are actually pretty nice:

https://www.aliexpress.us/item/2251832683673842.html

That will be essentially perfect, lighter, smoother, etc. than you can make. At that 19 ID/20 OD size, you can also use a Quest 20mm nose cone, if you want.

Doing composites on a very small scale like you're talking about and getting good results is actually more difficult than doing it on a somewhat larger size, IME. It's difficult to get the fibers to follow the tight radii you will need for it to look decent. I second the advice to start with something in the 2-inch range. A mandrel is more expensive, but maybe you can buy some tubing rather than solid.

You're obviously free to do what you want, but it seems to me you're making this so difficult you may never finish it, and far more expensive than it needs to be. If I was doing something like that rocket, I would do the tube linked above, a Quest nose cone, and make the fins from any of the many sources of flat carbon fiber plate on AX. Thickness of 0.5 mm should be plenty, especially if you plan to do tip to tip, but I'd probably just go 0.75mm or 1mm and epoxy them on with good fillets. There isn't a commercial motor made for which that won't be strong enough.

With the 19 ID/20 OD tube, you could also make a forward payload bay from Quest 20mm tubing with the Quest NC, and it will be big enough for an Eggfinder Mini. You need the cardboard to get the GPS signals in and the radio transmissions out.

Thanks for the input, after receiving the cloth I think you may be right, and I may take the wall thickness down some. I am looking to make the tube and fins stronger than needed so I can recover at higher speeds between 20-30ft/s and have a usable rocket. With the small ranges I have drift has been my enemy and having an overbuilt rocket that can withstand a higher recovery speed lets me get the altitude I want and still recover the rocket; before I get into dual deployment.

 

[SolarYellow]

If that's your logic, I'd go a different way. Buy a nice, long BT-50 rocket and an Eggtimer Quark and make it DD. I kinda came to that conclusion after a simple, 1300-ft whoosh-pop flight of my Cherokee-D sized "utility rocket" caught multiple thermals and drifted near-as-darnit a half mile before touching down. I had a visual on it, and it was on the ground exactly on the line I thought it would be on, but it was a long walk. Found someone else's Generic E2X that had been there a LONG time on the walk back.

I'm not intending to fly much at that field again without DD, even LPR stuff that goes over ~1000 ft. Don't necessarily need tracking, but need it to get back down.

 

[SkyDart]

Really appreciate the feedback guys, I'm stubborn and I like a challenge so we will see how this goes. I have done a couple of 2-inch-long test tubes and failed miserably; but I have learned from these failures. Learned some things like there is never enough mold release, you have to polish the aluminum tube, the edges unravel very easily, and saran wrap will not work as a compression wrap. Learning from these I have made some adjustments to my process and have my first attempt at a full-size tube rolled and curing. We'll see how it goes, I'm a little concerned about having enough compression particularly in the middle. Haven't found a source for the shrink tape locally and shipping is expensive so if this fails it will be a week to order something online and get it.


I have a piece of 1/4" balsa that I was planning on using to make the motor block and cap the nose cone, but I got to thinking and thought why not make the ejection baffles from it and just glue them directly to the inside of the tube, so I drew up what I was thinking in Onshape. The forward baffle will double as my shock cord mount. I have 1.5 calibers between the motor block and the first baffle but I'm not sure how much space to put between the baffles. Doing some research into that now.


If y'all are interested I will update as the build process progresses.

SolarYellow thinks for mentioning the Eggtimer Quark it seems like great economical option for getting into DD. Also, I lost my Generic EX2 to a tree testing a streamer recovery off of a C6-5 motor. The test on an A8-3 went great but when I switched to the C6-5 it weather cocked and flew several hundred feet upwind and recovered into a tree 30` up and feet from the edge of the tree line.

 

[SkyDart]

Little update, I have a usable body tube this morning. Had some issues getting it off the mandrel but I was able to salvage it. Just a small issue with a seam was left, I was able to grind that down


Did a small flat lay-up and hand cut some fins from it to get a better idea of what this is going to look like. They are just held in place with a couple of drops of super glue right now. I will replace them with CNC milled fins that are perfectly shaped once I get access to the CNC mill. And then just going with one layer Tip-to-Tip.




Once the rocket is done, I plan on putting a final coat of the epoxy resin on to smooth out the seam and any other imperfections and then polishing for a smooth low drag surface.

I'm a little concerned with how small the fins are but it sims out stable in open rocket even after adjusting masses to what has been measured and what is predicted for the parts to come. It is even bordering on over stable on the smaller engines.

The whole thing so far has been slightly challenging. Getting the body tube off was the hardest part. I underestimated how much thermal expansion and contraction played into releasing the carbon fiber from the aluminum mandrel and did it cold. Next tube I will be prewarming and placing in a warm oven so that it expands and makes removing the tube easier. Also shrink tape for compressing the tube and keeping it round, I used my mold release tape and it worked but there are some small <0.5mm deviations across the length of the tube.

Overall, this is coming together better than expected. The tube and fins are way stronger than needed, and the whole body with fins came out under weight at only 10g. I will be adding some weight with the fin to fin and the final coat of epoxy but right now I'm expecting a finished mass of ~30g with chute, nosecone, shock cord, baffle plates, motor block and shock cord mounts. Lift off weight on a C6-7 will be ~ 52.

Next update will come in 2-3 weeks, can't get into the makerspace until 5/24. Have all of the materials and designs ready to go though. Going to go with a 3d printed nose cone at first because it's easier and the weight will actually help performance.

 

[lr64]

Only 10 grams? With a 1.3 mm wall thickness, it's still probably strong enough to fill with sand and use for a sap or a mallet! I suspect a wall thickness of 0.2 mm in carbon would be more than adequate for a modestly proportioned rocket. Maybe even 0.1 mm. Carbon is stiff and strong, compared to cardboard. I'm sure you're learning a lot, though. It's possible to find carbon that's light enough for 0.2 mm wall or less. https://sollercompositesllc.com/product/3k-1k-carbon-fabric/
I have made several carbon fiber tail booms for RC gliders. I had a lot of trouble getting them off the mandrel, even though it was tempered. Maybe should have loosened them and then put back on before the post-cure.

If, by performance, you mean altitude, I think the winning strategy is probably to find a motor that starts with a lot of thrust to get good speed off the pad, and then burns for a long time at low thrust. In that case, the lighter the better. For motors that burn quickly, more weight may be optimal.

 

[SkyDart]

Only 10 grams? With a 1.3 mm wall thickness, it's still probably strong enough to fill with sand and use for a sap or a mallet! I suspect a wall thickness of 0.2 mm in carbon would be more than adequate for a modestly proportioned rocket. Maybe even 0.1 mm. Carbon is stiff and strong, compared to cardboard. I'm sure you're learning a lot, though. It's possible to find carbon that's light enough for 0.2 mm wall or less. https://sollercompositesllc.com/product/3k-1k-carbon-fabric/
I have made several carbon fiber tail booms for RC gliders. I had a lot of trouble getting them off the mandrel, even though it was tempered. Maybe should have loosened them and then put back on before the post-cure.

If, by performance, you mean altitude, I think the winning strategy is probably to find a motor that starts with a lot of thrust to get good speed off the pad, and then burns for a long time at low thrust. In that case, the lighter the better. For motors that burn quickly, more weight may be optimal.

I ended up reducing the wall thickness after I realized just how strong this stuff is. My cloth is 0.22mm and I was aiming for .66mm but tube came out closer to 0 .9mm though it didn't come out perfect so there are some spots ~0 .8mm and some closer to 1mm.

But yes, this thing is strong enough to be used as a mallet, I tested a piece I cut of the end and my 40lb 5-year-old can stand on it.

I am learning a lot. The cloth was cheap and easy to source off of amazon $30 for a 12"x5' roll with free next day shipping and the epoxy is just cheap craft epoxy. The point of this was to do something cheap and small to learn. I have enough materials to make another rock et or two at this same size or possible try a build using a 24mm or 29mm motor before I move onto building something for my L1. As much as I would like to do my L1 on carbon I think with the requirement of staying under 3k on an H motor means I am going with something that is low and slow.

By performance I did mean altitude although after realizing how close to Mach 1 it'll go on an Aerotech D24 I am interested to see if it is possible to do something that is light enough and strong enough to go transonic on an 18mm motor.

 

[lr64]

BTW, people usually refer to the weight of the fabric in ounces per square yard or grams per square meter. As you've discovered, thickness can vary depending on just how you lay it up. Also by the weave, and by the stiffness of the fibers. If I'm not mistaken, the ratio of the epoxy's weight to the reinforcement's weight can vary from something like 30:70 to the other way around, i.e. 70:30. On one end, you'd have unidirectional fiber laid up and vacuum-bagged by someone with skill, on the other end, regular weave wetted out with a paint brush 30 minutes before the end of work on Friday,. Numbers very approximate, and possibly optimistic for Friday.

Having a fiber content is a good thing, but sometimes it may require a bit more fabric, in order to retain a reasonable thickness. Otherwise, the part might end up much stronger in tension, but weaker in other ways.

 

[SkyDart]

I should have taken more build photos but here is some from the flat lay up for the fins

 

[SkyDart]

She is finished, came out a little heavier than wanted at 42.2g. I have updated open rocket with all the individual masses and my cg locations for the C6-7 and A8-3 motors match what I have measured in real life by balancing it. Rocket is still stable but not as stable as I was hoping for and flying the D24-10 might not work without replacing the fins with some slightly larger ones.

This rocket is way over built but I learned a lot building it. Had some issues installing the baffles and motor block. The first attempt failed my stress test of throwing the nosecone and the top baffle where the shock cord is attached came out with the nose cone. This led to me using the mandrel and hammering it through the BT to remove the other baffles, the tube didn't even think about crinkling or bending. Unless it zippers this thing will get lost before it gets damaged.

Should have held the ruler up for scale but the rocket with motor installed is 10.5 inches tall with a body tube OD diameter of ~20.2mm.
.

 

[lr64]

Are you SURE the ring in the nose cone is large enough. ;-)

 

[SkyDart]

Are you SURE the ring in the nose cone is large enough. ;-)

I'm certain it won't be a point of failure ;-D

 

[SkyDart]

Well guys, she flies. This will be the final update unless I try the D24-10 engine. Videos were too large to upload so here are the YouTube links
C6-7 Flight
Maiden Voyage