Approach

The approach taken here is to apply the concepts of unit-testing and refactoring to the design of control systems. For electronics, to unit test, one way would be to start with code, then replace it with a piece of hardware that does the same thing, for example, a filter. Refactoring might be recreating a completed project on a breadboard with a duplicate set of parts (to test) and then modifying (to refactor). Good examples are changing the output of a filter, voltage regulator or amplifier, or switching MCU.

Any control system needs a number of components.

flowchart

Always start by building the simplest possible device that works. For electronics, this is usually a blink program. For a control system, it would be processing a fake signal. While this seems trivial, it validates the whole toolchain: the power supply, the compiler, the programmer, the hardware, the wiring etc. Once that works, make one change at a time and test. If it fails, you know exactly where the problem is! In this way, you can build towards your goal.

Tools of production

Start with a small project and a budget. Buy the best set of equipment it affords for that project. Build your set over time.

These are the tools I have to date:

  • For PIC MCUs:

    • XC8 for compiling
    • a PicStart+ clone for programming pic16 MCUs (ZIF socket and ISCP) and a PicKit2 clone for programming pic16 and pic18 MCUs (ISCP only)
    • pk2cmd for uploading

  • For AVR MCUs:

    • avr-gcc and avr-objcopy for compiling
    • a USBasp clone for programming
    • AVRdude for uploading

  • GNU Make for managing compilation and writing of hex files.

  • For circuit design:

  • For circuit building:

    • Power supply
    • Fluke 179 DMM
    • Hakko FX888D soldering iron
    • Panavise 396 with Quadhands workbench mount
    • SS-02 solder sucker with silicone tip
    • Sidecutters
    • Tweezers
    • A 2mm drill bit
    • A set of electronics screwdrivers
    • Wire strippers
    • Components: solder, a resistor pack, capacitors, wires, stripboard, breadboards and dev boards

How to choose the next project

There are many projects and limited time. Use the following guide (in order of priority):

  1. Prefer projects that build on previous projects
  2. Prefer new projects that enable multiple future projects
  3. Prefer projects that assist your career
  4. Prefer projects of commercial application

Beyond this, being interested is the best motivation.

Next articles:

  1. LED dimming and toggle w/ AT90S2313 (PWM, interrupts, rotary encoder)
  2. LCD display "Hello world" w/ AT90S2313 (LCD)
  3. LCD control over serial link w/ 2x AT90S2313 (LCD, interrupts, UART)
  4. Circular buffer on AT90S2313
  5. Radio serial link

Some ideas:

  • a process for circuit board creation
  • bayesian classifier + IP blocking for website attacks
  • control the position of one servo motor with another
  • WebGL fabric modelling
  • Yet Another PIC Programmer
  • ngspice tutorial
  • temperature controlled oil heater with remote sensors