Launch 5

We launched LV2 with a complete avionics system over 5,000 meters on August 20th 2005, outside of Brothers Oregon.


Launch 5 was our “successful failure” launch. We had a great team, and a really well prepared rocket. We had a smooth countdown and launch. This was our first flight with 802.11 WiFi live telemetry from a linux flight computer. We got tons of data. The only thing to go wrong was the parachutes not opening!


For the first time in PSAS history we had a real flight computer. This was a x86 processor in an embedded format (PC104 form factor) running linux. We had more than enough computational power to stream and process all the sensor data in real time on the rocket. Since this was out first try, we mostly just recorded and streamed. It’s lucky that we were able to stream the data because this is one flight computer that we didn’t get back in one piece.

All the data we have from this flight was streamed over a simple WiFi card to the ground. As far as we know we’re the first amateur rocket group to fly WiFi on rockets and to have such a sophisticated avionics stack.

Download all the raw data and analysis python notebooks as a zip file, or view on github.


Our Inertial Measurement Unit (IMU) is a combination of accelerometers and gyroscopes in each direction $(x,y,z)$. We can use it to measure how much the rocket accelerates and how fast it rotates. This can also tell us about the speed and altitude of the rocket.


We start off looking at the vertical component (in body frame) of the measured acceleration during the flight, the accelerometers in the $x$ and $y$ (sideways) direction. The raw accelerometers are quite noisy, passing through a decimation filter helps.


A single integration of the vertical acceleration will give approximate velocity. There will be some drift as the tiny errors in the IMU readings slowly add up over the integration.

The peak velocity during the motor burn was 406 m/s (Mach 1.2). The final velocity before we hit the ground is -235 m/s (Mach -0.7).


Further integrating the velocity numbers should get us altitude. Hopfully we can fix this with filters that take into acconut several sensors, like pressure and GPS fixes, in order to de-bias our IMU drift.

According to the IMU the apogee was 5.56 kilometers. And then we “landed” at an altitude of 138 meters.


The rocket has a Conexant Jupiter GPS, a 12-channel OEM receiver based on the Zodiac chipset. It was set to send 1 Hz binary data.


We can compare the GPS reported altitude to the IMU integration.

There does appear to be a time lag for some reason between the two.

The maximum GPS altitude recorded was 5.53 kilomters. (IMU: 5.56 km). Though there was some forcing in the IMU calibration to match this numer. The final data point from the gps is 332 meters. At that time the integrated IMU reports 204 meters, a close match!


A similar story for velocity. There is also good agreement here, again with a time lag. Also the peak of the velocity curve appears missing from the GPS. This might have something to do with the ability to do fixes under high-g loads.

We continue to get IMU data for 0.28 seconds after the last GPS data point.