Finally I have a fully built unit of production quality. It has the revised board and a neato sticker on top which I designed myself using The Gimp. As you can see, my graphic design skills are not the best but at least it looks like a real product now and not something hacked together by a hobbyist.

Next stage is to do some real thorough testing on this unit before offering it for sale.

I’m not my best first thing in the morning. Some evil person has been taking pictures…

It is amazing me how much more effort it is to produce a product for sale than to just build something for fun. So far I’ve done all these things, none of which I would bother with for a hobby project:

1. Draw up schematic in Protel
2. Design PCB to fit the box
3. Find a suitable box for it
4. Find a good PCB manufacturer
5. Find best suppliers for the components
6. Get a prototype PCB made
7. Order parts
8. Build and test prototype board
9. Design and build chip programmer
10. Write boot loader and firmware update code
11. Test bootloader and firmware update app
12. Work out cost of parts and pricing of final product

But there’s still more to go:

1. Design drilling templates for the box
2. Write manufacturing procedure
3. Find some solution for easily making DB9-shaped holes
4. Design front panel artwork
5. Get front panel label manufactured
6. Register a domain name
7. Find a web host and set up a website for the product
8. Design web pages
9. Work out a good way to package the product
10. Write test procedure and/or automatic test software
11. Set up an online ordering and payment system
12. Write advertising copy and take photos of finished product
13. Buy some Google ad-words to promote the product
14. Set up customer support database and support forums

And probably more things I haven’t thought of yet!

I finally got my lazy arse into gear and assembled the PCB. I got the parts from DigiKey – it still amazes me that it is easier and cheaper to get electronic parts that were manufactured in China or somewhere shipped halfway around the world from the USA than it is to buy them here in Australia!

In any case, I’ve now assembled the board. I did it by hand-soldering this time instead of solder reflow. I think reflow is easier. I’ll do that next time.

Since I had so many spare boards, I hacked one up to be a programming adapter. This is a new feature to make it easy to set up the firmware when I eventually have to make dozens of these things. When the SAM7 chips come from the factory, they need certain pins pulled high to get the USB booloader started. I have put a header on the board to bring out all these pins to a ribbon cable, the USB lines are also brought out. All I need to do now is flip a switch to start the bootloader, then flip it back and download the firmware via USB.

Having all the USB-related parts on a separate board also means I don’t need to install those parts on the production boards which saves a little time and money since they are not needed for normal operation.

I was very pleased to see that apart from one little solder bridge (my bad, easy fixed), the board fired up perfectly and worked first time! I need to do some more thorough testing but it is looking very good so far.