Well, we’ve all got to start somewhere, and this is the first board I ever had fabricated. There were a lot of “ah, hell, there’s no way it’ll work” moments that went into it’s creation. Firstly, I was trying to figure out how to use Fritzing to design a schematic, second, I was trying to figure out how to actually draw a schematic in the first place, third, I didn’t really understand the functionality of the circuit and fourth, I had no idea what the hell a “gerber” file was.
The actual design process involved me sitting at work on my lunch hour and fooling around with Fritzing, because it seemed the most approachable EDA tool for a rank amateur. I’d tried quite a few: KiCAD, the free version of Eagle, Upverter and hadn’t been able to make heads or tails of them. Fritzing was just basic enough to let me figure out how to lay down components and hook them up. The previous night, I hooked up a basic Larson Scanner with 8 LEDs, 8 resistors, a 74HC595 and an Arduino. Looked good. Then I wanted to try something a little more advanced, so I set about putting together the LEDs in the pattern of dice pips and feed them from the Arduino random number generator. The Arduino would “roll the dice” and output the correct pattern to the LEDs.
So, there I am, bored at lunch and plunking this circuit into Fritzing when I think I’ll fool around with the PCB layout tool. I mean, it has a tab and everything, right? I click on it thinking, “What the hell”, and start moving footprints around on the board and running traces between them. I knew enough that I realized I could run traces on the back and on the front (mainly because there was an option to do so.) It seemed logical to put the “stuff going out” on the top and “stuff going in” on the bottom. I didn’t know that things such as “vias” existed, so I was perplexed in how to run some of the traces, considering that the QA pin (the first LED) on the 595 is stuck opposite all the others, between Vcc and SER. So I ran +5V on the back of the board along with SCLK, RCLK and SER, to get them out of the way of my precious, precious LED traces.
I sat there looking at the layout and saw that no traces conflicted anywhere, and then for the first time noticed the “Fabricate” button at the bottom. It quoted some outrageous price, € 21.00 if I remember right, and I thought, “well, that would have been fun.” Thinking about it a little more though, I remembered all the times The Amp Hour had mentioned OSHPark, so I looked around to see if Fritzing would spit out Gerbers, and it did! (“file”, “export”, “for production”, “extended gerbers”). I zipped the files and uploaded them to OSHPark. To my amazement the system accepted it! $12 dollars later, I had three boards being sent out for fabrication.
I was more amazed than proud of myself – I had zero confidence that the thing would actually work when I soldered it up. I was sure I had gotten the LEDs backwards, or the footprint for the DIP Socket wrong, or incorrectly routed Clock / Data / Latch. One of those, or all of those, had to have gotten screwed up, because no one like me should be able to just get a working PCB manufactured!
Except it did work. First time! I’d like to say it’s chunky and inelegant, nothing but massive through hole components and a striking disrespect for space management, but for a first effort, I think it looks as good as any “blinky LED” Velleman kit I ever put together.
This is now my benchmark design. As I learn new methods and develop better skills, I will iterate the circuit to bring it up to whatever my current capabilities are.