Rocket Engine Monitoring

by James O'Keeffe

Latest Updates...    

Modifications after the second static fire test


First flight test




The wireless vehicle health monitoring project was a great a opportunity to combine things I enjoy doing, programming microcontrollers, pcb layout and SLA rapid prototyping with my long held passion for rockets (or anything fast!). In the process, I learned a great deal about low power wireless sensors and debugging electronics in large vehicles. The system first flew in September 2007 onboard the Garvey Spacecraft Corp. P-8A rocket in Mojave, CA.


What is Vehicle Health Monitoring for Rockets?



Unlike other mechanical systems, rocket engines rarely fail gradually. When rocket engines do malfunction sensor data provides important clues about the cause. The vehicle health monitoring system relays pressure, temperature, voltage, strain and acceleration data back to the Mission/Launch Control center. Integrated Vehicle Health Monitoring (IVHM) goes a step further by providing onboard processing capability. Hence an IVHM system can often detect engine anomalies earlier and respond faster than a ground-linked system.

Goal: The wireless IVHM project replaced the standard MIL-STD-1553 databus with an 802.15.4 wireless link between groups of sensors and the Stargate flight-data-recorder. This system is being used as a platform to demonstrate intra-vehicle wireless transmission and power management software for long duration missions.






Components: For the current system we fabricated wireless pressure sensors (1 on the engine chamber, and 1 on the propellant tanks), 4 wireless accelerometers distributed throughout the vehicle and 2 thermocouples. The sensors were connected to Crossbow MicaZ radio transceivers. These small and versatile radios contain an ATMega128 microcontroller and provide power/control to the sensors. All sensors transmitted to a Stargate flight-data-recorder over an 802.15.4 link.





Radio Transceivers: Six 2.4Ghz Crossbow MicaZ radios, each containing an Atmel microcontroller, CC2420 zigbee transceiver and 512K serial flash memory, Typical battery life of 15hrs @ 100% duty cycle.

Custom Sensor Boards: Four layer Millspec boards, 2 wireless pressure gauges (1 on the engine chamber and 1 on the fuel tank), 4 accelerometers (Analog devices 2-axis 18g and 3-axis 3g),2 thermocouples (one on each of the propellant tanks).

Firmware: We used necC to program the Atmel microcontrollers. NecC is very similar to regular C but the code linking is optimized for easy access to the Chipcon2420 transceiver chip.

Flight Data Recorder: The Crossbow Stargate single board linux computer was a good choice because it is designed to interface with the Micaz radios. The Stargate is extremely rugged and features a 400Mhz Intel PXA255 ARM processor, 2Gb flash hardrive, Ethernet, Serial, JTAG, USB. We powered the Stargate with a 6.2V, 2800mAh lead-acid battery.




Pictures and Highlights  



Installed pressure sensor (May 2007)

SLA enclosures: After the first static-fire test it was obvious we needed a better way to mount the accelerometers to the wood bulkheads of the P-8A. I designed a case in Solidworks and had it fabricated by Advanced Prototyping Inc. The SLA material was ideal for prototyping but was far more brittle than standard ABS - the solution was to use a 5mm wall thickness.

Power Management Software (modification after the 2nd static-fire test)

Remote control power relay (modification after the 2nd static-fire test)

Voltage Regulators: Amazingly, Crossbow's micaZ module do not have an onboard LDO voltage regulator. This means that the reference voltage to the ADCs drifted over time. This was a particular problem for the accelerometers, which are not ratiometric and output a constant 1.42V at zero-g. The solution was to build an LDO voltage regulator onto the sensor board. Later we regulated the whole system by routing the battery through this regulator first.

Stargate Software: One of our summer interns was very skilled at compiling the Xlisten datalogging executable on the Stargate. Mike also wrote the bootup script.




Shock/Vibration Testing



A common failure mechanism for payloads is shock and vibration during launch. Here we tested the Stargate computer on a vibration table (electromagnetic ram) to 6.5g rms for 30 seconds on the X, Y and Z axes. This particular test mimics conditions during the space shuttle launch.

Note the serial cable: The Stargate was live and recording during this test.



Mike and Jon receiving instruction from the test coordinator Jerry Wang


Static-Fire Testing, August 2007





UHF Remote power switch



We learned a few important lessons from the static-fire tests:

Remote Switch: We add a UHF automotive power switch(shown in the picture). The unit allows the payload to be powered on/off remotely between tests and is similar to the remote locks on a car. For increased reliability, the switch was modified by removing the EM relay and replace it with a power transistor.

Audio Feedback: We added a piezoelectric buzzer to the Stargate and each sensor board. These provided an easy way to perform diagnostics at the test range.

Power Management Firmware: An important requirement was to put the radio transceivers into a low power mode when the flight-data-recorder is powered off. The Stargate was modified to generate a periodic heartbeat data packet. When the MicaZ radios fail to see the heartbeat they go into a low-power watchdog routine.



Flight Test, September 2007









Acknowledgements: I would like to thank all those who helped me to build/debug this project, including Dr. Alan Cassell, Mike Krug and Jon Mihaly. Special thanks to the team at Garvey Space Corp. Thanks to Marina for her help at the September launch.