Need a project idea!

[Christopher]

So. I am looking to design a fairly advanced PCB in my spare time. Something to show off to potential employers!

I have no ideas, though, as this has to do something fairly interesting... I'm an engineer not a customer!

What I am good at and want to show off:

• 4 (maybe 6) layer PCB, no wiring
• Linux (ARM Processor w/ ram?)
• High-speed ADC/DAC (To measure what? Length of half a bee's dick?)
• Test & measurement
• Fairly low power to run off a battery possibly
• DC-DC Converter

Thanks for reading my very vague question!

[Beeper]

How about hacking a toaster to make perfect toast?

[alanb]

I would start by considering what industries your potential employers are in. I would then choose a project related to that.

[jancumps]

Put on some fpga. They are all so pin-count-high that fanning'm'out takes a few layers.

[NiHaoMike]

Maybe a simple software defined radio? One of those single chip tuners can cover a very wide range of frequencies. Then add an ADC, FPGA, and a few parts for a user interface.

[janengelbrecht]

ARM, ADC/DAC, test & messurement, DC-DC converter , 4-6 layer PCB...hmm what about an USB PC Oscilloscope ?
Input buffer/Attenuator => ADC => FPGA => ARM => USB => PC
And DAC ? Well make an AWG/DDS generator also.....

Just to keep it simple lets say sampling frequency 100MHz then you can have an analog bandwidth at 10MHz.

Of course you also need some PC Software then.....

[Richard Crowley]

A compact, portable device that would record 8- or 16-channel audio (24bit x 48K samples/sec) into standard FAT format files on an SDHC card.
Should be relatively easy with modern components, controllers (like Xmos, et.al.)  and code libraries. 
I would think it would fit into a form factor not much bigger than 4 or 8 AA cells (like a thick "smart-phone")
It would involve several diverse areas of hardware and software engineering and system integration:

• Analog engineering -- line-level audio -> multi-channel A/D converters
• Digital engineering -- interface to A/D, multiplexer (if external from A/D), interface to UI, interface to SDHC card, battery power management.
• User Interface engineering -- user control buttons, user display of function, timing, and audio levels, etc.
• Programming -- interleaving the sample values, forming standard Resource Interchange File Format (RIFF), i.e. WAV, or AIFF, disk format and folder/file management.

[ehughes]

Usually,  part  and software selection comes after project specification :-)

If your target audience is a potential employer,  think about this...   

1.) A simple, well executed task will always be more impressive than something complex executed poorly.

2.)  When I was new to engineering,  my first place hired an "old guy" (had to come out of retirement to pay for medical care, etc) who was well known in our area.   Great at analog, etc.   He simply knew everything.   Type of guy that was in charge of many projects big and small while working with a diverse set of individuals.     His advice to every engineer:

"Any asshole can make something complex.   It takes someone very smart to make it dirt simple."


Instead of showing off a complex PCB,  consider a simple project that executed a task very well.   Then show that you can produce a design documented with some of these elements:

 -Proper Notes with the PCB.   I.E.    Dimension drawings,  detailed layer stackups,  material selection notes, etc.
- CCA Assembly Documentation.  I.E.   A fully populated BOM,  Detailed assembly procedures,   Cable assembly diagrams
- Post Assembly / Burn Test Notes I.E.  How is the board tested, calibrated, etc after assembly?
- Detail your change management and life-cycle process
- Detail the desired operational specification
- Detail a set of software requirements such that elements can be traced backed to them.
- Detail a software test/regression plan


I would honestly rather see a simple design that had these elements.   It shows that you can define a project, work through details and produce a set of artifacts that can be worked on by other long after you are gone.

There are many people who can throw together a board with the specs you mentioned.   Very few can approach a project with a controlled, thought process.   These things show that you can *manage complexity*.  After having been on the side of making hiring decision,  I'll take a guy who can do the above on a board with one 555 timer on it than someone who chooses to embed something complex with no process control.


That all being said,  make me a controller than can run an autonomous robot that will collect leaves for me.   I just spend 1.5 hours on my front yard (I am surrounded my Oak and Maple trees!)

[Alex Eisenhut]

How about hacking a toaster to make perfect toast?

Would you like some toast?

[calexanian]

Something Internet Of Things!! (Snicker snicker) is that not all the rage these days? Get our piece of junk on the internet so we can cash in on something we don't need or even really understand.

In all seriousness though. Whatever you do have good documentation. Show you can easily not only lay out a board, but have it produced, and can be later understood by anybody. The actual use for the board is inconsequential as whatever you will be doing for them is custom anyways. Just give them a general idea of what your work looks like and how you like doing things.