edwinshap1
Well-Known Member
- Joined
- Sep 17, 2010
- Messages
- 2,131
- Reaction score
- 3
Project 60K Failure Analysis
In May of 2012 Project60k was started as an attempt to break the “N” altitude record, and complete the CTI N5800 challenge. After much design and other collaborative work we had come up with a design that fit the goals and a reasonable timeline for completion. We then set off to find the money and other resources to put a project of this scale together. After speaking with many manufactures and vendors, we put together a full size budget for the project. After receiving many generous donations from vendors, and the community through donations, we were able to begin the build phase.
After leaping that hurdle, it was a matter of time before part started arriving for the project. We ordered nearly every component from small rocketry vendors. The entire rocket is actually made of purpose built amateur rocketry components, with the exception of the fin stock. Widely available body tube, nosecone, and electronics were used in this project.
The fins were designed in Solidworks. The bi-directional airfoil reduced the mass of the fin by half, and helped to reduce vibration at high velocity. The fin is machined from a ¾” thick piece of aluminum, and fillets are milled into the fin. The base of the fin was milled to be rounded so that the ¾” wide base could be seamlessly seated onto a 4” airframe. The fins were surface mounted onto the carbon fiber using Loctite 9430 Hysol Stuctural Adhesive, providing high hardness, tensile, and peel strength. Prior to adhesion, the fins were wet sanded using the epoxy to remove the oxide layer, as well as increase bonding surface area. The body tube and nosecone were coated in this adhesive as well, to be used as an ablative.
 
Points of Failure:
The source of failure was determined to be one of the fins. There was little to no epoxy on the first two inches of the fin, and air was able to lift the fin, peeling the tube apart. The area with no epoxy is circled in red, note that the rest of the fin, where the epoxy did sufficiently bond to the fin and tube, the top layer of the tube is torn off, and the tube and epoxy are still bonded to the aluminum.
After the first fin released from the rocket, it made a sharp turn to the horizontal and both other fins were torn off. It has a large curve on the leading edge from how it was torn, and the trailing edge was bent up into an S shape.
Pictured below is what we determined to be the last fin to be torn from the body tube. Below is the contour on the base of the fin. Note that body tube (black) is still attached to most of the fin. On the right is the shape of the fin post flight. The red line shows the contour of the fin after removal. This was likely the last fin to be torn off as it shows the least stress, so the rocket likely decelerated after the second fin was removed.
The failure of the body tube is an interesting one. We have heard many times over that bonding aluminum to carbon fiber is very difficult due to oxidation, material incompatibility, and a few others. We wet sanded the fins with epoxy to inhibit oxidation, and then cured the fins at 180F to reduce cure time and increase strength. We found that the Loctite resin used to bond the fins has a higher peel strength than the resin used to wrap the tube. It tore large amounts of fibers away from the tube, leaving exposed carbon fiber visible both on the fin bases and the tube itself, as seen in the images above.
Once the first fin tore off, the rest of the rocket quickly followed. The two inches between the top of the motor and the base of the coupler, where the vent holes were drilled, collapsed. That collapse caused the 3/8” threaded rod to snap, leaving the recovery electronics in the open. The electronics were not found.
The nosecone bulkhead was similarly destroyed, but it was not recovered for inspection. The bulkhead, ¼” threaded rod, and nose tip were not recovered, and they probably failed at the same time. We expect that when the bulkhead was stressed, it snapped, causing the nose tip to break its bond from the rest of the nose, leaving the cone and rod behind. The Garmin Astro was recovered intact inside the nosecone, and the Beeline transmitter was recovered 10 feet away, also intact, still recording.
The electronics bay bulkhead, however, was recovered. It was complete with U-bolt, Kevlar, and a CD3 ejection charge, which was never ruptured, though the charges were lost. Shown below with 3/8” rod removed, though they were recovered as once piece.
Additionally, the rocket was coated in the Loctite epoxy for an ablative layer. It was very rippled, and was sanded to a far smoother finish, though it was not smooth in comparison to the tubes found in many high performance applications. When recovered we found that the adhesive bonded so strongly to the nosecone that it actually pulled small chunks from the cone during ascent. We have recovered a filament wound cone recovered after coming in ballistic from upwards of 1000ft, but this pattern is not consistent with cracking from impact, and it was going far slower due to the lack of a tip.
Lastly is the parachute. It was in a tether, which was recovered, but at the high speeds it was released at, it was pulled out almost immediately. The parachute was torn in many different spots, but the most interesting problem happened at the swivel. When shown to Gene of Fruity Chutes we got a reaction that we did not expect. The swivel was rated to 1500lb strength, and it was bent straight. Gene had never seen that kind of damage before, which says a lot about the stresses of this motor.
Possibility for Improvement:
• Reassess our bonding technique, and practice before making another attempt. I believe that if the first fin to fail had been properly glued, it would have survived to the end of the boost, barring a nosecone or coupler failure.
• Determine a better way of drilling vent holes for electronics and pressure. The current method of drilling into composite tubes causes inner layers to be pulled from the wall, reducing strength, and exposing every layer of composite to the supersonic flow.
• Find high quality ablatives to coat the tube and nose with. The Loctite is far too strong, and when it does tear it tears the tube and cone with it. We may use Loctite in acetone for a good ablative due to the lower strength properties when dissolved.
• Find a stronger tube and nosecone. The filament wound tubes are strong, but every fiber is exposed due to the grinding of the tube to create a uniform finish. This leaves the fibers exposed to heat and aerodynamic drag, which reduces the velocity it can be flown in low altitude. We have a composites expert testing pre-preg carbon fiber tubes for durability, and we will run strength tests of our own when the tubes arrive.
• Test all components to flight stresses. This stems from the industry itself. Parts can be remade to close tolerance, but once a part is destroyed the issue can be far harder to resolve. We intend to test every part we receive to the maximum stresses it will encounter during the flight. If parts are broken they can always be replaced, and it will be far cheaper than rebuilding the rocket from the ground up.
 
Final Notes:
According to the rocket’s purpose, this flight was a failure, as we did not make it to 60000 feet, and we did not recover successfully. However we proved some hypotheses that had been doubted for a long time. We flew surface mounted aluminum fins to mach 3 at 10000ft MSL before a failure, though we believe that the failure was due to reduced adhesion on the leading edge of the failure point, not a failure of the fins or of the adhesive themselves.
It is also important to note that the epoxy did hold the aluminum to the tube, even when the fins were no longer attached to the body. This shows that bonding aluminum to carbon fiber does work as intended. Note that we did groove the base of the fin for increased adhesion, and we did mill the fins from ¾” plate to get aluminum fillets on both sides of the fin without welding and compromising the annealing properties.
We would like to thank all of our donators, as well as Wildman Hobbies, Rouse-tech, Fruity Chutes, CTI, Tender Rocketry, Whats Up Hobbies, Beeline, and any others I am forgetting at this moment.
Also, a special thanks to Kurt Gugisberg, who signed off on our flight, Tony Alcocer for his advice and assistance, and finally Aeropac for allowing us to launch.
Additional Materials:
High Speed Liftoff
[video=youtube;mBeYUV8wlW8]https://www.youtube.com/watch?v=mBeYUV8wlW8[/video]
Video by Carlos Rodriguez-Santiago
Launch View 1
[video=youtube;mrGn2gpkdGE]https://www.youtube.com/watch?v=mrGn2gpkdGE&feature=plcp[/video]
Video by Jared Shapiro (edwinshap1)
Launch View 2
[video=youtube;GdJAiXuDnbo]https://www.youtube.com/watch?v=GdJAiXuDnbo&feature=player_detailpage#t=29s[/video]
Video by Carlo Vaccari (Carvac)
PDF file of report below.
View attachment Project 60K Failure Analysis.pdf
Last edited: