I recently attended the National Association of Rocketry's 60th annual competition, where I undertook a personal goal to determine if printed models could be competitive. The models in this competition are usually comprised of paper, balsa, plywood, fiberglass, and/or vacuformed plastic. Other than a few printed nosecones, and details used in scale models, printing has seldom been used.
This model is 100% printed in ABS. It achieved a second place altitude of 274 meters with an Estes 'C' motor, just six meters out of first place. As this was the first model rocket of any kind that I'd flown in over 30 years, I was mildly pleased with this result. (For an old guy, 'mildly pleased' = 'wildly ecstatic'.)
The intention of the Eggloft event is to fly a raw large egg to the highest possible altitude and return it unbroken. This model is intended for 18 mm model rocket motors, which should allow it to be flown in 'B', 'C', and 'D' motor class events.
Flying with a little more experience and a bit of fine-tuning of the model, I expect it can do better, and I expect that printed models will find a place in competition.
My other Things: https://www.thingiverse.com/garyacrowellsr/designs
My other Model Rocket Things:https://www.thingiverse.com/garyacrowellsr/collections/model-rockets-vcp
My Model Rocket Collection: https://www.thingiverse.com/garyacrowellsr/collections/model-rockets
My Collections: https://www.thingiverse.com/garyacrowellsr/collections
All of this model was printed on an XYZ DaVinci 1.0, upgraded to Marlin 0.92. Slicing was done with Simplify3D, and it's the Simplify settings that I'll be detailing. Each component will often have different settings, though the primary process used was Corkscrew (AKA Spiral or Vase) mode.
I suspect that every other flying model rocket or model rocket component that you've seen on Thingiverse has been of very sturdy printed construction. That approach is simply not viable for a competition model - it would be much too heavy to be competitive. OTOH, a competition model really only has to survive for one flight. So we sacrifice durability for weight. Corkscrew printing mode produces single-wall components that are light enough (barely) for competition. Print two. (Or three. You'll probably break one during assembly.)
Note that this is printed in ABS. I'm pretty certain that trying this in PLA would be a disaster (though maybe the new 'super' PLA's might work). PETG would probably work as well.
Note also, that as a competition model, there is no provision for a conventional 'launch lug'. This model is intended to be flown out of a tower, with a motor piston. Some people with more experience with a piston can fly a 'naked piston' without a tower. If those sentences didn't make any sense to you, you need to ask around on one of the model rocketry forums or read the technical notes regarding pistons and towers from the NAR or a manufacturer (https://www.apogeerockets.com/).
========== Nose Cone and Nose Cone Tip ==========
The first component is the main body of the Nose Cone. The STL is a solid shape, which becomes single wall when printed in single outline corkscrew mode. Layer Height was 0.15, (No infill or support is permitted with this mode.) I print most parts with a 2 outline, 2 layer brim, with an offset that allows the brim to be easily pinched off, while still adding some adhesion assistance. No solid Top or Bottom layers.
Now the Nose Cone is a blunt elliptical design, which is good for subsonic model flight, but you will notice that the Nose Cone has a flat, truncated tip. Since there is no support allowed in corkscrew mode, I don't think that there is any printer or material that could complete the tip as a single piece. Therefore, there is a separately printed Nose Cone Tip. To get a reasonably smooth tip, this was printed at LH = 0.15, 3 Outlines, 3 Top Solid Layers, and with support. Although it's small, this is a component that could probably use some work to lighten it up. 0.1 LH would probably be appropriate for the last few layers. The Nose Cone Tip is inserted into the Nose Cone from the inside, where an inner ridge holds it in place. A quick wipe of Acetone, MEK, or CA seals it in place. The Nose Cone is also the upper part of the capsule that holds the egg payload.
========== Shoulder ==========
The Shoulder is the lower half of the capsule that holds the egg. Its large diameter is sized so that it's edge just fits within the Nose Cone, to a depth of about 6-8 mm (depends upon the egg and the amount of padding you put around it). This was printed corkscrew like the Nose Cone, but I used two processes. Due to the overhang angle on the small end, I printed the first 5mm with an LH = 0.15, and the rest at LH = 0.2. Because this is an interior part, we don't care about the smoothness of its surface. Otherwise same as the Nose Cone; small brim, no solid top or bottom.
========== Altimeter Capsule ==========
You've noticed that the Shoulder component is open at the bottom (just like the Nose Cone was open at the tip). This component closes that opening, and provides a place to house the altimeter. This is sized to hold an Altus Metrum MicroPeak altimeter, which just fits within the little box. The capsule is glued to the open end of the shoulder, with your favorite glue. This was not printed corkscrew, but with LH = 0.2, 3 Bottom Layers, 2 Top Layers and 2 Perimeter Shells, with 10% infill and support. The purpose of the two open rings will be apparent later when we finish assembly and prep for flight.
========== Transition Body ==========
This is the part that makes printing shine. With other materials, it's very difficult to make a smooth aerodynamic shape that transitions from the size of an egg to the diameter of the 18 mm rocket motor. Most people settle for a conical transition, but that leaves non-aerodynamic sharp edges at the top and bottom of the transition. I'd like to say that the shape here was designed after hundreds of hours of subsonic CFD computations... but it's not. It's the TLAR* method, which often works remarkably well in these situations.
The Transition STL looks like a simple solid, ready for corkscrew print, but it holds a secret. At a single layer, this body is pretty fragile. I broke several of them while doing the finishing steps described later. Although I think it could fly fine as is, I found a neet way to strengthen it. The STL actually has some micro-sized cuts in the surface, running lengthwise, at three locations. View the sliced image to see what I mean. While the slicer is still doing corkscrew, it has to observe the cuts, which then form interior ridges that strengthen the part. I've never broken a part with ridges. The ridges could use some work to optimize their size. (Yes this does leave a little 'pucker' on the surface - see the finishing details later.)
Printing the Transition is a straight corkscrew, LH = 0.15.
*TLAR = That Looks About Right
========== Fin Can ==========
(A single unit that incorporates a short section of rocket body and the attached fins is called a 'Fin Can'.) This is a pretty little unit that I've used on several models and in several sizes. While the aerodynamic details at this scale aren't critical, it's easy to print in several cool features. The fins are a modified elliptical shape, with rounded leading and trailing edges (not really printable at this scale), with full fillets that are tapered fore and aft. The leading and trailing edges lead into the body smoothly. It is also tapered root to tip, and at the leading and trailing edges.
Printing the fin trailing edge requires support, and the slicer wasn't really doing a good job of it, so support is built-in. It's spaced so that it should snap-off easily. But therein lies another trick that I found. Like most of the other components, this is also printed in corkscrew mode. But wait... If selecting corkscrew, you 'can't' have support, yet I do. The slicer sees four individual closed objects, the body and three supports, and it obviously can't do all of them as a single continuous corkscrew extrusion. So it doesn't. I don't know what other slicers do, but Simplify3D treats all four objects each as a single wall outline and prints them a layer at a time. When it gets high enough that the supports are complete, it transitions smoothly into corkscrew for the remainder of the fins and body. Neet! Same as other components, LH = 0.15, no tops, no bottoms. It's also important under the Simplify 'Advanced' tab to select 'Perimeters only' for External Thin Wall Type. If a single wall is allowed, the outer edges of the fins can transition to a single-walled fin tip. The corkscrew mode naturally produces a two wall fin (one side on the extrusion path out, one side on the path back in). One wall might work, but the transition from double to single wall leaves an indentation in the surface and a weak point (experimentation is needed).
There is one other detail to take care of. It needs a motor mount and some strengthening at the transition joint. For this I use the Simplify3D Advanced process to stop printing at a height of 63 mm. A second process is added starting at 63 mm. This process is not corkscrew, LH = 0.15, Perimeter shells = 2. The STL is set up at this point so that this should produce a short motor mount and a body diameter step-in to create a shoulder that fits inside the Transition. Some work is actually needed on the CAD model here to move the motor mount down, so that the motor extends farther from the bottom of the fin can, to facilitate using a piston (to be explained later).
Print time. I was printing most parts really slowly, often at 40 mm/sec. That's a couple of hours for the fin can, four hours for all the nose cone and capsule bits, and four or five hours for a transition. You can do better. Total printed weight was 21.8 grams, without egg, motor, altimeter or parachute.
That's all the bits to print. See the post-processing to find out how to turn these bits into a flyable model.
A smooth surface finish is critical to flying a low-drag competition model, and I know of three ways that this might be achieved.
- Sand and fill: Sand off the tops of the ridges and other imperfections, then fill with putty or a high-volume spray primer and sand again. Repeat until smooth. This works fine for scale models, but in a performance competition, it adds too much weight.
- Acetone vapor finishing: I've done this before, but I didn't take the time to see how it would work on a model like this. I was worried that the single layer walls might warp. I need to give it a try and see what happens.
- What I did... sand, sand, sand. ABS sands nicely and you can sand aggressively, starting with 150 or 180 grit. I just flat sanded the ridges off completely. The benefit to this is that I measured a 15% weight loss from the material removed. Keep sanding until there is no 'click' when a fingernail is run across the surface anywhere. Then finish off with some 600 to get rid of the scratches from the heavier grits. Finally buff with some 1000-1200. After that, you could really get by with nothing else, but I did give everything a coat of liquid wax (Pledge Floor Care) to give a bit of a gloss.
When sanding the Fin Can, sand one side of a fin - then flip it over and sand the opposite side of the fin the same amount; and sand all three fins to the same degree. You've printed perfectly aligned fins, now you don't want to sand in an airfoil asymmetry. Sand the interior of the Fin Can with a rolled-up piece of sandpaper until you get a nice slip fit on an 18 mm motor.
Remember those Transition ribs and the little 'pucker' that they leave on the exterior? Well, you can almost sand that pucker out. Maybe if you sand more aggressively than I did you could get it out completely, but I'd be worried about thinning the wall too much. That small indentation is in-line with the airstream, so it shouldn't have much of a drag impact. If it drives you crazy, it probably could be filled with a thin filler.
The interior of the Nose Cone and Transition will need to be sanded or scraped for the shoulder to fit properly. Also the exterior of the shoulder of course. It should be a very close fit.
The interior of the bottom of the Transition must be scraped, and the upper shoulder of the Fin Can must be sanded so that they will have a close fit. (Also the Fin Can must be slotted, see below.)
After the final assembly (including the retention wire) and with the Nose Cone / Shoulder assembled and inserted into the Transition, do a final sanding of the Nose Cone / Transition joint. Make matching pencil marks on each piece so that they can always be inserted again with the same alignment. Since this joint is exposed to the airflow, your objective is to sand away any ridge at the joint. Having a mark to re-align them the same way when prepping for flight will ensure a ridge doesn't appear from misalignment.
Also, wear a dust mask when sanding this stuff. It can't be good for you.
There are a couple more things to do for the final assembly before the model can be flown:
The Transition must be glued to the Fin Can. There is one small modification needed to the Fin Can, and that is that the upper shoulder must be slotted in three places to accept the interior ridges in the Transition. This just took a few seconds with a cutoff wheel in a Dremel Tool. I put the slots in line with the fins. I think I used CA to glue these together and it worked fine.
Finally, there is what I call the 'Retention Wire'. This is a piece of 15-thousandths (0.015") music wire. Start with a piece about 10-12 inches long. Loop it in the middle (do not bend) and insert it into the Nose Cone, so that equal lengths of wire extend out of the Nose Cone. While holding the wire in place, glue it to the interior. I used two drops of Gorilla Glue Gel, one on each side, as near as possible to the tip. Try not to become one with the Nose Cone. When that has set, insert the Shoulder into the Nose Cone to a depth of 6-8 mm, with the wires routed to the exterior of the Shoulder. By-the-way, before gluing that wire in, you did sand/scrape so the Nose Cone / Shoulder fit together, didn't you? And, you've got the altimeter Capsule glued to the bottom of the Shoulder, right?
So now you should be able to perceive the madness. Bend each wire over to one of the loops on the bottom of the Capsule. With needle-nose pliers, form a ~1/8" diameter loop in each wire, wrapping the end of the wire around itself a couple of times so that the loop can't unravel. The wire loops should be even with the loops on the bottom of the Capsule.
Clip off the excess wire.
Glue the retention wire into the interior of the Nose Cone.
First, you'll notice that there is no provision for a conventional shock cord attachment. Competition models rarely use them. Instead, you use a 4-5 foot length of 100 lb. Kevlar cord. Fish the cord through the Transition and Fin Can and tie a slip knot in the bottom end. Insert the motor and shim with tape for a fairly tight fit. Slip the cord loop over the exposed end of the motor and pull it tight from the Transition end. That's your shock cord, anchored by the loop around the motor. For extra insurance, you can put an elastic snubber in-line with the cord.
There's room in the Transition for a 16 to 18-inch mylar parachute. That's usually sufficient to land an egg safely. The parachute shroud lines are tied to the shock cord about 18 inches from the end of the cord.
Now it's time to prep the egg. It's common practice to put the egg in a cut-off condom to contain any accidents. Place the safe-sex egg into the Shoulder section with shredded paper or foam as necessary so that when the Nose Cone / Shoulder is closed the egg cannot move. Now, taking care that the Nose Cone / Shoulder does not separate, use a short length of cord to tie one of the wire loops to a printed loop on the altimeter Capsule. Use the end of the shock cord to tie the other wire loop to the other printed loop. Use a short length of cord tied through a mounting hole on the altimeter, with the other end tied to a wire loop. Now you've got everything tied together; if your knots are secure, the egg can't get out, and your parachute and altimeter are tied to the Capsule / wire loops. Nothing is coming apart and it will all come down together.
Insert flame-resistant wadding into the bottom of the Transition. Pack the parachute and coil up the shock cord and insert into the Transition. Turn on the altimeter and insert it into the altimeter Capsule. Close up the Nose Cone / Transition, and align the marks that you applied when sanding. Insert an ignitor and the model is ready for flight.
The wire loops are tied to the Capsule printed loops.
Most printed model rockets use a paper tube insert to protect the plastic from the heat of the motor and its ejection charge. I didn't. Too much weight. But I'd never flown a printed model before this, so I wasn't sure what would happen.
In all of the flights I made at this competition, I never had any problem with single wall motor tubes in contact with the rocket motor. The motors were easily removable and the Fin Cans were readily reusable. Recall that all of this is ABS plastic.
The ejection charge was another matter. From the photo, you can see that about two inches of the Transition forward of the Fin Can was 'wrinkled'. It's still structurally sound and could be flown again, but the wrinkles would probably kill its performance. A couple of things to note though. Where the Transition and Fin Can shoulder overlapped, effectively two wall thicknesses, there was no damage. Also, the only wrinkling was between the Transition ribs.
Another point is that I flew several other models, also single wall Transitions, in bright yellow filament with no ejection charge damage. Could the color make a difference, with the darker more susceptible to heat? However, those other flights were with 13 mm motors, and that might have been the difference. Experimentation is needed.
So there are several potential solutions:
- Different color filament (white or yellow), and/or different filament type; perhaps PETG.
- A coating on the lower interior of the Transition; perhaps thinned epoxy or a heat-resistant paint.
- A small flameproof paper or foil insert in the lower Transition area.
- Double wall thickness in the lower Transition area.
- Added ribs in the lower Transition area.
With that many options to try, I'm pretty confident that this heat issue can be solved without resorting to a full paper stuffer tube.
Ejection charge heat damage extends 2" above Fin Can
First off, the retention wire widget worked great - that's not going to change. You'll notice that this arrangement created only a single break in the model surface that served as both the egg capsule and ejection/parachute separation. However, the altimeter Capsule was a pain to get the altimeter in and out. It's really not much protection, so it's just easier to fly the altimeter 'naked', wrapped in some wadding, and just sitting in the top of the Transition but still tied to one of the wire loops. So the Capsule can go away, replaced by a much smaller base to the Shoulder piece. It will probably have a single loop that both wire loops are tied to. This will result in a weight reduction and more room in the upper Transition area for the parachute.
For the Fin Can, the upper body/engine block needs to be shortened so that more of the motor is exposed beyond the base of the Fin Can. This will facilitate the use of a piston. I also think I'm going to look at replacing the motor mount thrust ring double wall section with ribs similar to the Transition. The ends of the ribs would then serve as the thrust ring.
The Transition ribs need to be optimized. A lot of excess rib material currently in the lower Transition area can be removed. Also, note that the Transition / Fin Can could be printed as a single piece if your printer volume is tall enough (about 10").
And finally, fix the heat issue with one of the methods described earlier. With these changes, I'm aiming for a Version 2.0 model dry weight under 20 grams.
The flight of this model was perfectly straight in a light wind. It did, however, seem to exhibit a slight low-amplitude, high-frequency 'shimmy' on the way up. I suspect that some work can be done to optimize the dynamic stability, perhaps by adjusting the length or fin size.
I flew several other models at the contest that also performed well, and could have done better if not for my clumsiness, inexperience, and other technical issues. I also have several others that haven't flown yet. As they are tested and refined I'll post those additional competition models here.
Additionally, most components of my launch tower and pistons were also printed.
It's my shameful little secret that I design in OpenSCAD. I should be using SolidWorks or Fusion360 but I'm really attracted to the programming format of OpenSCAD - I can whip things up pretty easily. It does have limitations though. It's freeware if you're interested: http://openscad.org
If I can get the code cleaned up enough so that it's not too embarrassing, I will eventually post the OpenSCAD source for this model here.