“Okay, Houston, we’ve had a problem here.”
With this blog post we want to review our launch last Wednesday. Several very impressive engineering feats were performed before final take-off. We would like to give some more in-depth information about the problems encountered and the creativity that let us solve them. “You can’t have a perfect launch” seems to apply to both Stratos II and Stratos II+ and definately also Stratos III. Regardless of the final outcome, these feats are something to be proud of. We’ll go a little more in depth on these as it is cool and there are some clear things to learn.
As is now public knowledge, our rocket took of at around 3:30 a.m. but suffered an anomaly around 20 seconds into the flight. Here starts the long process of looking at our data and trying to find the root cause, the smoking gun, that quite literally broke our dream is many small pieces. One of the things that we really appreciate about DARE is the willingness to communicate about the failures. This is quite rare in the aerospace industry but results in a lot of recognition from other students and engineers around the world (this is a hint for some of the other student teams). This leads to some great exchanges and is a comfort even when things don’t go as planned. Lastly, we will finally provide some data and look ahead to the future. It really isn’t all that bad.
During the first attempt on Tuesday we saw that the high altitude winds exceeded the allowable margin. We decided to pack up and put everything back in place for a new attempt the next day. We needed to recharge several batteries and put a new CO2 cartridge in the rocket that is needed for the drogue parachute deployment device. In the end Tuesday turned out to be a nice practice for the actual launch since going through the procedure went really smooth. Everybody was positive about the second attempt that was done on Wednesday.
Pre-flight assembly (or checklist Alpha) started around 9 a.m. and we were ready to roll out to the pad just after lunch at 14:00. This meant we had a lot of slack in our schedule for the actuator checkout and attaching the 9 bottles at our filling station. However, just after the flight termination system checkout was completed, a problem was encountered. One of the engineers noted that the pressure sensors (there is one in the combustion chamber and in the tank) showed faulty measurements.
A pressure sensor transfers the pressure on a membrane to a small current. The problem was that even though a current was measured it did not show any kind of noise. Typically there will always be some sensor noise either from small pressure fluctuations or electronic noise from the systems around it. A few fluctuating digits at the end is thus a good sign. Launching without working sensors was not an option, as these measurement devices are critical during filling operations for tank pressure monitoring. Therefore, the countdown had to be stopped and the rocket partially disassembled to solve the issue. After the problem was fixed, Stratos III was loaded back in the tower and the countdown was resumed. Unfortunately, the faulty sensor readings and the additional work that our engineers had to do to fix it, delayed the planned lift-off from 9 pm to 11 pm.
Our rocket is too big
During filling the second problem poped up. Again a sensor stopped changing its value which made it impossible for us to know how much oxidizer was already in the tank. A loadcell is mounted at the rocket base that measures how much N2O has been loaded onto the rocket. At 150 kg of N2O it however stopped. It took a quick few minutes for one of the engineers to register that the limit of the measurement system was exceeded with the full weight of the rocket. In order to make sure the tank could still be filled to a safe level, a infrared camera was used. With the camera the engineers could see through the outer layer of the tank and watch the level of liquid oxidizer in the tank.
Even with this method, the tank was slightly overfilled which meant the oxidizer has to be bled away. Otherwise, the rocket would be too heavy and not go as high as planned. This is why there was vapor coming out of the rocket, as you could see in the live stream. After the mass of oxidizer in the tank reached the desired value, the team began pressurizing the oxidizer with infrared lamps. This took way longer than expected as it was a new method of pressurizing the tank, which caused a significant delay. In the end, the rocket lifted off at 3:30 in the morning instead of 11 pm.
The rocket trajectory was a straight line from the 85 degrees launch rail and fit our simulations very well. The anomaly occurred at around 20 seconds into flight. Stratos III was flying at around 11 km with a speed of nearly 900 m/s or close to mach 3 at the time of the anomaly. In addition the failure occurred so fast that data transmission was cut of almost immediately. Some sideway 2g acceleration was registered by the on-board inertial measurement units however this still needs to be confirmed as the rocket was also spinning slightly.
So what do we know and the way forward
We are sure that during the first 20 seconds the rocket performed nominally. We are very lucky with our launchsite as they have very high quality radar systems and infrared optronics systems to give us a suggestion about what happened. We were only able to get the footage today and even more is expected to come in during the next few days.
Forensic engineering is a master course at the TU Delft because it is hard. To draw evidence based conclusions a full reconstruction needs to be made that combines all the different types of data. A timeline can than be made of the events that led to the failure. This will take several weeks in the coming months.