Now that you have done several of these how would you judge them and what would you do differently. Have you standardised on a basic design in order to minimise risk in both development and operation? Or do you always have to try for something new because new modules are constantly being produced.
Projects are usually too different to standardise much ( except some code, sometimes). Similarly you always find things that could be done better next time, but again you rarely get the chance as next time is so different.
There are some common factors, like :
Using RS485 for data (because it's simple, easy and cheap)
Making sure everything is protected from any conceivable abuse (mostly shorts between data and power)
Avoid putting hardware in inaccessible places
Make spares of everything
Don't buy a PSU from a cheaper brand than mean Well, and don't run at more than 75% of rating.
Allow time for burn-in tests (which invariably disappears because something screws up, causing delays)
Make things modular, so if one works, there is negligible chance of problems when there are more of them.
How much does the available technology influence the designs (the creatives concept) or do they come to you saying we want this make it happen?
It works both ways - in several cases what happens is I find an interesting part, knock up a demo unit & show it around to plant the seeds of some ideas, and projects then come out of it in various ways.
one client has a wall in their studio covered in various of my demo pieces, which they then show to their clients to see if anything takes their fancy.
For example, this project happenned a couple of years after me finding some white LED matrix displays & playing around with greyscale drivers :
http://www.jasonbruges.com/projects/uk-projects/mirror-mirror (actually the panels were destined for another project that fell through, & they got recycled as an artwork for the V&A!
And this came pretty much directly from "PSP displays - how cheaply can I make something to drive them"
http://www.cinimodstudio.com/project/flutter/And this from"I need something to make up the free-delivery value at Mouser, ooh look - 0402 white LEDs, it would be fun to make a tiny 32x32 matrix"
http://cinimodstudio.com/project/baby-angel/Obviously it's not entirely that straightforward but none of the above would have happened (at least not in the form they ended up) had I not randomly decided to play with something & make up a demo unit.
What is the brief that you are asked to fulfil?
Varies a lot - most of my customers involve me at an early stage & I often have to steer them away from impractical/ridiculous ideas. I have occasionally rescued projects where people had just got hopelessly out of their depth trying to scale up something they got running on an Arduino with no understanding of Ohm's Law.
Wish I could do something like this.