Our high school has tasked my and a friend of mine with with building a web connected controller for a hydroponics system. The school has budgeted 5k for us to spend as we like, and we are intent on making our own single board computer (we know this is way overkill) based on the TI AM3874. We both have experience making custom pcbs with AVR based controllers and the like but never with somthing so complex. Additionally we have to have this project completed by late december.
- Does this sit within the realm of possibility?
- What cad package would be the best? all of TI's reference designs are in OrCAD (we can get the high speed version for ~$500, we don't think Eagle or Kicad will cut it)
- What are the pitfalls we should watch out for?
As others have commented, your aim of making this kind of SBC is probably far too high for your experience, equipment, schedule, and maybe even budget.
But you are young, you think you can surely do it, all those old folks don't have the spirit, etc.
I suggest you have it both ways: get a RaspberryPi (or something similar, but a PI is as good a choice as the alternatives), and use it to develop a (fully working) proof of concept for the total system. If your SBC project is feasible, you should be able to do the Pi version in ~ 10% of the time and budget. And it will give the other teams in the project a chance to experience a version of the full system early on. Never underestimate the value of an early prototype for user feedback!
If the PI prototype is successful and you have sufficient time, budget and motivation left you can continue with your SBC. By that time we will surely see you back here on SE with more detailed questions. If the SBC development fails (of course it won't, but just in the off chance), the hydroponic system is not without a controller, and you are not without a good result to get your marks for.