The best bits to drill holes in your FR-4 board are tungsten carbide bits. They are, however, more expensive. You can also use HSS (high-speed steel) bits, but they will quickly dull. To drill larger holes start with a small bit and then progressively go larger.
Have a good working light shining at the board. This is crucial to get the holes well centered, especially for smaller bits.
It is time to solder through-hole and surface-mounted components.
I use Chip Quick solder paste for SMT components. When not in use, keep it in the freezer to extend its shelf life (it should not dry up). Chip Quick has a very low melting point and it’s perfect to use for SMT since you can apply tiny drops of it through a syringe it comes with. Apply small drops to pads and then position SMT components on them using tweezers.
Even tinned, the PCB traces will still corrode. Soldered joints will also oxidize since the solder is exposed to air.
After soldering, the board could be further protected by applying a coat of clear lacquer. The lacquer dries fast and can be soldered through afterward if needed.
I use this lacquer for final board protection:
http://www.testors.com/products/137179
This particular board did not need an enclosure. Instead, I mounted it on raisers. Holes for raisers were already drilled.
If you do have an enclosure, you may want to try to fit it: drill the holes and/or cut openings using a tool like a nibbler.
This is general guidance for the initial testing.
Thoroughly examine the board for physical defects: are all traces complete or there are breaks?
Are all pins soldered and is the quality of the solder joints acceptable? The soldering point may be missing (forgot to solder it), may have a bridge or a short, or it may be a poor joint (cold point). Just redo it.
Is any component missing? (Something blunt but may happen!)
You may need a magnifying glass or a loupe with a strong light.
Check again the orientation of components: LED, diodes, connectors, chips – where is pin 1 located. Check the orientation of tantalum and electrolytic caps. On SMT components, this can sometimes be confusing.
The quality of the tools that we use to write software for projects like this is very important when selecting which MCU we may want to use. Atmel has a very robust IDE based on Microsoft’s Visual Studio (which is arguably the best development environment out there). Coupled with a JTAG interface, writing and debugging firmware with the Atmel toolset becomes a pleasure.
Their ASF (“Atmel Software Framework”) is a good collection of ready-to-use projects and components. Investing time to learn that environment is the best way to become very efficient when using it.
There are a few things that I did not get to try this time around:
If a project is large enough, one could add a board ID that is readable by the software. That could be done by GPIO pins connected to, or “fused”, to certain voltage levels creating a short ID. This may be overkill for simple boards, but still, it is a known industry practice.
When printing from Eagle (a PCB CAD software), you can print only one layout per page. Usually, I need two since when I print to transparency, I like to overlap two identical images to get a better “black” for photo-etching.
Using this method I can compose several layouts into a single page to be printed. This also results in less waste.
I got hold of an old Sinclair ZX81 in apparently good and working condition!
ZX81 was the first personal computer I owned. I was 13 years old. At that time, I copied its 8K ROM, byte by byte, into a notebook, and hand-disassembled it (a consequence of which I still suffer from: I still remember some Z80 opcodes), but I never opened it. Finally, now I can do what I missed 🙂
This post shows how it generates TV images and how its Z80 CPU boots. I instrumented it and captured scope and logic analyzer images as it was powering on.
What a beauty!
This article is a follow-up to the “RS232 and Raspberry Pi” (link)
The same MAX board that I had also worked with Arduino at 5V although it was designed for 3.3V operation (capacitors onboard were sized for 3.3V operation – page 10 of MAX3243 datasheet). Still, it seemed to work fine up to the 28800 baud rate. Any faster and the software starts detecting data framing errors and the data becomes corrupted.
However, that was purely a design limitation of that particular RS232 board that I had.
