• Figure out where the rocket is
• Control the chute ejection systems
• Control staging and ignite secondary engines
• Record flight data
• Provide system status information
• Upload/download to a pc
• Control on-board devices (e.g. cameras)
• Do real-time math (e.g. Kalman filtering.)
• Be easily swapable amongst a variety of rockets.

We may also have various self-contained items: a small transmitter to aid in recovery, a corner reflector for ground based telemetry.

We consider a number of solutions...

• A PC/104 system.
This is quite attractive: the rocket would be able to run Linux. Unfortunately, the form-factor of the cards is just too big: these will not fit into anything less than a 4" diameter rocket.
• An Intel 80x96-based system.
This is a nice chip, but it's not cheap and I can't find a free C-compiler for it.
• Multiple Basic stamp microcontrollers.
These are small enough to fit in a 2" tube, and cheap, and run a high-level language (*cough* Basic.)
• Multiple SX-28 microcontrollers.
These are small, very fast, and cheap. These talk a custom machine language.

The SX-28 multi-processor set up looks like the way to go. Darren goes to work designing "rocket-bus." This will be a modular system of microcontroller cards talking to each other via a token-ring network. I like the modular idea, but dislike the token-ring aspect: lose one card, and we lose the entire system. Even with watchdog timers and in-flight reboot, I'm skeptical. I also want electronics modules in different stages of the rocket: a token-ring approach will need to deal with in-flight network partitioning. I figure Darren will come up with a good solution here.

Regardless of the final bus configuration, we need a number of system modules. These can be linked together to produce a custom package for a specific mission:

• A power/clock module. Must provide power and timing to all other cards. Needs master/slave modes for multi-stage rockets.
• A data recorder. The simplest approach is just to record everything that comes over the bus. 64M would be more than enough. It might also be nice to have a flash-ram-in-a-can recorder. This would be useful for post-mortem analysis.
• Analog input interfaces.
  • Accelerometer. Seems useful, especially for apogee detection. DMR believes double-integration for altitude will be totally unreliable.
    • Analog Devices single-axis accelerometer.
    • Summit Instruments' tri-axial accelerometer.
  • Inclinometer. Nice to have, hard to achieve.
  • Barometeric pressure sensor. Well, it worked in Darren's rocket for chute deployment, but I don't want to rely on this for altitude estimation.
  • Stain gauges. I want to know where the weak-points in the rocket are.
  • Thermometers. Not important for small engines, but I'm going to need these for the bigger ones.
  • Gyroscopes. May need hi-altitude active controls. Gyro based input may be good here.
• Digital I/O devices
• External RS232 interface. Needed for final systems check, pre-flight parameter downloading, and post-flight data extraction.
• External Status, Arming Interface. At the least, I want to have a happy/unhappy LED on all rocket bodies.
• GPS Receiver. An all-round good idea. Hard to find one that fits in less than a three inch tube though.
• Radio link to ground system. Would be nice to know what the rocket is doing when it's out of sight.
• High power output. Must provide high power to ignite motors and ignition charges.
• Controller boards. Listen for input from other devices, and kick out commands/status info when specific conditions are met. This card will control the various output devices.

Notes