3D Printing 3D printed tooling strategy for high performance nosecone build

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Adrian A

Well-Known Member
TRF Sponsor
TRF Supporter
Joined
Jan 21, 2009
Messages
3,152
Reaction score
2,795
Location
Lakewood, CO
I've been daydreaming about getting a 3D printer and making nosecones and av-bay parts. Today I ordered a Bambu P1P with the automatic material system. I have a build strategy I wanted to run by you guys:

I'm planning to make a 38mm, high speed (let's say Mach 3) nosecone for a 2-stage project. I think the final cone will need to be made out of composites, and I'm prepared to do that. I have made fiberglass molds before, and for me they are a PITA last resort, so I have been planning thinking about printing a disposable interior made out of PVA, laying up carbon fiber and fiberglass over that, and also printing an outer surface sanding tool to enforce a perfect outer shape.

I bought a 2' section of 1/2" steel hex bar from McMaster that can go down the centerline during construction to keep all the PVA sections aligned with each other. It will also chuck straight into a drill so I can spin the assembly like a lathe. I'll print the nosecone sections with a hex hole down the center for the bar, and then glue them together. Then I'll lay up the nosecone shell, with mostly uni-directional carbon fiber but with fiberglass where the tracker needs to see out. Vacuum bag, room temperature (plus a bit) cure of high-temperature epoxy.

The outside will start out pretty rough after vacuum bagging, which is where a printed outer mold line (OML) sanding jig will come into play. The jig I'm envisioning will have 2 arms that have the OML diameter and shape between them, plus a small allowance for the sand paper thickness. Considering the build volume of the Bambu P1P, this will likely need to be made in several pieces and glued together, also. Once I get done shaping the outside, it's time to remove the interior, and here is where using PVA will hopefully let me just dissolve it away. After that I'm left with a thin composite shell that I plan to post cure in an oven.

Questions:

1. I'm also thinking about printing in a structural feature or two that will go on the inside of the shell to help hold an av-bay in place. I'm wondering if I could print it in nylon-CF or PET-CF as a materials change while I'm printing the rest of the material out of PVA. The printing temperatures may not be compatible. I don't need incredible strength, but I would like to be able to post-cure the nosecone shell to about 200C with this printed feature in place. I suppose I could just print it separately and glue it in afterward, but I like the idea of avoiding misalignments by building it in from the beginning.

2. Should I print the whole disposable center out of PVA, or maybe print most of it with a better material and then just have the outer shell be PVA, maybe with some axial channels in between so the water can get access to dissolve the PVA? Is the PVA strong and stable enough to use as the disposable core by itself?
 
Last edited:
Subscribed. Have you thought about Polycarbonate blend with carbon Fiber?
 
I didn't see that in the Bambu options, but possibly similar is PET-CF, which had the best properties. However, it's not compatible with their automatic material system, so I couldn't use it in a print mixed with PVA.
It is a Prusament filament. You would have to work to get it to work with a Bambulabs
 
With SLA printers, they make a meltable wax-like substance for dental and jewlery molds, I believe (never purchased it, but I think that's its purpose). Is there something similar with filament printers and would that be helpful for this project?

Sandy.
 
Interesting. For me the hardest part of making a nose cone mold is producing the plug. My lathe skills are feeble. For cones that aren't too big, I've started 3D printing the plug. A little bit of fill, finish and polish produces an decent plug. Making the actual mold is tedious, but straight forward.

I've also started using the bladder technique described in TVM's videos: https://www.apogeerockets.com/How-to-Make-Carbon-Fiber-Tubes?m=education Inflating the bladder to just 30psi gives me double the compression of a vacuum.

I wonder. I have never post-cured my stuff. Would your thin nose cone stand up to a post cure cycle without deforming? My understanding is that the original mold is included in the post cure to prevent that problem.

Is sanding the vacuum-bagged layup any less work than sanding and finishing a mold surface? How many nose cones are you going to make? I'd guess that after 3 or 4 nose cones, the mold-making overhead doesn't look too bad...,.

As I said, interesting!
 
It is a Prusament filament. You would have to work to get it to work with a Bambulabs
I was surprised that the PET-CF from bambulabs was incompatible with their automatic material changer. Maybe it's too brittle?
With SLA printers, they make a meltable wax-like substance for dental and jewlery molds, I believe (never purchased it, but I think that's its purpose). Is there something similar with filament printers and would that be helpful for this project?

Sandy.
As far as I know, the PVA is the filament equivalent to this. There is another material that dissolves in limogene.
Interesting. For me the hardest part of making a nose cone mold is producing the plug. My lathe skills are feeble. For cones that aren't too big, I've started 3D printing the plug. A little bit of fill, finish and polish produces an decent plug. Making the actual mold is tedious, but straight forward.

I've also started using the bladder technique described in TVM's videos: https://www.apogeerockets.com/How-to-Make-Carbon-Fiber-Tubes?m=education Inflating the bladder to just 30psi gives me double the compression of a vacuum.

I wonder. I have never post-cured my stuff. Would your thin nose cone stand up to a post cure cycle without deforming? My understanding is that the original mold is included in the post cure to prevent that problem.

Is sanding the vacuum-bagged layup any less work than sanding and finishing a mold surface? How many nose cones are you going to make? I'd guess that after 3 or 4 nose cones, the mold-making overhead doesn't look too bad...,.

As I said, interesting!
I tend to get frustrated with making the part line barrier, and end up needing to refinish the inside of the mold, which is a pain because it has tight concave curves. Also when you lay up in each half there is going to be a seam or overlap that is thicker than otherwise needed, and that can get in the way of cramming lots of stuff in there. I find sanding the outside surface of a cone much easier. I'm hoping that the sanding jig I have in mind will make it even easier and more precise. And it's only one surface I need to perfect, rather than 2.

My next favorite way to do this is to finish the plug, put it inside a fiberglass or polycarbonate cylinder, and then cast clear silicone around it. No re-finishing of the mold is required, or even possible. And no part line barrier is needed. I slit the cylinder lengthwise with a saw and pry it open to help slide the silicone mold out, and then just push the mold off of the plug. The silicone has natural mold release that's good for about 3-5 uses and then it's a crap shoot about whether silicone spray adhesive is adequate after that. In my experience, no. The bad part about this is that to lay up the part, you need to do it just through the open end, which is a royal PITA.

I do usually post-cure in the mold, up to whatever is the mold's limiting temperature, and then bump it up higher with just the part. I like to think it's quite rigid by that point, but I think I did tweak my fin can once during the post-cure of a tip-to-tip layup when I had very large, thin, uni-carbon fins.

I should check out that bladder video.
 
I was surprised that the PET-CF from bambulabs was incompatible with their automatic material changer. Maybe it's too brittle?
The CF in filament makes them abrasive and would very quickly wear through components in the AMS system and the tubing. In fact it will still wear through any PTFE tubing that it goes through and its always a good idea to have extra on hand.
 
The CF in filament makes them abrasive and would very quickly wear through components in the AMS system and the tubing. In fact it will still wear through any PTFE tubing that it goes through and its always a good idea to have extra on hand.

If you use a hardened nozzle, it is not that dramatic. I have used the same nozzle, PTFE tubing, and heatbreak in my Prusa MK3S for over 12 months.
 
If you use a hardened nozzle, it is not that dramatic. I have used the same nozzle, PTFE tubing, and heatbreak in my Prusa MK3S for over 12 months.
Not talking the hotend but rather if a PTFE tube is used between the tool and hotend (including bowden setups). The curves in the tubing will cause the filament to rub on the sides. In the hotend, the PTFE is straight and it is also a short segment.
 
[snip]

I should check out that bladder video. [/snip]
The way I was taught to make lightweight fiberglass nosecones was from a guy who does gliders and the method he put forward was a 2 piece mold with a bladder. It was tedious for sure, but after a handful of failures and refining production methodology, they were pretty repeatable and very light.

One of the keys for me was learning to make the bladders shaped correctly and sealed. In my brain, I would have assumed a balloon would have been fine, but the best results I got were making the bladders by heat bonding plastic that was shaped somewhat correctly and then turning that inside-out so that the heat bond line was inside, not outside. A lot of failures during that learning process for sure. I have also made split silicone molds and feel like they would be too soft for a pressurized bladder system, but I could be mistaken.

With the disposable plug method you're talking about, would you be using cloth or just fibers for the layup. I imagine it could be really cool to epoxy the plug and lay the fibers vs. using cloth, but I've never attempted anything like that so I have zero clue if it is viable or not. It seems like it would be tedious to get the final finish with either cloth of fibers vs doing it in a mold, but again, never tried it.

I look forward to your ultimate success and also the failures in the process of learning (you might not look forward to the last part, but its cheaper and easier for me to learn from your mistakes than making them myself!!!).

Looking forward to seeing more!

Sandy.
 
The way I was taught to make lightweight fiberglass nosecones was from a guy who does gliders and the method he put forward was a 2 piece mold with a bladder. It was tedious for sure, but after a handful of failures and refining production methodology, they were pretty repeatable and very light.

One of the keys for me was learning to make the bladders shaped correctly and sealed. In my brain, I would have assumed a balloon would have been fine, but the best results I got were making the bladders by heat bonding plastic that was shaped somewhat correctly and then turning that inside-out so that the heat bond line was inside, not outside. A lot of failures during that learning process for sure. I have also made split silicone molds and feel like they would be too soft for a pressurized bladder system, but I could be mistaken.

With the disposable plug method you're talking about, would you be using cloth or just fibers for the layup. I imagine it could be really cool to epoxy the plug and lay the fibers vs. using cloth, but I've never attempted anything like that so I have zero clue if it is viable or not. It seems like it would be tedious to get the final finish with either cloth of fibers vs doing it in a mold, but again, never tried it.

I look forward to your ultimate success and also the failures in the process of learning (you might not look forward to the last part, but its cheaper and easier for me to learn from your mistakes than making them myself!!!).

Looking forward to seeing more!

Sandy.
The glider guys do amazing work. Often more skillful and sophisticated composites manufacturing than what I have seen in the space industry.

The silicone molds that I have done just have a split in a strong tube that surrounds them and the silicone is continuous. The tube (fiberglass or thick-walled polycarbonate) keeps the silicone mold straight and contained. When the silicone doesn't have anywhere to go it doesn't compress as much. I would put the whole thing into a vacuum bag, and then have a formed inner bladder that goes inside the cone. For the inner bladder, silicone painted over the same plug and turned inside out has the right basic shape but is a little too large, and it's tough to get vacuum bag sealant to stick to it. Heat bonded plastic could work better that way.

I like using unidirectional carbon, but it's more beneficial on fins than on a nosecone because on fins you know exactly what direction the stresses are. Braided sleeves (+/- 45 degrees) conform nicely to a cone and are easy enough to work with that I can almost sometimes get them to lay down correctly on the inside of an un-split mold. But the stresses on a nosecone are probably mostly bending/buckling and hoop stresses, so 0/90 would probably a little more efficient. I'll probably lay down a thin layer of hoop fibers, then unidirectional fibers aligned with the rocket axis, and then more hoop on the outside. Where I need RF transparency I may just overbuild with FG braided sleeve. I guess I would need to predict and account for the extra thickness of the FG area in the disposable plug design if I don't just make it one consistent layup. But I'm probably over-thinking this. I am looking forward to having the layup be accessible, though.
 
Having traveled down this road, I would recommend printing with HIPS instead of PVA. I had major trouble with PVA getting crushed in the vacuum bag process unless you print with a very, very high infill.

HIPS is also dissolvable, but uses D-Limonene which you can buy by the gallon on amazon.

I switched to 3d printed HIPS mandrels when I was using my filament winder 10 years ago. The details are in this thread: https://www.rocketryforum.com/threads/filament-winding.120441/post-1525926

Only thing I would add is you can shoot your mandrel/mold with some primer to remove the 3d printed layer lines and it will still dissolve, but leave a much nicer finish if that matters to you.

Alex
 
Adrian, I like your original idea of using composites over a printed core, which I have done with some success. I'd never thought of printing a sanding jig to get it symmetrically finished, but the idea is brilliant.
Assuming that the composite skin will bear all or most of the stresses, the core can be sacrificial or so light that you don't care. I worked on the idea of using water-soluble PVA filaments (often used for supports), but could never get them dissolved satisfactorily, and the resulting nose cone was noticeably weaker after removing the core. Of course, you can always add more material to the skin, but I like the idea of the core contributing some of the strength.
Most of my successes recently have used foaming PLA filaments, which reduce weight by at least 50%. As the filament passes thru the hot end of the printer head, gasses are formed in the extruded filament, which is laid down like a bead of PU foam. The foaming filament was developed for printing model aircraft wings. The core is still fairly strong, and light enough to be negligible.
I've settled on as thin of a printed skin as possible, usually 2 or 3 nozzle thicknesses, and a 5-10% 3D fill using the cubic fill pattern from the Cura slicer (imagine a cube balanced vertically on a point), which distributes the stresses equally in all directions. The skin is often multiple layers of light fibreglass cloth (25 g/m2) and epoxy, using the layers to vary the direction of the weave, and sanding by hand. Larger nose cones (98mm) I build in sections using bulkheads, because my printer only prints to 30cm high, and regular 80 g/m2 fibreglass cloth.
Someone expressed concern about anchoring an eye bolt or attachment for a shock cord. I print a vertical hole in the core, and glue in a hollow CF rod to work as an anchor. The CF rods are commonly available from kite shops, and come in different sizes. I glass over the bottom of the nose cone with 2 layers of 80 g/m2 cloth, which bonds to the CF rod and gets integrated to the exterior bearing skin. I tap the CF rod to fit the eye bolt, and Bob is your uncle.
 
Adrian, I'm curious which profile you favor for a Mach3 nose cone. Most of my nose cones are Von Karman profiles, often 7:1 slenderness with a spherically blunted tip, which is good to about Mach 2. Beyond that, I understand that some of the power series profiles might be more efficient.
I'm no aeronautical engineer, though I can appreciate the math. I'm mostly concerned about keeping the boundary layer cool, especially close to the tip.
Von Karman profiles seem to be a good compromise for sub-Mach, thru-Mach, and Mach less than 2, which covers most of the rockets we fly.
 
Last edited:
Not talking the hotend but rather if a PTFE tube is used between the tool and hotend (including bowden setups). The curves in the tubing will cause the filament to rub on the sides. In the hotend, the PTFE is straight and it is also a short segment.
I have not experienced that degradation.
 
Adrian, I'm curious which profile you favor for a Mach3 nose cone. Most of my nose cones are Von Karman profiles, often 7:1 slenderness with a spherically blunted tip, which is good to about Mach 2. Beyond that, I understand that some of the power series profiles might be more efficient.
I'm no aeronautical engineer, though I can appreciate the math. I'm mostly concerned about keeping the boundary layer cool, especially close to the tip.
Von Karman profiles seem to be a good compromise for sub-Mach, thru-Mach, and Mach less than 2, which covers most of the rockets we fly.
I'm planning for Von Karman. Not sure if I want the minimum drag per volume or the minimum drag per diameter version. I have had good success with the drag-per-volume version a long time ago. I would rather have a longer taper to the cone than have any extra cylindrical length in the rocket so everything ut the motor will go into the cone, and the front end of the notor case will be the nosecone sliding interface. I'm interested in minimizing the wall thickness and avoiding any unused volume. I may go for a higher L/D ratio than 7:1. I'll know after I play with options in RASAero. The rocket spends little enough time above Mach 2.5 that I think VK is still the best overall. Intuitively I think a skinnier VK will do better at the higher Mach numbers than the VK nosecones you typically see, but I'll trust what RASAero has to say about that.
 
I had an old pdf somewhere of someone's empirical results of achieving the best rocket for altitude: things like 3 fins, 70 degree leading edge taper how big the fillets should be, and Von Karman nose cones at 5:1 L/D ratio. While slenderer nose cones are better, there's not much more to be gained, according to the paper.
I've understood that high Mach flights have more to worry about where the Mach waves hit the rocket, local heating at nose tip and fin leading edges, how the boundary layer of air clings to the rocket. Most sim programs can calculate drag without considering some of these secondary effects. I've found that RockSim almost always sims higher using conical nose cones than VK nose cones at higher velocities, but I don't understand enough about the resulting Mach wave to say it's the best design for higher velocities. Also, 4 fins seem better for resisting the effects of Mach waves, though I don't understand why.
It seems with your sub-mindia design that one of the controlling design parameters will be having just enough space to house your electronics and chute, and making it long enough to house it all. A long thin nose cone will have less wetted area than a short nose cone with a bit of airframe, but the longer and thinner it is, the less useful space you will get, unless you need some nose weight. It would seem a classic iterative exercise with a sim program to determine the sweet spot (min wetted area/max usable volume/maximum altitude) assuming a VK profile and a minimum slenderness ratio. I could wrap my head around that kind of challenge :)
 
I've been daydreaming about getting a 3D printer and making nosecones and av-bay parts. Today I ordered a Bambu P1P with the automatic material system. I have a build strategy I wanted to run by you guys:

I'm planning to make a 38mm, high speed (let's say Mach 3) nosecone for a 2-stage project. I think the final cone will need to be made out of composites, and I'm prepared to do that. I have made fiberglass molds before, and for me they are a PITA last resort, so I have been planning thinking about printing a disposable interior made out of PVA, laying up carbon fiber and fiberglass over that, and also printing an outer surface sanding tool to enforce a perfect outer shape.

I bought a 2' section of 1/2" steel hex bar from McMaster that can go down the centerline during construction to keep all the PVA sections aligned with each other. It will also chuck straight into a drill so I can spin the assembly like a lathe. I'll print the nosecone sections with a hex hole down the center for the bar, and then glue them together. Then I'll lay up the nosecone shell, with mostly uni-directional carbon fiber but with fiberglass where the tracker needs to see out. Vacuum bag, room temperature (plus a bit) cure of high-temperature epoxy.

The outside will start out pretty rough after vacuum bagging, which is where a printed outer mold line (OML) sanding jig will come into play. The jig I'm envisioning will have 2 arms that have the OML diameter and shape between them, plus a small allowance for the sand paper thickness. Considering the build volume of the Bambu P1P, this will likely need to be made in several pieces and glued together, also. Once I get done shaping the outside, it's time to remove the interior, and here is where using PVA will hopefully let me just dissolve it away. After that I'm left with a thin composite shell that I plan to post cure in an oven.

Questions:

1. I'm also thinking about printing in a structural feature or two that will go on the inside of the shell to help hold an av-bay in place. I'm wondering if I could print it in nylon-CF or PET-CF as a materials change while I'm printing the rest of the material out of PVA. The printing temperatures may not be compatible. I don't need incredible strength, but I would like to be able to post-cure the nosecone shell to about 200C with this printed feature in place. I suppose I could just print it separately and glue it in afterward, but I like the idea of avoiding misalignments by building it in from the beginning.

2. Should I print the whole disposable center out of PVA, or maybe print most of it with a better material and then just have the outer shell be PVA, maybe with some axial channels in between so the water can get access to dissolve the PVA? Is the PVA strong and stable enough to use as the disposable core by itself?

Seems doable, only one way to really find out though :)
Post your process and results.
 
I have been printing rocket parts (lots of motor retainers and rail guides, some nose cones and fin cans) from polycarbonate as of late. Polymaker PC max - they have a CF version, also have a PC-ABS blend. The regular PC seems to be quite good (have annealed some per the instructions, but mostly just left it as is).
 
With SLA printers, they make a meltable wax-like substance for dental and jewlery molds, I believe (never purchased it, but I think that's its purpose). Is there something similar with filament printers and would that be helpful for this project?

Sandy.
I have used a wax filament for fdm printing. Its used for jewelry making and lost wax casting.
 
Back
Top