Choosing what to build?

[eujean]

Hi there, I am an Electronics Engineering student, and have the following problem.

As part of a course at the university, in a group of 3 we have to design a system, which is completely modular, and designed from the ground up. There are a few limitations and the system can be described as a good system when it complies with the following:

• Does the system address a clear requirement / need?
• Is the system divisible into multiple functional blocks?
• Is the system divisible into multiple architecture blocks?
• Can the system be demonstrated in a basic undergrad electronics lab?
• Can the system's functionality be evaluated by measurable performance parameters

So this sounds simple enough right. But there are so many possible options. I was wondering if anyone on the forum has any ideas? If we do it, it will be completely open source/open hardware.

I also have a few examples:

• Access Control System
• Automated Animal Feeding System
• Unexploded Ordnance Detection Buggy
• Optical intercom
• etc...

Also the system has to be designed with an eye on production/manufacturing.

A guaranteed way for the project to be rejected according to the lecturer includes:

• Systems that work with liquids or pollutants
• Purely embedded / software designs
• Purely mechanical designs
• Designs that require a single IC, or a simple interconnect of a few ICs.
• Systems that cannot be disassembled into testable constituent subsystems.

Thanks in advance, and let the project ideas roll!

[tron9000]

As you already listed: "Can the system be demonstrated in a basic undergrad electronics lab?" - so your project must be producible within the capabilities of your labs means. So it must be compatible with the lab equipment you have access to.

I'm a Lab tech at a university and I ALWAYS get students asking for things we don't have or we cannot buy (due to financial, health and saftey, or ethical reasons) or assumed we had them in stock: like obsolete components.

So go to your lab, talk to the technical staff, ask them what they have lying around. Usually inspiration for projects can come from an older project lying around or some odd or unique parts in storage, begging to be used. Also ask them about the equipment they have, if your planning to analyse a 3GHz signal and they only have a spectrum analyser that goes up to 2.4GHz, they're not going to get one just for your project.

Budget: do you have a budget?

sourcing components? does your department have a preferred supplier? Here: we are not allowed to order anything from eBay or non-approved suppliers. Do you have access to a store of components?

I'm confused about the terms of rejection?

Designs that require a single IC, or a simple interconnect of a few ICs.

so it has to have more than one IC with as many interconnects as it can in order for it not to be rejected? cos its going to be pretty hard otherwise to make anything fancy without IC's?

[eujean]

I'm confused about the terms of rejection?

Designs that require a single IC, or a simple interconnect of a few ICs.

so it has to have more than one IC with as many interconnects as it can in order for it not to be rejected? cos its going to be pretty hard otherwise to make anything fancy without IC's?

What was meant was that you cannot simply propose a project consisting of an LED turning on and off by use of one simple 74-series logic chip or something simple like that. You can use IC's without limits, you just can't make it super super basic. If that makes sense.

As far as components sourcing goes, anything that is practically available is sufficient, however as mentioned, there has to be some thought about production/manufacturing, so as you said eBay simply won't do.

Will talk to the lab techs later, maybe they can light the lightbulb.

[george graves]

I feel like any help we would give you would be undermining both you as a student, and the spirit of what your professor intended.

It's time to put on your big boy pants.

 

[tautech]

In a uni here lecturers offered students industry projects (IP) and I had a small one done for me for as part of a project I was working on at the time.
Simply it was a PV powered charge pump, VCO controlled with values optimised for the PV's capabilities and the whole project part of a uP application. 6V @ 100mA was the PV max and 150V was the target for the charge pump, zener regulated.

Try that and I'll mark it against the finished project I received. 

Edit
Schematics, calculations, inductor design, BOM and PCB layout required for top marks. 

[charlespax]

Here are some design considerations.

- A single person can carry the entire project across campus. Nothing too heavy or big.
- The project can fit through a doorway without being tilted.
- The project can run without additional hardware or software design. Avoid making any kind of simulator.
- No dependencies. Each modular component can be tested independently. You can demonstrate progress and stay motivated.
- Keep it clean. Like your professor said, no fluids. Also avoid dirt, sand and other junk.
- Project can easily be setup and reset. No Rube Goldberg machines.
- Can easily make multiple copies. If you make something cool, each team member can have one!
- Small enough to store away. If it's small, you can bring it to job interviews or just keep it on a shelf to enjoy.
- The parts you make pop out. Avoid using individual components that overshadow your won work
- All aspects of the project are known before making hardware. Requires no original research or significant learning required.
- Keep it simple. No complex machinery.

Bad ideas:
- Working with something automotive. Cars are big and you can't carry them around. If you're working with the OBDII interface and you don't have a car, you'll have to deal with a simulator. Just too much extra work.
- Solar tracking system. Too big to carry. Individual components are larger than your own work.
- 3D printer that prints glass. Nobody's done this. Don't be the first. R&D required. Too complex of a machine.

Good idea:
- Automatic fish feeder. Specific need. Specific requirement. Simple mechanics (auger, motor). Simple electronics (microcontroller, RTC, motor output, timer configuration input), simple software (turn motor after some amount of time). Can be small. Cheap.

[Kilrah]

- 3D printer that prints glass. Nobody's done this. Don't be the first. R&D required. Too complex of a machine.

It's actually been done already

http://www.dezeen.com/2015/08/26/neri-oxman-3d-printing-transparent-glass-sculptural-structures-mediated-matter-mit-media-lab/

[A.Huggy]

Ideas:
Home security system
Automated functional test system for anything practical. (functional test system for computer mouse, or maybe a multimeter)

My favorite: Design a flashlight with several modes (search flashlight forums for design hints), and then an automated functional test system for the light.  This might violate the rule about no liquids or pollutants because the light will probably have a battery, but maybe that could be worked around by having the test system power it from it's power supply.  The test system would test light output, correct function of the switches and modes, proper function of low voltage shutoff.


I despise vague assignments like this, it seems like such a waste of time to make the students think of the problem that needs to be solved.  Instead of telling us to design a problem, and then design the solution, why not give us a problem so we can focus on designing a solution?

Every time I ask an instructor why an assignment is so vague rather than giving us some design requirements and having us build something that meets them, they cite creativity.  They want us to come up with projects no one has done before.  While there's something to be said for that, wouldn't it be more productive to give us clear design goals and have students compete to come up with the best solution? 

[MarkF]

I just finished building a function generator based around the Analog Devices AD9834 chip.  It includes a small OLED display and rotary encoder to control all functions.  The processor I used is a pic18f2550 MCU.  If you have PCBs made for the team, you will have a tool for testing other projects down the road.  The functions I implemented are Sine wave, Triangle wave, Square wave and Freq. Sweep (up and down).

The design process covered a lot of different disciplines.  You have an analog section, digital interfaces, MCU graphics library (write your own functions), GUI layout, software development, PCB layout and manufacture.

Or maybe you can come up with another piece of test gear that will provide utility for other projects...