I decided a few weeks ago that I would make my own circuit boards (PCB) for a while. At least for the first Beta.
I succeeded, but the time and money invested make it a very “not profitable” adventure. I had to deal with nasty chemicals, capricious sensitized copper coated boards, unfriendly software and a long process that requires a lot of precision.
This guy has a detailed page about how to go about it. In theory, the whole process is pretty simple:
- Use a CAD software (I used Eagle)
- Print on laser printer onto translucent material
- Expose board to printed material (with a lot of light)
- Develop the photo (reminded me of being a teenager learning darkroom techniques)
- Etch (remove copper with nasty chemicals)
- Drill holes to insert parts
Yup. That’s exactly what I did. Over, and over, and over again.
Eagle, from cadsoft.de, is extremely non intuitive. Damn! It took me days to produce anything! Maybe it’s me. It must be me. Damn!
In the end, I must admit, the software does an excellent job. It must be me…
My first few tries at printing on the laser printer didn’t go too well. The blacks weren’t black enough. They have to be opaque. I found a bunch of old inkjet transparencies in a box. I ended up printing on my Canon inkjet with very good results. I only managed 300 dpi resolution, but it was enough for my simple circuits. Mike recommends 600 dpi. It would be necessary if I was trying to get the wire traces to be 8 mil wide (0.2 mm). But I decided to use 20 mil traces (0.5 mm). I did manage to produce an 8 mil trace board, but it was not very reliable, and the traces were way too fragile.
The first board I printed was 2″ x 4″.
Yes, the traces a 8 mil. The board is slightly overdeveloped. So some of the traces are actually less that 8 mil. You can see a broken one near the middle, just left of center. That little board was cheap. $2. It required a “full noon sun on a clear day” (as the included instructions specified) exposure of ten minutes. It also required the use of steel wool to remove the photo sensitive layer before I could solder. Hence the broken trace.
I then printed 5 boards to a 6X6 pre-sensitized copper board. That board had a defect. There were bubbles in the sensitive layer. It never looked good. Then I tested a pattern on a new board, at different exposure times and different development times. I found a combination that made sense.
Success came when I started to use MG Chemical boards and products. Exposure is 9 minutes using a 15 watts standard fluorescent tube. Development time is about 4 minutes. Etching took 20 minutes.
Then, of course, I had to drill some holes. A number 68 drill bit is necessary. 0.031″, or just about 1/32 of an inch. 0.8 mm. Pretty standard hole size. The drill bit IS fragile. Yup.
You’d think that only one size hole would do, right? I had to run to the hardware store to buy a 3/64 drill bit. Number 56, at 0.465″, or 1.2 mm. I bought a box. They are fragile too!
It took me half a day just to drill and install components.
But now that it’s done, I think I could get another pair of boards within half a day, all included. Actually, I’m planning to produce 4 pairs. But it’s a secret. The Beta is coming…