Skip to main content
Engineering LibreTexts

15.2: Input/Output

  • Page ID
    35871
  • Consider a typical embedded application such as a programmable or “intelligent” thermostat. Unlike a normal electro-mechanical thermostat, these devices allow the home owner to automatically change temperature at different times of the day in order to save energy. After all, why have the heat or air conditioner running when no one’s home? Certainly, these devices do not come with a monitor or keyboard. In their place may be a small LCD display with a few fixed messages and a two digit numeric display for the temperature. For input there may be as few as two or three buttons for programming (set item plus up and down). By comparison, a microwave oven will probably have a complete numeric keypad with a series of special function buttons along with multiple seven-segment displays, or possibly several alpha-numeric displays for short messages. In any case, these devices are far different from the standard desktop computer. Consequently, a programmer’s approach to input and output processing will be much different in the embedded case.

    To start with, it is unlikely that there will be printf() and scanf() style functions. They are largely worthless in this world. What use would printf() be if all you have for output hardware is a bunch of LEDs? For input, you often need to read the state of switches, pushbuttons, and possibly some form of level control such as a rotary knob. For output, you often need to simply light an LED or set a value on a seven-segment display. For “fixed” style messages, these also need only a single signal to turn them on or off, such as an LED. In more advanced applications, a multi-line alphanumeric display may be available so setting individual letters is a possibility. In almost all cases these chores are handled by setting or clearing bits on specific output or inputs ports on the microcontroller. Some ports may be set up as a byte or word. Further, some ports may be bi-directional, meaning that they can behave as either input or output depending on some other register setting. Ports are little more than pins on the microcontroller that are hooked up to external circuitry. Thus, if a port is connected to an LED circuit, setting the output of the port HIGH could light the LED while setting the port LOW could turn off the LED. The obvious question then is “How do you read from or write to a port?” In many cases ports will be memory mapped. That is, a specific address in the memory map is allocated to a given port. You read and write from/to it just like any other variable. Further, development systems sometimes disguise these addresses as pre-defined global variables. They might also include a library of specific routines to utilize the ports. Thus setting a certain port (let’s call it the “A” port) to a high might be as simple as PORT_A = 1; or set_portA(1);. Reading from a port might be something like a = PORT_A; or a = get_portA();. Consequently, embedded code is often all about reading and writing to/from ports and then branching to the requested chores.

    There are some tricks to this. For example, how do you know if a key has been pressed? Calling get_portA() tells you the state of the switch connected to port A at the instant you call it. There is no “history” here if this is a simple momentary pushbutton instead of a toggle switch. In a case like this you might “poll” the port waiting for something to happen. This involves using a simple loop and repeatedly reading the port state in the loop. The code breaks out when the state changes:

    while( get_portA() );
    

    This will keep looping until get_portA() returns 0. Of course if you need to monitor several ports as is typical, you’ll need to read a value from each for the test. This form of monitoring while waiting for something to happen is called an event loop. It may not be evident, but your house and car are probably filled with devices running event loops, just waiting for you to do something! These loops execute fairly fast so a time lag between your push and the resulting action is not noticed. On the output end, a port normally stays at the value you set it, so there is no need for a loop to “keep it set”.

    For more complicated displays such as a seven segment or alpha-numeric device, you may need to create a table of values indicating bit patterns for each numeral or letter to be displayed. These patterns, or words, would then be sent out various ports that are in turn connected to the displays.

    For variable input devices such as a volume control, the external device (knob, slider, sensor, etc.) will be connected to an analog to digital converter or ADC. This might be a separate circuit or the controller may have these built-in. The converter turns the analog voltage into a numeric value that you can read from a port. For example, a volume control may be just a potentiometer connected to a fixed voltage. As the knob is moved, the voltage at the wiper arm will change. The A/D converter may encode this into a single byte. A byte ranges from 0 to 255 in value. Thus, if the volume is at maximum, the port will read 255. If the volume is at halfway the port will read about 127, and finally 0 when the volume is off.

    • Was this article helpful?