1/26th Scale Space Shuttle (full stack)

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Polycarbonate fins are your friend, at least to start. I would start with a center tank mounted motor, get balance and trustline worked out as well as the orbiter and booster separation and return automation then work up to gimbaling, and motors in the srbs, that will still be a huge challenge to start and then start adding more of the complexity instead of just going for the full meal deal in one shot..

I don't think anybody's done a full space shuttle stack with motors mounted and burning in the shuttle itself as well as the srbs...
 
Polycarbonate fins are your friend, at least to start. I would start with a center tank mounted motor, get balance and trustline worked out as well as the orbiter and booster separation and return automation then work up to gimbaling, and motors in the srbs, that will still be a huge challenge to start and then start adding more of the complexity instead of just going for the full meal deal in one shot..

I don't think anybody's done a full space shuttle stack with motors mounted and burning in the shuttle itself as well as the srbs...
Sounds good. It was never going to be a "Hail Mary" one shot deal.

I worry about the fins on the SRBs being that close to the orbiter though. I presume that's why they were pushed back on other models so as not to slice through the orbiter at separation. Leaving the SRBs on the whole time and only separating the orbiter, reduces the risk a bit.
 
Sounds good. It was never going to be a "Hail Mary" one shot deal.

I worry about the fins on the SRBs being that close to the orbiter though. I presume that's why they were pushed back on other models so as not to slice through the orbiter at separation. Leaving the SRBs on the whole time and only separating the orbiter, reduces the risk a bit.
You would want to release the boosters before the shuttle, I would think they would drop back and away and not forward into the shuttle, given the tail on the shuttle you could rotate the fins so that one fin is on the backside of the stack facing away from the shuttle tail and one straight out away from the srb parallel to the shuttle wing and get the maximum stabilization and that would clear the shuttle wings in that case.
 
There is a good picture in Nadine's calendar for Woerner's Shuttle for the month of December 2020. The actual flight was in 2005. Andy's Space Shuttle was 6'7" tall and weighed 105 lbs. The SRB's were powered by 2 Aerotech K550's and the Orbiter had 3 Aerotech J180's. To me the Shuttle takes off very fast in the video and fish tails a little bit. The still pictures on Nadine's calendar makes the model look like the real thing. From the video the Orbiter is gliding, but the descent angle looks very steep and the sound of crowd sounds startled upon the Orbiter landing. I think observers say that the landing was hard. Andy sold some of his models through Skunkworks, which was later changed to Polecat. If you search on TRF you can see some references to the Space Shuttle being sold. The price is in the thousands. This probably explains why there is not much documentation plus most of Andy's projects did not have much construction details.
 
There is a good picture in Nadine's calendar for Woerner's Shuttle for the month of December 2020. The actual flight was in 2005. Andy's Space Shuttle was 6'7" tall and weighed 105 lbs. The SRB's were powered by 2 Aerotech K550's and the Orbiter had 3 Aerotech J180's. To me the Shuttle takes off very fast in the video and fish tails a little bit. The still pictures on Nadine's calendar makes the model look like the real thing. From the video the Orbiter is gliding, but the descent angle looks very steep and the sound of crowd sounds startled upon the Orbiter landing. I think observers say that the landing was hard. Andy sold some of his models through Skunkworks, which was later changed to Polecat. If you search on TRF you can see some references to the Space Shuttle being sold. The price is in the thousands. This probably explains why there is not much documentation plus most of Andy's projects did not have much construction details.
Thank you. I have watched that video so many times it's dangerous. I've tried to forage as much information as possible from others who have done this journey. I got my level 2 on a Polecat Aerospace Bullpup (7.5"). It was a great kit. Andy, in my opinion, went the furthest down this road for something as close to the original as possible. George Gassaway has also contributed hugely to this. Anyway, here are a few pictures that I have for research purposes. Copyrights to whomever had the originals.

245338-18e7b2736cdca5bd916fd5bbf10cc5b4.jpg 11910-10724346a7731b2cc4516ec5a795d1b0.jpg

My general premise on this ...
  • Things happen so fast that having a human trying to control the stack at ascent, separation and landing is difficult at best - thus automation and computation become needful. We will have to collect a lot of data to make the whole thing happen.
  • Can we allow the shuttle to separate, make the turn and land under autopilot - including landing gear.
 
You would want to release the boosters before the shuttle, I would think they would drop back and away and not forward into the shuttle, given the tail on the shuttle you could rotate the fins so that one fin is on the backside of the stack facing away from the shuttle tail and one straight out away from the srb parallel to the shuttle wing and get the maximum stabilization and that would clear the shuttle wings in that case.
If I understand you correctly. You describe something like this from the back...

1670543031898.png
 
I used that fin config on my three piece falcon heavy model rocket' the srb's separated as they burnt out, leaving the center section in flight. Which then shredded due to the light weight and g-70 engine. It simmed okay, but it's hard to model after separation.
But, that fin config flew better than the V- config at both ends of the stack. I finally ended up with a 4 fin setup on the srbs And the main, which was tight, but stable.
 
If you're using OpenRocket to estimate CP, I recommend running it without the elliptical tail on the ET. The CP calculations overestimate the effect of tail cones, and the estimates get worse as the tails get shorter and fatter. In the real world I expect the airflow to separate from the rear and create base drag rather than follow the ellipse and produce destabilizing torque.
 
A More Stable Stack

Just to close off the thread from the weekend. Taking Frank and Grog6's advice, we end up with a more stable stack with the following configuration:

stabilized back.jpg

and then to follow up on Jeff's point about OpenRocket treating CPs differently with tail cones included or excluded. Here's two images of how the CP changes according to OpenRocket. It looks like it only slides back a few inches. As to how it reacts in the real world...I guess we'll find out.

shuttle with ET back transition.jpgshuttle without ET back transition.jpg
 
Plans Going Forward

As I've mentioned, this will take a long time to achieve - multiple flights, probably multiple mistakes -- which are opportunities for learning. So, I've decided on a few early tasks that I can start now and share with everyone as I get through them.

Fabrication
There are a number of parts on this model that will require 3D printing. I'm planning on learning a lot more about Fusion 360 for printing SRB skirts, nosecones and an External Tank nosecone that will be quite large (and will probably require me to use an outside service that has a larger print bed). Additionally, the stand offs that the shuttle sits on away from the External Tank will be made of lightweight aluminum tubes that I got from Grainger. I've not done a lot of welding/soldering outside of electronics, so that will be a challenge. I may use a local machine shop to get this right - I'll do an initial sketch build using cheap PVC.

Mechanisms and Control Surfaces
The separation and fall away features of this model will require solenoids and active/passive traction to hold the models together doing boost. I'll be building these out and getting them tested. The orbiter control surfaces - two elevators per wing, a body flap and a split rudder will require a lot of time to figure out. I'll probably be pinging the RC guys on this thread quite a bit to come up to speed on this.

Avionics
I'm going to be using a GPS, an IMU and an Arduino to determine position, orientation and attitude for the orbiter. Learning about this and figuring out how to get the control surfaces to respond correctly with Ardupilot will be a challenge. I've decided to build a simple rig of the orbiter fuselage to get this all figured out.

So early studies and cheap proof of concepts to start to de-risk some of this project is in order. The fastest way to get done is to get started.

I'm not planning on having everything working all at the same time - some things will mature faster than others. Early flights will be all about collecting data and confirming assumptions for the flight controller.
So much to figure out:
  • determine proper CG/CP/Thrust
  • boost to separation of SRBs
  • to separation of orbiter
  • to recovery of ET and SRBs
  • to "righting" the orbiter for glide back,
  • following way points for a return glide, making the turns using the auto pilot,
  • measuring altitude, distance to landing and airspeed and responding with control surfaces
  • and finally landing.
I will make mistakes, hopefully we will learn together.
 
An Early look at Avionics for the Orbiter

Over the holidays, I'll be looking at the various sensors and programming to connect them to an Arduino Uno R3. I put together the following diagram to help understand what each sensor will be used for and the breadth of things to worry about (there's so much to worry about).

So, I have the following sensors and some rationale behind each of them:

  • A GPS for determining position in space - this will be necessary for the auto pilot and flying to waypoints that will be laid out to perform the glide, the approach and the landing. If two GPS points and an altitude represent the landing strip, that vector can be used to determine way points, turns, etc.
  • an IMU for determining orientation in space - this will be necessary for "righting" the orbiter after boost and separation (and even determining boost). The IMU will help the Arduino adjust the control surfaces to achieve level flight or articulation of elevons to reduce or slow the sink speed of the orbiter.
  • Altimeter for obviously determining distance from the ground, but also for the ground altitude.
  • Airspeed is one of the most expensive sensors (the others are $20 or less, with the exception of the Lidar). This will be useful for better precision on controlling the elevons and rudder as well as determining time to waypoints for the glide down.
  • Two distance sensors will be valuable - one for short distance and one for longer distance. The ultra-sonic sensor is useful for distance from the External Tank during separation and re-orientation and will also be useful as the ground gets less than a meter away. The Lidar, the most expensive sensor on the vehicle ($120), will give us distance up to 30 meters away from the ground. Using this in conjunction with ultra-sonics will give us better control at the time we really need it.
  • A solenoid will be used to separate the orbiter from the stand offs on the External Tank.
  • A servo driver board will let the system drive up to 16 servos - including elevons, body flap, rudder and gear. The solenoid to separate the orbiter could be controlled by a servo, but more likely will be handled independently - it's really powered to "pull out" - no real motion outside of a release.
  • Finally, a Micro SD Card reader/writer will allow the system to record and persist telemetry for later analysis.
I'll be talking in future posts about each sensor and will share the wiring and the programming for them.

1671038979529.png
 
Last edited:
Fabricating engine bells and the body flap with Fusion 360

So, I have some progress to share on my Fusion 360 exploration. In the past, I've only played with Fusion 360 as a way to make small parts - electronic sleds and such. But recently as I've delved deeper into the product, I'm quite impressed about what can be done and in so short a time. I created profiles for the orbiter's main engine bells and the small maneuvering thrusters on the pods of the aft fuselage. Then by taking that profile and revolving the surface, I can get the accurate look of the engine bell. I also modelled the body flap that sits underneath the main engines including a built-in control horn and holes to pivot the flap. While these are obviously just low-quality plastic renderings, these can also be milled, or 3D metal printed - titanium, aluminum and other materials are available from many online services. You just upload the design and a few days later they show up in the mail.

As I continue to build out the prototype rig for the orbiter, you'll see these characters again. Just wanted to share what I did. Even if you don't print the parts, Fusion 360 can help you visually think through and design things that normally rocketeers just fumble their way through. I encourage you to add it to your bag of tricks - it's a free trial for 30 days and then $70 a month subscription.

C30526A1-2E35-4C9E-BC52-DACE0F4C16D7.jpeg
 
Last edited:
SRB Mockup

Here is the mockup SRB. It's a 5.5" Blue Tube and the Nosecone and skirt are 3D printed. Still getting used to Fusion 360 and these are very low-quality prints, but it would be a lot more time and effort to do that with fiberglass. I'm going to be experimenting with various ways to separate the SRB from the ET. I have solenoids and servos and will share the results as I get them.

A1499A6D-8B26-4399-B282-C5A23F13EB19.jpeg
 
Last edited:
Playing with Sensors and the Arduino Uno

I'm starting to come up to speed on the Arduino Uno and a bunch of sensors I got in the mail recently. I've worked though how to wire up an Altimeter/Barometer, a GPS and an Airspeed senor complete with a pitot tube. I'll be going into a lot more detail on each of these and sharing the wiring and code behind them. I still have an IMU/Gyro that will handle orientation in space and an ultrasonic sensor (low distance to ground) to integrate into the stack. As a Software Engineer by day, getting into this much hardware and electronics is a different world, but I'm having a lot of fun.

3D982A9F-28B7-440E-B296-2AEE81F9AACC.jpeg2C97BBC1-DBCF-4C4F-869C-484F5014B0F0.jpeg

1F73AD48-76B7-4BFC-B1C2-09C2ADB9B11E.jpeg
 
Last edited:
Andy Woerner had a spectacular HPR Space Shuttle years ago.

This is one of my childhood memories. My dad and I were behind the flight line at the ice cream truck, and while we didn't have a good view of the boost, we had a great view of the impact. It caved in the door of a car, I think it was a Beetle.
 
Are you planning on flying the orbiter as a boosted glider to get it's flight computer and programming worked out ?
Or a smaller boosted glider with the flight computer for testing ?
 
This looks like a really cool project, but I think it's going to end up being a lot heavier and take much bigger motors than you think it will. All that electromechanical stuff adds a lot of weight, too.
 
Are you planning on flying the orbiter as a boosted glider to get it's flight computer and programming worked out ?
Or a smaller boosted glider with the flight computer for testing ?
Hi Charles - Merry Christmas. It will probably be a cheap, foam board version as close to the real size as possible. Probably more EDF to get everything tested, rather than boosted. It will be a long process of testing the orbiter - that much is certain.
 
This looks like a really cool project, but I think it's going to end up being a lot heavier and take much bigger motors than you think it will. All that electromechanical stuff adds a lot of weight, too.
Hi Cerving - Merry Christmas.

Actually, the Arduino and sensors are relatively light - I can measure that now - it will be the batteries and electronic retracts that will add the weight. It's pretty clear to me that the orbiter will need to be formed from carbon fiber and fiberglass, with as few bulkheads/plywood as possible. Weight is a real problem, because I think I can prove that OpenRocket cannot sim this stack accurately. I will probably test the SRBs (just one) and then the ET with a motor with SRBs attached, but not loaded - no separation of orbiter to capture data and then, finally, the whole stack.
 
You can use your 3d printer to print female and male mold halves then press in the carbon fiber..
This would allow for mass production of spare parts..(which will come in handy later)
 
This is one of my childhood memories. My dad and I were behind the flight line at the ice cream truck, and while we didn't have a good view of the boost, we had a great view of the impact. It caved in the door of a car, I think it was a Beetle.

We were there, too! Andy certainly never did anything half way. Good old days.
 
Some decisions to alleviate weight

Happy new year y'all. I've been thinking a lot about this project and have decided on two major things: weight of the External Tank (ET) and the weight of the orbiter.

On the ET front, I've decided to move to a LOC/Precision Phenolic tube that is 11.41" - this is still roughly scale when compared to the 12" Sonotube I was going to use. The SRBs are exactly half of the diameter of the ET, so the LOC tube is actually more scale and will not be as thirsty for Epoxy as Sonotube reportedly is.

To state the obvious: this project is part rocket, part RC glider. And you have to deal with each with the technologies that assist them. The SRBs and ET are very straightforward to me and familiar. Yes, there will be challenges there, like the separation mechanisms and getting the CG just right. The orbiter, on the other hand, is an airplane - specifically a glider. It cannot be built with normal high powered rocketry techniques. I've gotten used to building things heavy duty (like the Nike and the ALCM), but in this case, the glider needs to be as light and strong as possible.

Therefore, I've decided to use 3D printed molds to build a carbon fiber/fiberglass shell. By using Aramid Honeycomb, I can avoid bulkheads and tubes (typical HPR building techniques). In this world, the entire orbiter model is built in Fusion 360 and "subtracted" from mold blocks. The printing of these blocks takes a long time - in my first test of printing a fuselage core that was 8.5 x 8.5 inches, it took 3 days to print. The full fuselage is around 53" long and the cabin is 13" long. All told, just printing the molds will take nearly two months. Then the prep of the molds (filling and sanding) and the final casting. The good news is that when the molds are done, if I don't destroy them getting the form out, can be reused again and again. I'm not deceived about this project; I know I will destroy some orbiters figuring this whole thing out.

I've been doing research about this technique online and the best and closest practitioner I can find on the subject is Ramy RC - a brilliant RC builder in Germany. Here are some links - and don't be surprised when after you've viewed a few of these that you are inspired in new ways to build rockets.







Let me know your thoughts...
 
I collect spores, molds and fungus

As I mentioned, the orbiter will be built from fiberglass and carbon fiber built from molds, not built up as I have in the past. This will result in extremely light and strong structures that should enable a good glide from apogee. In the past, doing molds was a lot more complicated than today - it's still complicated with the advantages we have, but assailable by the mere moral. The advantage that we have today is that we can use a program like Fusion 360 to store an object and then "subtract" that object logically from another body. This allows us to build molds around shapes of any kind. I decided to learn how to make these molds by creating one for the orbiter's rudder.

Here's a full 3D mesh of the orbiter - available in many places online for free, or you can purchase one if you want. It's an STL model - which means it's polygonal - described by vertices in 3D space and faces that are combinations of these points.

1673186194503.png

You can then draw a block around any shape you want to mold using the sketch tools and then extrude that shape into 3D.
I've selected the shuttle body here, so you can see the outline of what's "inside the mold."

1673186310875.png

Here's where this gets interesting. You can then "subtract" the shuttle from the mold block using the Combine feature. This allows for the shape to be transferred inside the block. This is a huge time savings.

And here's what the mold looks like when I use a cutting plane to remove areas we don't need to 3D print.
1673186493133.png

I then exported this from Fusion 360 as an STL and printed it on my 3D Printer at high detail mode - at 50% scale - around 6" tall, it printed in 8 hours. In order to do this right, you have to print both halves, but it's quite a step up. In the past I've ignored the full "air foil" of fins, by just beveling the edges, but this would allow anyone to model an airfoil easily.

Have you ever made a mold with Fusion 360? Let me know your thoughts.
 
From Digital to Analog

After sixteen hours of print time, the two molds completed nicely. Lots of sanding to do to make my first part using this approach.
My cutting planes should have been done identically on both halves - doesn't impact the final delivery but would have been a nicer look.
Lesson learned. All those internal "ridge" lines will need to be sanded smooth - as will the prominent rudder hinge.

Update: As many of you that follow my builds know, I really like the services of places like SendCutSend to perform the cutting services for CNC purposes. The economics are quite different in the 3D printing world. In the CNC world, the cost for the wood and the prep time for cutting largely is the same or more expensive than dealing with company like SendCutSend. I requested a quote for this rudder part (just half the mold) and Xometry came in around $380 - delivery in a week. Delivery in 2 days was around $600. Buying a 3D printer with a large print bed (like Creality) is around $500. So, printing your own molds and investing in a larger print bed is a no-brainer.

ED0363BC-31D5-4A0D-B17E-A019AD8E194B.jpeg
 
Last edited:
Back
Top