98mm Min-Diameter Composite Mach 3 Build

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I have a part sitting on my desk which is from an SLS machine which I can at least wrap my head around being stiff enough to use for a tool. I believe it's aluminum+nylon. For my personal projects, I probably still wouldn't go this route but my requirements are likely different from yours.

Another couple comments on your mold:
  • I would not print male keys into mold halves. They're going to be a pain to finish and finish around, and look like they have the potential to break. It's also going to be really hard to both put mold halves together and take them apart because the mold will only move along the axis of your keys. I also wouldn't use a square profile, it will be hard to finish. For tools which are directly cut, I've used both the marble method @G_T shows pictures of in my thread and had holes through each half which pins push through. Both have been successful.
  • Your keys are getting kinda close to your part edge. Epoxy tends to squeeze into these areas and stick to your keys if you don't have a plan to deal with it.
  • If you 3d print your mold, do you have a plan to keep your nosecone shoulder feature sharp while you're finishing and polishing? I omitted this altogether and built the nose and lower tube as a single part.
I guess you could say this is a learning experience for me. I'm new to the world of DIY composites and this is our group's first time manufacturing nose cones, so our designs are pretty fluid right now. I'll definitely be updating the molds, thank you.

We're still talking through the specifics of the shoulder, but I'll propose your method. I need to do a little bit more research on different techniques to solve this problem.

I really appreciate all the input, and thanks for being patient with some of the flaws we're working through. There's a reason trial and error is the gateway to opportunity!
 
I've done this flight (custom CF rocket with M2245) and it folded up right around Mach 3. I rolled my own airframe (4 layers 2x2 200gsm twill) and had a bare (rolled CF airframe only) section between the forward closure and the aft end of my AV bay. That's where the vehicle folded. I was able to recover the NC and AV bay but the rest of the rocket (basically everything up to the top of the casing) just kept scootin.
I used a COTS FWFG nose cone for this and it survived going sideways at nearly M3. It strained some of the filaments but it didn't pop. A nominal flight with a COTS FWFG NC should be fine. No need for reinforcement, ablative, etc. You'll only just kiss M3 most likely and then decelerate quite quickly.
What would you have done differently to strengthen the airframe that failed?
 
We're also experimenting with compression molds. One of the ideas that are floating around is whether or not we should make forged carbon fins. The mold would allow us to give an exact profile to the fins that would otherwise be difficult while machining G10. (The machine shop supervisors here also told us that they discourage the use of G10 b/c it wears down their tools so fast) We already planned on doing a tip-to-tip carbon fiber layup once the fins are on the airframe, as we've had success with their strength on our larger competition rocket.

Here's the draft of that mold. Probably needs some tweaks too.
Fin Mold2.png
 
What would you have done differently to strengthen the airframe that failed?
I'm going to rebuild this offseason and give it another try when I can get back over to Queensland. Got the motor and hardware already sorted. My plan this time is to have some CF inside the airframe below the AV bay to effectively make the wall thicker at that weak point. Everywhere else should be fine and once I stiffen that one weak point I expect it'll hold up to Mach 3.
 
Started a little experiment today and 3D printed a compression mold for the fins. This test was using shredded fiberglass, and if the part turns out decent we'll move on to carbon fiber. We'll crack open this mold tomorrow, as well as start a fiberglass layup to prototype our nose cone.

Mold.JPGMold3.JPGMold2.jpg
 
Man, I'll be happy to survive 3. I'm finalizing the design this week and then we'll be getting to work. It's upwards of 9ft. long... 🙃
 
98mm diameter and ~9 feet long. We're using an N-motor (we had one lying around already) and dual-deploy chute setup, so that combination alone requires upwards of six feet. This isn't supposed to be simple. Our whole purpose is to challenge ourselves so that we get better at building our actual competition rocket for Spaceport.
 
Last edited:
A little update for everyone.

Fin Compression Testing

I last showed the start of a fiberglass compression mold that we were testing for our fins. It turns out that we didn't shred the material small enough, and paired with the fact that we had a little bit less resin than we probably should've used, the edges of the fins were almost entirely resin. The fibers didn't get compressed and fill the shape as well as we hoped.

101622.6.jpg

We ended up redesigning the two mold halves and printing them with much smaller layer heights (0.8mm) to help minimize the visible steps in the curve. I sanded them down way more than previously and almost entirely eliminated the transition between the visible layer lines.

101622.9.JPG

This mold (pre-sanding picture above) is currently curing with another test piece. Should turn out a lot better with the smaller fibers we used this time around.



Nose Cone Molds

I redesigned the molds with some of everyone's suggestions in mind, and this is what the finished (53% scale) product looked like. They were printed in PLA on a Raise3D Pro3, but the full-size version will be printed out of PCTG on an Essentium HSE 280i HT. Also, we decided to eliminate the shoulder from the NC design, as we will be using the e-bay as the coupler to the airframe.

101622.7.PNG101622.5.JPG101622.8.JPG

Again, I sanded down the molds to eliminate any noticeable transition between the layers. We pre-cut four layers of tight-weave fiberglass for each half then squeegeed resin into the cloth before laying them inside, trimming excess as we went.

101622.13.jpg101622.14.JPG

I applied a layer of peel-ply over the glass, and then three layers of breather material for our DIY vacuum bag. We plan on bagging every part we make.

101622.15.jpg
Ran out of attachment space so to be continued...
 
Last edited:
The NC came out a lot cleaner than we were expecting, but it still needs a few improvements. We only used four layers of the tight-weave fabric, and the pieces came out WAY thinner than expected. We're doing another round with the mold right now with eight layers, but I have a feeling we'll need to do more for a final piece. Still trying to figure out exactly how we want to join the two halves.

101622.23.JPG101622.21.JPG101622.2.JPG
101622.3.JPG101622.25.JPG



Body Tubes

I thought I would try a new (at least to me) process for making body tubes. We've had some issues in the past with clean separation from the mandrels (Blue Tube), so I tried shrink-wrapping a thick plastic wrap around the tubes after waxing the mandrel itself.

101622.10.JPG

I then waxed the exterior of the plastic using Partall Paste and sprayed the corresponding release agent film. Not only did I use the same tight-weave fiberglass as the NC, but I also only used four layers of it so the tube came out way too thin to be usable. I also made a vacuum bag for the tube.
101622.11.JPG

101622.12.JPG
101622.16.jpg
 
Last edited:
101622.19.JPG101622.20.JPG

As I said, the test piece was way too thin to be usable, so we'll use double the number of layers the next time around. However, the new shrink-wrap worked flawlessly and separated easily.

In addition to this high-speed rocket, we're also making some standard L1 certification rockets, so we used the same shrink-wrap technique to make a full-length body tube. Used four wraps of standard fiberglass and then two wraps of tight-weave fabric. Vacuum bagged as well, with the second NC test pieces with it.

101622.24.jpg

We'll break open these molds tomorrow and see if our changes made a difference, so I'll post an update with that.
 
Also thought I would share our (almost) finished design. We're going to be using a Cesaroni N4100. Going to do some CFD on Ansys to verify everything is good, and once we validate all of our test pieces we'll get to work making the finished product out of carbon fiber.
V4.png
The body tube length isn't exact, but this is our best approximation of the packed chute and shock cord length. We laid it out with a 98mm body tube and this was the minimum length needed for the chutes, but we might end up switching to smaller chutes, or even single-deploy which is unlikely.
 
Nice job! A couple thoughts:
  • nosecone - don't treat the seamline/bond between nosecone halves as an afterthought. This is the hardest part to get right on these split molds. I've never attempted to post bond mold halves together. One time I folded one of my nosecones with a subpar seam at about mach 2.5; the rocket kept going up for an otherwise nominal flight and I was able to find all the pieces of my nosecone.
  • tubing - pay close attention to both the tolerance of your mandrel (OD should not vary along length) and the length of the tubes you're making (short tubes are much more forgiving than long tubes). I've stuck tubes on mandrels from both these issues. I'm also a big fan of aluminum mandrels because of their thermal coefficient. I've been able to unstick stuck tubes by putting them in a freezer, putting ice water inside them, and (in a real act of desperation) putting acetone in dry ice in one.
  • EDIT: Stability - presumably your Cg will continue to move around, but your stability margin looks quite low.
 
We're also experimenting with compression molds. One of the ideas that are floating around is whether or not we should make forged carbon fins. The mold would allow us to give an exact profile to the fins that would otherwise be difficult while machining G10. (The machine shop supervisors here also told us that they discourage the use of G10 b/c it wears down their tools so fast) We already planned on doing a tip-to-tip carbon fiber layup once the fins are on the airframe, as we've had success with their strength on our larger competition rocket.

Here's the draft of that mold. Probably needs some tweaks too.
View attachment 540462

Just FYI there's a variation of this technique. It creates a sandwich structure.
I don't have much info on the exact process but these images should be enough to reproduce the process. The resulting fin is designed to be bolted to the airframe through the centering rings. I suppose you could get rid of the bottom section and directly bond it.
Expanding epoxy foam is spread on the wet layup, and the mold is closed. The expanding epoxy foam is left to expand: This creates pressure on the layup while it's curing. In later revisions, a recess was made in the mold on the leading edge to accept a thin aluminum leading edge cover.
These types of fins are used on the SERA 1, 2, 3, and 4 rockets for the PERSEUS project. Of course, you'll have to make the fins thicker than usual. Maybe the reduction of mass is enough to counter the increase in drag... idk...
Personally, I don't think it is worth the trouble on rockets <4" in diameter. But on larger ones...
1666715894181.png1666716165733.png1666716719165.png
 
@robopup,
  • We have three small-scale cones that were post-bonded together, and as you mentioned, we aren't comfortable with their strength—currently making a full-size NC prototype but bonding correctly.
  • Although we don't have full-size aluminum mandrels, we have brand-new cardboard ones that have served us well in the past. We haven't had issues (yet) with our release agent applications or separation process. Since my last post, we've made three six-inch-long 98mm diameter CF body tubes as test pieces to run material analysis on and ensure that we have enough layers, proper technique, etc. We're happy with how the pieces have come out, but I'm curious about what you did as a finishing touch on your tubes. I'll post pictures later, but I sanded them down to about 2000 grit and have an extremely smooth surface, but they look a bit dull. Did you clear-coat your tubes at all? I assume standard aerosol can clear coat won't hold up to the aerodynamic friction and heat, but are there alternatives? (I was thinking something like an automotive engine enamel)
  • The stability in that initial OR file I posted isn't final by any means. I've been holding off to post updates because we're constantly revising small aspects of the build that ultimately change the layout of the entire rocket. For example, that picture has the e-bay functioning as the coupler, but in the time since I posted we decided to package the e-bay as far forward in the NC as possible. This is intended to be practice for building a space-shot rocket next year, which would require us to make certain design decisions that we want to test with this (slightly) less complicated build. I remember reading one of your posts about preferred stability for mach flight, but in the context of this build what would you expect that margin to be? The higher-ups that we're working with want that number right around 2, but I'm pretty sure what matters is that it doesn't fluctuate a large amount for the duration of the flight.
@Zertyme,
  • Thank you for sending that! Although that exact process won't work for us since we're doing a min-diameter rocket, the concept seems viable. I haven't posted an update yet, but our two-piece compression molds flopped. We're having a hard time getting consistent results, so we're definitely open to trying something new. We haven't started the layup yet, but we have a three-piece mold ready to go that looks very similar to the one in your pictures.
  • I didn't consider an aluminum leading edge. How did you get the exact profile that you wanted? Surely those small pieces aren't CNCed, right? No matter what our fin core is going to be, we were planning on doing a tip-to-tip layup since its worked in the past for us.
 
Finishing tubes
I used to care about the cool naked carbon fiber look. Using my method (if I remember it correctly), I simply sanded them down to ~300 grit, cleaned them up really well and shot them with a fancy two part clear coat which I can't remember the name of. If I was going to go for this again, probably I'd try Jim Jarvis's method. I'm lazy though, so now I do a couple layers of Interlux 404 Primekote primer (brushed on), followed by an Interlux Perfection topcoat (also brushed on).

In general I've had really good luck with all the two part paints I've used. The only damage I've seen is a little bit of chipping and bubbling along the fin leading edges, perhaps a little on imperfections on the nosecone and some small bubbling behind the fins (motor heat, not aero heating). I've had really bad luck with all the rattle can "high temp" paint I've tried. Bbq paint, engine enamel, I've tried a bunch of them and they all seem to bubble of the fin leading edges, nosecone and base of the fincan just like regular cheapazoidal spray paint.

Stability margin
If I had to pick a single number, I would say at least 2 (through the flight profile and including mass change due to motor burn, not just subsonic). On all my builds I'm trending towards 2.5 or so (as built, not as designed) unless I'm really pushing things like with that 3" build I did. Stability margin does a few things for you, all of them important:
  • Allows for a little slop during the build because it can be hard to estimate and then actually hit a design Cg on your as built rocket unless you have a lot of experience and are deliberate about your override mass numbers.
  • Allows for flight anomalies - winds, shorter tower than you wanted, off axis Cg, Tony's nozzle story etc.
  • Allows for modeling error - as they say, all models are wrong, some are useful. I think Rasaero, and to a lesser extent (at these speeds) OR give a reasonable estimate of the Cp location, but I take both with a grain of salt. CFD has its own issues (user error usually dominates actual model error here unless you really know what you're doing).
 
Thought today would be a fitting time to post an update.

First, on an unrelated note, I finally was able to launch my scratch-built 4in. diameter, 5ft tall L1 Cert rocket. I've had this thing built for over a year, but finally found the time to launch it today at Hutto. Perfect cert flight on a CTI I345, dead straight, up to about 3,800ft. Felt like spoiling myself, so spent some more money on a 54mm Aerotech I215R. The second launch was also perfect; I couldn't have asked for better outcomes. Also, finally got to meet @JimJarvis50 and show off our first NC prototype in person.

IMG_3819.JPGIMG_3815.JPG
^Me and my orange rocket with two other members of our club also successfully launched for L1 Certs. The Navy rocket had chute deployment issues on the first (technically second, but that's a long story) attempt and snapped a fin. We quickly glued it back on and launched it successfully. The silver rocket flew straight as an arrow to cap off a great day at Hutto.

View attachment L1 Launch.mp4

Safe to say that this is just the beginning of a long addiction. Looking forward to getting my L2 and 3. Now down to the business,


Body Tubes

I made considerable progress in layup methods for our CF tubes. The first three tubes had some issues with resin veins and surface finish, but this fourth tube was perfect. Five layers of 2x2 twill 3K CF. The next test piece I make will likely have three layers of fiberglass between two layers of CF to increase the thickness to match our NC. Still debating whether this is a good idea or not.


View attachment Tube.mp4

IMG_3778.JPG
^Tube before trimming and sanding.


Nose Cone

I haven't been working much on this myself, but we were able to use the 3D-printed molds to produce a 20in. fiberglass cone that we are extremely happy with. Laid up the two mold halves individually with five layers each, then joined the two molds and used the excess material to form the bond between the two while it was still wet. Then inserted a type of balloon into the open mold and inflated it to act as a plug to squeeze out any excess resin. Worked flawlessly, although we're going to update the mold shape to include 4in of body tube length. We weren't able to get many pictures of the layup process because it was so technical and all hands on deck, but I'll make sure to take some next time. This is the finished product, notice the almost perfect seam line on the right side.

IMG_3790.JPG

The interior of the NC was almost perfect as well. There was a small air gap from the balloon in the tip, so a small flap of fabric folded down. We were able to send it down and fill any imperfections with shredded fiberglass. We're hoping that the next prototype is even cleaner than this one.


E-Bay

We've made a ton of progress on our e-bay design. Taking @robopup's design into consideration, we're going to be SLS printing this piece to squeeze into the NC. The aluminum tip is threaded into a spacer above the sled, and we're hoping to integrate temperature probes into the tip itself. We're going to be using a CO2 ejection charge rather than black powder due to concerns about ignition at 50k feet.


110322.4.JPG110322.5.JPG110322.6.JPG

This is a model of the assembled NC. We're rearranging a few of the electronic components, but this is the general layout of everything.
 
Last edited:
We also printed a prototype of the e-bay and mocked up the electronics.
IMG_3783.JPGIMG_3784.JPGIMG_3785.JPGIMG_3792.JPGIMG_3793.JPG


Fins

I'd love some advice on our fins. I mentioned a while back that we weren't successful at making a compression mold, so we decided that machining the fins was the way to go. Unfortunately, we aren't allowed to machine G10 on any CNCs here on campus, but we would be able to use aluminum with no issues. Since our fins have a really precise airfoil shape, it seems like machining them out of aluminum guarantees symmetry that manufacturing them by hand wouldn't be able to produce. We have a rough estimate of the fin flutter that aluminum fins would experience. Are there any practical issues with using aluminum? We would be doing a CF tip to tip over whatever material we decide on. Also discussing the possibility of a fin can, but that has a long way to go...
 
Do NOT trust an eyebolt to be held in place with loctite or JB weld. Use a jam nut. You may need a long extension and a crows foot to tighten in place.
I forgot to mention it but jam nuts were definitely in the plan. Read my fair share about bad experiences with eyebolts.
 
A few quick comments:
  • Congrats on the cert!
  • To give credit where credit is due: My ebay configuration was inspired by (stolen from?) @watheyak. I've been very happy with it...
  • EDIT - as plugger said, I'm not very familiar with EasyMegas but most altimeters are particular about the axis they are mounted.
  • Fins - If you have the capability to machine thin aluminum to some specific profile, I would probably choose the aluminum. For reference... aluminum E=~10e6psi. A quick comparison of a quasi-iso CF layup ([-45/0/45/90]ns) of typical uni properties has E1=E2=~.75e6psi, and if it's all 0/90's ([0/90]ns) E1=E2=~.95e6psi. Aluminum density is ~0.1lb/in3, Typical uni CF is ~0.056lb/in3. You also have none of the good quality laminate/peeling concerns with the aluminum if you don't tip to tip, no heating concerns with the composite matrix (having said that, I'm no expert in metallic heating...). If you choose the aluminum route, make sure you have a plan for galvanic corrosion.
 
Last edited:
here's a link about aluminum fins and things learned

https://www.rocketryforum.com/threa...eter-m2245-project.143884/page-5#post-1814886
on another project ....I RSO'd a 98mm min dia with aluminum fins. They said the fins were welded on by a professional.

all 4 fins came off before motor burn out. Don't just tack the fins on and cover them with filler.

The N5600 has kicked the arses of many experienced fliers.

I've seen a few teams not even make it off the rail without major issues. Follow the instructions..if your not sure ask questions..if you think you know...ask even more questions.

Tony
 
here's a link about aluminum fins and things learned

https://www.rocketryforum.com/threa...eter-m2245-project.143884/page-5#post-1814886
on another project ....I RSO'd a 98mm min dia with aluminum fins. They said the fins were welded on by a professional.

all 4 fins came off before motor burn out. Don't just tack the fins on and cover them with filler.

The N5600 has kicked the arses of many experienced fliers.

I've seen a few teams not even make it off the rail without major issues. Follow the instructions..if your not sure ask questions..if you think you know...ask even more questions.

Tony
One thing I noticed with that fin layout was that he only relied on epoxy between the base of the fin and the airframe (I can't tell if the base of the fins had epoxy filets that sheared off or if that is part of the aluminum). We have a reliable tip-to-tip method that we've used on multiple rockets, but we're still deciding if it's necessary. Based on that error you sent, I'm thinking that it's in our best interest to make the fins as strong as we possibly can, but I'm not sure if we need to worry about the leading edge with a tip-to-tip layup.
 
another good read..

https://www.rocketryforum.com/threads/98mm-n5800-md-rocket-for-balls.40831/page-15#post-404831
There's several of these...good reads ..

One interesting thread I'm not finding is Jerry O's thread about CNC aluminum fins out of Aluminum H or I beam. It saves a ton of time and money..as compared to making them from solid blocks.

Going back to Chris Atterbery's flight. His fins were very thin. The leading edges were rolled over..not sure if it happend prior to the "yard sale" or after via the shock cord rolling over the leading edges?

Tony
 
Back
Top