1/26th Scale Space Shuttle (full stack)

[bguffer]

Maybe this is useful, maybe not. 1/72nd prototype north coast rocketry orbiter build, with some detailing swapped from plastic model kit:


At least shows how internals were beefed up a bit, but for considerably smaller scale.

Once i get video/photos of launch, full build video will become available. Full build video will show SRB detailing, and painting/decals based off https://axmpaperspacescalemodels.com STS-5.

Bob

 

[burkefj]

Maybe this is useful, maybe not. 1/72nd prototype north coast rocketry orbiter build, with some detailing swapped from plastic model kit:


At least shows how internals were beefed up a bit, but for considerably smaller scale.

Once i get video/photos of launch, full build video will become available. Full build video will show SRB detailing, and painting/decals based off https://axmpaperspacescalemodels.com STS-5.

Bob

That's a new prototype north coast might kit? 12 oz for the orbiter sans paint?

 

[bguffer]

That's a new prototype north coast might kit? 12 oz for the orbiter sans paint?

old prototype. got it back in 2010 from gassoway auction. 12oz without paint, and with some plastic model parts, and with the nose weight added.

 

[sambatterman]

Using Fusion 360 to think and plan your parts

As I mentioned a few threads back, I'm learning Fusion 360 for fabricating many of the parts of the Space Shuttle stack. One of the things that I really love about the product is that it helps me think through designs that otherwise I would "bumble" my way through and waste materials and time.

As an example, I'm using Fusion 360 to work out how to use traction for retaining the SRBs to the External Tank and also how to separate them.

For reference, here's a George Gassaway (Copyrights to George) drawing showing the scope of what I'm working on...the SRB nose cone and the adapter, which contains the separation mechanism.

Here's a perspective view of the entire component so far...


Each of these parts can be hidden and worked on individually. So, let's look at just the retention mechanism: a flange embedded in the SRB coupler.



You'll notice the 1/8" hole for a solenoid plunger to pass through inside the ET and three 1/4" pin holes for aluminum pins inside the adapter. Let's take a peek at that...notice the top stop has an "inset" hole that contains the flange internally.


And here's the top stop with the flange embedded and pins installed...


Now, let's connect the bottom stop, which has pin holes for registration to the flange. This entire assembly would be epoxied together, forming a very strong point to retain the front part of the SRB to the ET.



This entire stop would then be epoxied inside the SRB adapter like so...



and the nose cone attached to the front of the adapter.


Some lessons I've learned doing this:

Separate all the components from one another - don't try to draw one big drawing. You'll notice the SRB nose cone is "linked" in from another drawing all together. This helps in other ways - the relationship of parts to one another and ensuring the scale is correct.

I'll be working on the lower retention component and the solenoid flange inside the ET later this week but wanted to show some progress. What are your experiences with using Fusion 360 - not just for printing parts, but the overall design of the rocket or its sub-assemblies.

 

[Raketenolli]

Amazing project!

A few remarks from my experience with Shuttles:
- The full stack is stable without any extra fins with sufficient noseweight - as in, at least as much as the stack weighs empty. Would not recommend for a project your size ;-) I did fly a 1/100 model for a science fair project in 2001.
- Rolf Stabroth flew a full stack successfully a number of years ago with passive stabilization: https://rollirocket.raketenmodellbau.org/Shuttle Startvorbereitung & Flug.html Takeoff mass was about 15 kg IIRC

I would suggest you go for lighter construction for ET and boosters, just like you are planning to do with the orbiter. Use a Sonotube or 11.4" PML tube as a core and make a fiberglass-balsa or carbon-balsa sandwich tube. Then go for low thrust, long burn motors. I don't have much experience with active stabilization, but going from what Joe Barnard and Joachim Lang have done, I think you'd want a thrust:weight ratio just a bit above 1. Unless of course you've already made up your mind and are definitely going for fin stabilization for the full stack and "only" autonomous landing of the orbiter.

One caveat I can think of with the IMU: if you have too much drift between takeoff/initialization and separation of the orbiter, it might not know well enough where "up" is. I'm not sure you can figure it out with only the IMU once you have separation. All the other sensors seem quite useful.

Oliver

 

[sambatterman]

SRB Full Body Design

This week I worked on the full body design for the SRBs and how to connect them securely to the External Tank (ET).
Here's a bird's eye view:


Just like the real thing, there are two attach points to the ET - one near the nose and one aft. Both will use solenoids and flanges.


And here it is nestled up against the ET. I'm still working on a number of things --- nothing's ever finished here, but wanted to share some progress.

 

[sambatterman]

Meeting the IMU

I've been working on getting an IMU to tell me its secrets and I'm finding some success. I decided to use the Adafruit IMU BNO055 Sensor because it's cheap ($25-$30) and easy to program. An IMU contains an accelerometer, gyroscope and magnetometer - and those devices give you position in 3D space. This will be super important for this project for determining boost, release from the ET, re-orientation (getting right side up after release from the ET), heading for waypoints along the vector to landing (along with the GPU) and probably most importantly: auto-pilot - keeping the craft centered and avoiding stalls.

Here's the hardware set up I'm using and the code to talk to the sensor. There will be a lot more about all of the sensors being used as I move forward, but I was excited to see data come streaming by after about 10 minutes of wrangling with it.

 

[sambatterman]

Early look at the External Tank

Still working through the design on Fusion 360, but thought I would share these cool, two views of the orbiter and the external tank.
The ET is an 11.41" x 48" phenolic tube, the nose cone is 19" tall and the tail cone is around 4.1".

 

[Raketenolli]

Looks like your design is coming along nicely.

How are you planning on using IMU data for pitch control in the autopilot? Angle to the gravity vector? Or are you going to add another sensor like dynamic pressure or angle of attack?

Oliver

 

[sambatterman]

Looks like your design is coming along nicely.

How are you planning on using IMU data for pitch control in the autopilot? Angle to the gravity vector? Or are you going to add another sensor like dynamic pressure or angle of attack?

Oliver

Hi Oliver!
Thanks for your question. While I'm still working through everything, I currently have pitch as an angle and roll as an angle, not yaw (yet). The IMU I'm using has an accelerometer, a gyroscope and magnetometer on board. I'm using the accelerometer and the gyroscope to cancel noise along with a low-pass filter. I have an air pressure sensor as well. As I said, still working through all of this. Here's the code I'm using on the Arduino:

Code: #include <Wire.h> #include <Adafruit_Sensor.h> #include <Adafruit_BNO055.h> #include <C:/Users/Sam Batterman/Documents/Arduino/libraries/Adafruit_BNO055/utility/imumaths.h> #include <math.h> float thetaM; // Theta = Pitch float phiM; // Phi = Roll float thetaFnew = 0; // New Theta Filtered Value (Low) float thetaFold = 0; // Old Theta Filtered Value (Low) float phiFnew = 0; // New Phi Filtered Value (Low) float phiFold =0; // Old Phi Filtered Value (Low) float thetaG = 0; // from Gyro float phiG = 0; // from Gyro float dt; unsigned long millisOld; #define BNO055_SAMPLERATE_DELAY_MS (100) Adafruit_BNO055 myIMU = Adafruit_BNO055(); void setup() { // put your setup code here, to run once: Serial.begin(115200); myIMU.begin(); delay(1000); Serial.println("*************************************************"); Serial.println("* Booting up IMU Sensor *"); Serial.println("*************************************************"); int8_t temp = myIMU.getTemp(); Serial.println("Current Temperature"); Serial.println(temp); myIMU.setExtCrystalUse(true); millisOld = millis(); } void loop() { uint8_t system, gyro, accel, mg = 0; // Calibrate the Sensor myIMU.getCalibration(&system, &gyro, &accel, &mg); // Read Vectors of IMUs accelerometer, gyroscope and magnetometer imu::Vector<3> acc = myIMU.getVector(Adafruit_BNO055::VECTOR_ACCELEROMETER); imu::Vector<3> gyr = myIMU.getVector(Adafruit_BNO055::VECTOR_GYROSCOPE); imu::Vector<3> mag = myIMU.getVector(Adafruit_BNO055::VECTOR_MAGNETOMETER); // Tilt on X-Axis // Normalize to Units of 1 G (9.8) // to degress instead of Radians thetaM = -atan2(acc.x()/9.8,acc.z()/9.8)/2/3.141592654 * 360; // Tilt on Y-Axis // Normalize to Units of 1 G (9.8) // to degrees instead of Radians phiM = -atan2(acc.y()/9.8, acc.z()/9.8)/2/3.141592654 * 360; // Filter the results; phiFnew = .75*phiFold+.25*phiM; thetaFnew = .75*thetaFold+.25*thetaM; Serial.print("Phi: "); Serial.print(phiM); Serial.print(" (Filtered: "); Serial.print(phiFnew); Serial.println(")"); Serial.print("Theta: "); Serial.print(thetaM); Serial.print(" (Filtered: "); Serial.print(thetaFnew); Serial.println(")"); // Gyro calculations // Time to seconds... // Calculate between measurements... dt = (millis() - millisOld)/1000.; millisOld = millis(); thetaG = thetaG + gyr.y() * dt; //phiG = phiG + gyr.x() * dt; // reverse roll phiG = phiG - gyr.x() * dt; Serial.print("Gyro Theta: "); Serial.println(thetaG); Serial.print("Gyro Phi: "); Serial.println(phiG); // Update the Low Pass Filter... phiFold = phiFnew; thetaFold = thetaFnew; delay(BNO055_SAMPLERATE_DELAY_MS); }

 

[Raketenolli]

Ok, so currently you're integrating the values from the gyroscope to obtain the attitude, but you're not fusing those values with the accelerometer output in any way? From what I understand the accelerometer only gives you valid attitude information when you're in stabilized flight. I'm mostly concerned about the phase between separation of the orbiter and gliding descent. You need to make sure it will, without any control input (because you won't know where up or down is at that moment), enter a stabilized glide and then can you start controlling airspeed through attitude.

Have you verified that your sensor's axes are well-to-perfectly aligned with its outside geometry/the mount/the orbiter? That is an issue with some sensors.

Oliver

PS: I see you are an MDRA member, looking forward to seeing this in real life sometime!

 

[sambatterman]

Ok, so currently you're integrating the values from the gyroscope to obtain the attitude, but you're not fusing those values with the accelerometer output in any way? From what I understand the accelerometer only gives you valid attitude information when you're in stabilized flight. I'm mostly concerned about the phase between separation of the orbiter and gliding descent. You need to make sure it will, without any control input (because you won't know where up or down is at that moment), enter a stabilized glide and then can you start controlling airspeed through attitude.

Have you verified that your sensor's axes are well-to-perfectly aligned with its outside geometry/the mount/the orbiter? That is an issue with some sensors.

Oliver

PS: I see you are an MDRA member, looking forward to seeing this in real life sometime!

HI Oliver, still working through everything, but yes, will be using both the accelerometer and gyro. Gyro only gives velocity changes, so that has to be resolved - which is what the code above does. I'll have an update this weekend. Yes, the sensor will be mounted appropriately/accurately in the orbiter. Thanks for the feedback and ideas.

 

[sambatterman]

Determining release from External Tank using an Ultrasonic Sensor

Using an ultrasonic sensor, we can easily determine when the orbiter is clear of the External Tank, or even if the SRBs have jettisoned correctly. In this case, I envision using the sensor on the orbiter belly, just below the nose gear pointing directly at the ET. This would be a distance of 2-3" during boost. At the appropriate time, a solenoid will release two pins - one fore and one aft on the orbiter, allowing the orbiter free from the attachment points on ET.

This sensor is really cheap - 5 sensors for $12 on Amazon.

Here's a code example. And pictures of the sensor and the graph of what happens during a separation are shown below.

Code: // Define Trig and Echo pin: #define trigPin 2 #define echoPin 3 // Define variables: long duration; int distance; void setup() { // Define inputs and outputs: pinMode(trigPin, OUTPUT); pinMode(echoPin, INPUT); //Begin Serial communication at a baudrate of 9600: Serial.begin(9600); } void loop() { // Clear the trigPin by setting it LOW: digitalWrite(trigPin, LOW); delayMicroseconds(5); // Trigger the sensor by setting the trigPin high for 10 microseconds: digitalWrite(trigPin, HIGH); delayMicroseconds(10); digitalWrite(trigPin, LOW); // Read the echoPin, pulseIn() returns the duration (length of the pulse) in microseconds: duration = pulseIn(echoPin, HIGH); // Calculate the distance: distance = duration * 0.034 / 2; // You can also use distance = sonar.ping_cm() or distance = sonar.ping_in() // which returns the measured distance in whole centimeters or inches. // With this function you do not need to take a duration measurement and calculate the distance. // Print the distance on the Serial Monitor (Ctrl+Shift+M): Serial.print("Distance = "); Serial.print(distance); Serial.println(" cm"); delay(50); }

Here's the sensor on the breadboard and its view is occluded by this book - constant measuring of 3"



Removing the book quickly, you can see the sensor is now monitoring distances greater than 3" - indicating separation.


As part of the flight software, this would break the orbiter from a state machine of "Boost" to "Recover" - ensuring the orbiter is right side up and beginning to orient itself to the landing area.

 

[sambatterman]

Controlling 16 servos

Okay, so my 16-channel servo driver showed up and it's quite easy to drive 16 servos from a single Arduino Uno. Basically, you tell the driver board, which channel (servo) and what angle to move the arm. This will be useful for the outer and inner elevons (4), the rudder, the gear doors (6) and the landing gear (3).

 

[ReynoldsSlumber]

Having an unpowered orbiter would be scale... for Energiya/Buran

 

[Ez2cDave]

Having an unpowered orbiter would be scale... for Energiya/Buran

 

[Art Upton]

But it was JET powered on reentry / landing

 

[Ez2cDave]

But it was JET powered on reentry / landing

Art,

A very interesting point . . .

Being turbojets and not ramjet engines, they are not have been started at extremely high airspeed, certianly not during re-entry.

I'm assuming that Buran was "deadstick", until lower altitudes and airspeeds were reached . . . Does anyone know at what point the jet engines were started ?

Something I never knew . . . A test-version of Buran was capable of lifting off from a runway ! https://en.wikipedia.org/wiki/Buran_programme

QUOTE :

"A test vehicle was constructed with four jet engines mounted at the rear; this vehicle is usually referred to as OK-GLI, or as the "Buran aerodynamic analogue". The jets were used to take off from a normal landing strip, and once it reached a designated point, the engines were cut and OK-GLI glided back to land. This provided invaluable information about the handling characteristics of the Buran design, and significantly differed from the carrier plane/air drop method used by the United States and the Enterprise test craft."

END QUOTE :

https://en.wikipedia.org/wiki/OK-GLI

http://www.primeportal.net/space/hans-hermann_buhling/buran_ok-gli/index.php?Page=4

Dave F.

 

[alexzogh]

Determining release from External Tank using an Ultrasonic Sensor

Using an ultrasonic sensor, we can easily determine when the orbiter is clear of the External Tank, or even if the SRBs have jettisoned correctly. In this case, I envision using the sensor on the orbiter belly, just below the nose gear pointing directly at the ET. This would be a distance of 2-3" during boost. At the appropriate time, a solenoid will release two pins - one fore and one aft on the orbiter, allowing the orbiter free from the attachment points on ET.

This sensor is really cheap - 5 sensors for $12 on Amazon.

Here's a code example. And pictures of the sensor and the graph of what happens during a separation are shown below.

Code: // Define Trig and Echo pin: #define trigPin 2 #define echoPin 3 // Define variables: long duration; int distance; void setup() { // Define inputs and outputs: pinMode(trigPin, OUTPUT); pinMode(echoPin, INPUT); //Begin Serial communication at a baudrate of 9600: Serial.begin(9600); } void loop() { // Clear the trigPin by setting it LOW: digitalWrite(trigPin, LOW); delayMicroseconds(5); // Trigger the sensor by setting the trigPin high for 10 microseconds: digitalWrite(trigPin, HIGH); delayMicroseconds(10); digitalWrite(trigPin, LOW); // Read the echoPin, pulseIn() returns the duration (length of the pulse) in microseconds: duration = pulseIn(echoPin, HIGH); // Calculate the distance: distance = duration * 0.034 / 2; // You can also use distance = sonar.ping_cm() or distance = sonar.ping_in() // which returns the measured distance in whole centimeters or inches. // With this function you do not need to take a duration measurement and calculate the distance. // Print the distance on the Serial Monitor (Ctrl+Shift+M): Serial.print("Distance = "); Serial.print(distance); Serial.println(" cm"); delay(50); }

Here's the sensor on the breadboard and its view is occluded by this book - constant measuring of 3"

View attachment 562341

Removing the book quickly, you can see the sensor is now monitoring distances greater than 3" - indicating separation.
View attachment 562342

As part of the flight software, this would break the orbiter from a state machine of "Boost" to "Recover" - ensuring the orbiter is right side up and beginning to orient itself to the landing area.

I tried using an ultrasonic sensor (HC-SR04) a few years ago for something very similar - worked perfectly on the bench and in LPR testing. Once I went to HPR, I could never get it to work - the laminar flow around the rocket created a turbulent-boundary layer flow right at the sensor creating what looked like static (white noise) in the readings. No amount of filtering/Kalman smoothing could pull the signal out of the noise.

I switched to a laser rangefinder - TOF10120. You can find it at the usual places for around $10.

Pro's
Much faster response times (light vs sound)
immune to wind, noise, and vibration
completely sealed from the environment (vs. the SR04)
about a quarter of the size making it easier to place.
Can use 3.3v or 5v

Con's
More expensive by about $8.
Narrower field of view (which might be a + for you)



There are lot's of libraries out there for the TOF10120, it's a staple of the drone community

 

[burkefj]

But it was JET powered on reentry / landing

As Dave said or was implying the jet engines were used only for testing purposes those wouldn't have been used and weren't used on the orbital vehicle.

 

[sambatterman]

Reference Material

A turntable view of the Polecat Aerospace/Andy W. Space Shuttle Stack


And the Ames Research Center Modelling Lab (circa mid-1970s).

 

[sambatterman]

Adventures in mold making - Part 1

It's been a minute since I posted about this project. Work was really complicated this year and I could only focus on a limited number of complicated things. One of the determinations I decided on back in January was that the Orbiter would need to be mostly composite - not bulked (foam/plywood sandwich) up construction like the Nike, ALCM and Jayhawk (my newest build). It needed to be super light, super strong and repeatable. By repeatable - and I'm saying this out loud - I will need to have many orbiters to complete this project - they will crash as I learn how to land it autonomously.

So, I'm getting into mold making in a big way. I've made simple fiberglass/carbon fiber molds for hatches, but nothing with compound curves - and certainly nothing approximating the complexity of airplane parts.

First step - start simple. Printed a complex compound curve shape on my 3D printer - it took 4 days. The shape was printed at high detail in order to have small preparation before casting the mold. I sanded it down with 80 grit to get all the "frizzles" off the surface and then sprayed it with normal primer (like you would before you paint any rocket). Sanded it with 120 grit. Notice in the picture below the print anomaly "line" on the right side of the curve. This would normally be taken care of with putty. But what kind of putty? Bondo? SuperFill?

I was in a hurry to get results, so I skipped the putty stage.




One of the best choices and discoveries I made was buying a mold release that is easy to use. This is from Fibreglast and you just wipe it on the mold surface with a towel. Wait five minutes and you're ready to go.



I used two layers of fiberglass and for the first time used a blot sheet - you normally see these with vacuum bags, but I decided to try it out.
I used West Systems and let it dry overnight - 12 hours to cure.



Next morning, I pulled the blotter off - came of easily - love that stuff and it's cheap in comparison to fiberglass and carbon fiber. It's worth it.
The fiberglass shape literally popped from the mold with just a little effort.
Since it is fiberglass and only two layers, it's flexible, but holds the basic shape, but it's not strong rigid.
Notice the anomaly transferred to the mold.

Next step: adding Carbon Fiber on a more complex shape. I will also recast this one with Carbon Fiber to see the strength difference.

 

[Blast it Tom!]

Great Scott, what a project! I just found it. And I had to skip page 2 for now. A tremendous engineering effort on all fronts, wish you all the best. And I agree with what you said in the OP about the breadth and depth to this hobby/sport, and the revolutionary effect of electronics. Go man go!

 

[sambatterman]

...And the Ames Research Center Modelling Lab (circa mid-1970s).


By the way, the Ames model looks to be about 1/26th scale - maybe a little smaller.

 

[sambatterman]

Adventures in Mold Making - Part 2

Okay, with my first casting under my belt, I decided to do something specific to the Orbiter - the cockpit.
Using Fusion 360, I drew a block around the cockpit of the shuttle model and positioned it at exactly half point longitudinally - as if a knife was sliced down the center point of the shuttle. I then used the combine tool to remove the shuttle from the block - this forms half the mold. Printed it to my 3D printer as an STL. That print takes two days at high detail.


... here's the analog counterpart:



Like the first part, I sanded the inside with 80 grit to remote the "frizzles" and then painted it with a coat of gray primer. No putty.



Put the mold release on the surface and let it dry for a few minutes and then put a layer of fiberglass as the outside skin, followed by Carbon Fiber and the blotter.

The Carbon Fiber was very rigid, and I could see that covering the areas like the tip of the nose and the windscreen was going to be problematic. I cut the Carbon Fiber into smaller strips and laid them up against the fiberglass and overlapped them slightly. Each layer was wet out with West Systems. I let it dry overnight (12 hours).



The next morning, I pulled the blotter off - that is something that's a winner every time. And carefully worked the sides of the shape until it popped free of the mold. I did not trim the edges yet - and I probably should have done that in leather stage to avoid hard work afterward.



Overall, I love the result - it's strong - much stronger than the first part. To complete this part, I am printing the mirror image of the cockpit mold. My 3D printing software lets met click "Mirror" and reflect the shape whichever direction I want, so that's easy. Waiting two days for the result is not easy.

Next up is printing the rudder - an airfoil shape.

By the way, these shuttle parts are scaled to 65% of the original size because my 3D printer doesn't have the volume to do the full size. I will have to buy a new 3D printer to do the orbiter project. Until then though, learning about these molds and getting the technique down is worth it.

From a basic back of the envelope calculation - it will take two months to print the shuttle orbiter molds. Building it by plywood bulkhead and foam sandwich would take at least that long, be much heavier and not repeatable - at least not economically. Mold is the way to go.

This is the way

 

[tsmith1315]

Very cool build!

Did you consider positive alignment bits to have the two nose molds fit together to help with assembling the two halves?

 

[sambatterman]

Very cool build!

Did you consider positive alignment bits to have the two nose molds fit together to help with assembling the two halves?

Yep, that's next...

 

[sambatterman]

Putting the molds together

I put the two carbon fiber molds together to experiment with the adhesion side of this problem. It's not perfect, so I'm putting Bondo putty on the whole seam and will sand it down this week.

To attach the two sides, I epoxied a carbon fiber ribbon all along the inside seam of one side. Then fit the two sides together and used tape to register everything as I epoxied the opposite side of the carbon ribbon and pressed it down inside the shell of the cockpit.



I applied Bondo putty and sanded it down the center of the mold fit and then used SuperFill for the surface imperfects.
Primed and ready to wet sand.

 

[sambatterman]

Planning the Body Flap

I'm waiting for my new 3D printer to show up. It was clear I needed a much bigger volume to do this project AND there was a sale!

So, in the meantime, I'm working through airfoils on this model. As I've said before, it's a rocket (actually three rockets - ET and two SRBs) and an airplane. I can't use the techniques on the orbiter that I used on rockets (some might say I build furniture that flies).

Starting with the body flap because it's a simple wing shape.


While covering will be carbon fiber on the wings, the ribs and spars will probably be 3D printed or use 1-8" to 3/8" bass wood. Ribs and spares are not on my radar, and I have to make my mind go there.

So, Fusion360 to the rescue! It's a great tool to help you plan.

First, I measured out what the body flap would need to look like. With three holes - one for the pivot point and connection to the aft fuselage and two for internal support spars (1/8" diameter). I chamfered the angles, and it sure makes the model nicer.


Next, I used the rectilinear pattern tool and projected this body 8 times.
Also, added the spars and connecting rod on the back. The total width of the body flap is roughly 8".



For covering the surface, I decided to use the Ramy RC technique of a single layer of carbon fiber on a flat surface and then cut to fit.
A cool feature of Fusion360 is you can apply whatever material you want to the shapes. Here's a rendering with the top and bottom surfaces covered in carbon and a 3D printed backplate.


Once I tackle the shell of the orbiter the wings will have to get figured out and that's a place I plan to spend a lot of time working.

My goal is to have the orbiter weigh only about 12 pounds - that includes electronic retracts. At nearly 55" long that means that carbon fiber and aramid honeycomb will be necessary materials to avoid heavy bulkheads (although there will have to be a few internally).

 

[Blast it Tom!]

I might like a 3D printer that could do that well. Do you mind divelging what you bought? I'm a complete n00b, but I'm trying to learn and evaluate!