Originally published by Nov 29, 2016
Today we will be discussing Arduino communication protocols. Devices need to communicate with each other to relay information about the environment, express changes in their states, or request auxiliary actions be performed. With any serious hobby electronics work, you’re bound to run into one or more of the main communication protocols in use, be it when working with different sensors, or with modules such as the ESP8266. In this tutorial, we hope to introduce the standard communication protocols that electronic devices use and explain each of them in detail using Arduino Uno.
Inter-device communication occurs over digital signals. Before we discuss communication protocols, we’ll first discuss how these signals are transmitted.
In a digital signal, data is transferred as a sequence of high to low and low to high switchings that occur very rapidly. These highs and lows in a digital signal represent 1s and 0s respectively that, when put together in sequence, carry information that can be interpreted by microcontrollers. Ever heard of bits and bytes? Individually, these 1s and 0s are bits, and when they are in groups of 8, they are called bytes!
A byte might look something like this: 10111001
As it turns out, this sequence of eight represents a number the same way a number like 597 represents five hundred ninety-seven. Each of the digits occupies a place-value, and the 1 or 0 in that place-value indicates how many times the place-value is counted.
In the example of 597, the 5 indicates that there are 5 hundreds, the 9 indicates that there are 9 tens, and the 7 indicates that there are 7 ones. Put together, this means 5 hundreds + 9 tens + 7 ones (500 + 90 + 7)…or five hundred ninety-seven. Since ones is 100, tens is 101, hundreds is 102, etc. this is called the base 10 system! In the base 10 system, each digit can have the value 0 through 9 (0 through 10-1).
Now we return to our number 10111001. We know this is the base 2 system because each digit has value 0 through 1. We can then say that each digit is a power of 2, which would mean that 10111001 is actually 1 * 27 + 0 * 26 + 1 * 25 + 1 * 24 + 1 * 23 + 0 * 22 + 0 * 21 + 1 * 20, or 185 in base 10.
As you can imagine, you can have number systems based on any number! Some common ones in math are base 2, base 8, base 10, and base 16. Mathematicians give these common number systems names for simplicity—base 2 is binary, base 8 is octal, base 10 is decimal, and base 16 is hexadecimal. Each number system follows the same principal: each digit represents a number of times a power of that base is counted, and the value of each digit can only be between 0 and base-1 inclusive.
Knowledge of different base number systems is useful because bytes and data are often represented in different ways. As you can imagine, it is easier to write out B9 (hexadecimal) than 10111001 (binary). In software, binary numbers are prefixed with 0b, octal numbers are prefixed with a 0, and hexadecimal numbers are prefixed with a 0x. Decimal numbers are not prefixed.
Knowledge of converting between bases is also useful because representing numbers in binary often makes for some cool math tricks as well. For the purposes of this tutorial, however, we’ll leave it at just that. Check out the appendix to this tutorial below for articles on these tricks!
The electrical engineering community decided to standardize electronics around three communication protocols to ensure device compatibility. Centering devices around a few protocols meant that designers would be able to interact with any device by knowing a few basic concepts about each communication protocols. These three protocols, UART, SPI, and I2C, differ in their implementation, but ultimately serve the same purpose: transferring data at high speeds to any compatible device.
The first communication protocol we’ll cover is Universal Asynchronous Receiver/Transmitter (UART). UART is a form of serial communication because data is transmitted as sequential bits (we’ll get to this in a bit). The wiring involved with setting up UART communication is very simple: one line for transmitting data (TX) and one line for receiving data (RX). As you may expect, the TX line is used to for the data to send device, and the RX line is used to receive data. Together the TX and RX lines of a device using serial communication form a serial port through which communication can occur.
The term UART actually refers to the onboard hardware that manages the packaging and translation of serial data. For a device to be able to communicate via the UART protocol, it must have this hardware! On the Arduino Uno, there is one serial port dedicated for communication with the computer the Arduino is connected to. That’s right! USB, which stands for Universal Serial Bus, is a serial port! On the Arduino Uno, this USB connection is broken out through onboard hardware into two digital pins, GPIO 0 and GPIO 1, which can be used in projects that involve serial communication with electronics other than the computer.
You can also use the SoftwareSerial Arduino library (SoftwareSerial.h) to use other GPIO pins as Serial RX and TX lines.
UART is called asynchronous because the communication does not depend on a synchronized clock signal between the two devices attempting to communicate with each other. Because the communication speed is not defined via this steady signal, the “sender” device cannot be certain that the “receiver” obtains the correct data. Therefore, the devices break data into fixed-size chunks to ensure that the data received is the same as the data that was sent.
The UART data packet looks something like this:
Devices that communicate via UART send packets of pre-defined size that contain additional information regarding the start and end of the message and confirmation of whether the message was received correctly. For example, to begin communication, the transmitting device pulls the transmit line low, indicating the start of a data packet. Long-term, however, this means that UART is slower compared to a synchronized form of communication because only a portion of the data transmitted is for the devices’ applications (the rest is for the communication itself!)
When implementing UART serial communication on most embedded platforms such as the Arduino, the user does not have to deal with communication at the bit level. Instead, the platform often provides higher level software libraries that are the only aspect of the communication process the user has to deal with. In the Arduino platform, users can use the Serial and SoftwareSerial libraries to implement UART communication for their projects.
Here’s a brief C++ reference for Arduino Serial and SoftwareSerial initialization and use.
|Serial and SoftwareSerial Method||Purpose||Code||Explanation|
|Constructor (SoftwareSerial only)||Define the GPIO pins that will serve as the UART RX and TX lines||SoftwareSerial comms (2 , 3);||Defines a serial connection with RX line on GPIO 2 and TX line on GPIO 3.|
|begin||Define the baud rate (transmission speed) for the serial connection in range 4800 to 115200||comms.begin(9600);||Communication on the “comms” serial port will occur at 9600 baud.|
|Write byte data converted into human-readable characters over serial connection||comms.println(“Hello World”);||Writes bytes equivalent to Hello World (human-readable characters)|
|write||Write raw byte data over the serial connection||comms.write(45);||Writes byte with value 45.|
|available||Evaluates to true when data is available over the serial connection||if (comms.available())||Enter if statement if there is data available to read over the serial connection|
|read||Read data from the serial connection||comms.read();||Reads from serial connection|
One important aspect to note about UART communication is that it is designed for communication between only two devices at a time. Because the protocol only sends bits indicating the start of a message, the message content, and the end of a message, there is no method of differentiating multiple transmitting and receiving devices on the same line. If more than one device attempts to transmit data on the same line, bus contention occurs, and the receiving devices will most likely receive garbage unusable data!
Furthermore, UART is half-duplex, which means that even though communication can occur bidirectionally, both devices cannot transmit data to each other at the same time. In a project where two Arduinos are communicating with each other via a serial connection, for example, this would mean that at a given time, only one of the Arduinos can be “talking” to the other. For most applications, however, this fact is relatively unimportant and not disadvantageous in any way.
The next communication protocol we’ll cover is Serial Peripheral Interface (SPI). SPI is different from UART in several key ways:
The hardware connection diagram for SPI is slightly more complicated, and looks something like this:
The first feature about SPI to note is that it follows a master-slave model. This means that the communication relies on defining one device as a master, and other devices as slaves. This creates a hierarchy between the devices signifying which device is effectively “in control” of the others. We’ll discuss this master-slave model shortly when we describe what a sample communication between a master and slave device might look like.
We noted earlier that multiple slaves can be connected to a single master device. The hardware diagram for such a system might look something like this:
SPI does not require a separate transmit and receive line for each slave that is connected to the master. One common receive line (MISO) and one common transmit line (MOSI), along with one common clock (SCK) line are connected between all the slaves and the master. The master device decides which slave it is communicating with through a separate SS line for each slave. This means that each additional slave that the master communicates with requires another GPIO pin on the master’s side.
SPI is synchronous, which means that communication between master and slave is closely tied to a clock signal (a square wave of fixed-frequency) defined by the master device. Here we see one of the direct effects of the master-slave model. The master device drives the communication by specifying the communication speed through the clock signal, and the slaves respond by communicating at that speed. This definition of speed works for any communication speed the master decides on (within the maximum speed tolerable by the slave devices).
In SPI, two specific features of the clock signal determine the start and end of data transmission: the clock polarity (CPOL) and clock phase (CPHA). CPOL refers to the idle state (either low or high) of the clock signal. To conserve power, devices will put the clock line at the idle state when not communicating with any slaves, and the two options available for this idle state are either low or high. CPHA refers to the edge of the clock signal upon which data is captured. A square wave has two edges (the rising edge and the falling edge), and depending on the CPHA setting, data is captured either on the rising edge or the falling edge.
There are four distinct combinations of CPOL and CPHA, that can be summarized in a table.
|CPHA = 0
(“first edge” of clock signal)
|CPHA = 1
(“second edge” of clock signal)
|CPOL = 0
|CPOL = 1
Wikipedia does a great job graphically representing the CPOL and CPHA settings and their relationship to the data being transmitted!
Now we’ll focus on SPI implementation on the Arduino using the Arduino as the master device (SPI.h). The SPI digital pin connections for SCK, MOSI, and MISO are predefined on Arduino boards. For the Arduino Uno, the connections are as follows:
Any digital pin can be also used for the SS pin. To select the device, this digital pin must be driven low.
Here’s a brief C++ reference for Arduino SPI initialization and use.
|Constructor||Define the clock speed, data bit order (most significant bit first or least significant bit first), and SPI mode||SPI.beginTransaction
(SPISettings(14000000, MSBFIRST, SPI_MODE0));
|Defines a SPI connection at 14MHz in SPI Mode 0 and data transmission occurring with the most significant bit first|
|digitalWrite||Select the slave device attached to this GPIO pin||digitalWrite(10, LOW);||Drive the GPIO pin low to select the slave device. Drive the pin high afterwards to de-select the slave.|
|transfer||Transfer the byte to the selected slave||SPI.transfer(0x00);||Send a value of 0 as a byte|
|endTransaction||Ends the SPI transaction (should be called after a digitalWrite(high) is called on the SS line)||SPI.endTransaction();||Ends the SPI transaction|
SPI is full-duplex, which means that communication always occurs bidirectionally, even if the application only requires data transfer occur in one direction. SPI full-duplex data transfer is usually implemented with a shift register. (Check out the appendix for a tutorial on shift registers!) This means that as each bit is read, the bits before it are shifted over by one place. The foremost bit is subsequently popped off and sent back through the SPI connection to the other device!
Inter-integrated circuit (I2C), pronounced either “i-squared-c” or “i-two-c,” is the final communication protocol we’ll cover in this tutorial. Though its implementation is the most complicated of the three protocols, I2C addresses several drawbacks in the other communication protocols, giving it an advantage over the others in some applications. These include:
At the hardware level, I2C is a two-wire interface—the only two wires required for an I2C connection are a data line (called SDA) and a clock line (called SCL). The data and clock lines are pulled high in their idle states, and when data needs to be sent over the connection, the lines are pulled low through some MOSFET circuits. For the purposes of this tutorial, we will not discuss the MOSFET operation inside I2C circuitry. The important fact to take away is that the system is open-drain, which means that the lines can only be driven low by the devices. Thus when using I2C in projects, it is crucial to include pull-up resistors (usually something like 4.7kΩ) to ensure that the line is actually pulled high in the idle state.
I2C is unique because it solves the issue of interfacing with multiple slave devices through addressing. Just like in SPI communication, I2C makes use of a master-slave model to establish the “hierarchy” of communication. However, instead of selecting slaves through separate digital lines, masters select slaves through their unique byte addresses. These addresses might look something like this: 0x1B. This means that connecting more slaves to a master will not require additional digital lines. As long as each slave has a unique address, the application will still be able to differentiate the slaves on these addresses alone. You can think of these addresses like names. To call a slave’s functionality, the master merely calls its name, and only that slave responds.
The address and data portions of the I2C communication line might look something like this.
Notice the ACK and NACK bits on the communication line. These bits denote whether or not the addressed slave responds to the communication—a way of periodically checking whether or not communication is occurring as expected. These bits, of course, are unrelated to the address or data bits being sent, but in the grand scheme of the communication, they add very little extra time to the communication compared to the many start and end bits and pauses that occur in protocols like UART.
I2C gives slaves freedom to decide what communication requests might look like. Different slaves require different bytes to be written over in different orders over the SDA line to write and request information from them. In some accelerometer modules, for example, bytes are written to denote which hardware register the master wants to read before any read requests are made. For these specifications, the user must refer to the slaves’ datasheets for device addresses, register addresses, and device settings.
On the Arduino, I2C implementation occurs through the Wire library (Wire.h). The Arduino can be configured as either an I2C master or slave device. On the Arduino Uno, the connections are as follows:
Here’s a brief C++ reference for Arduino I2C initialization and use.
|I2C (Wire) Method||Purpose||Code||Explanation|
|begin||Initiate the library and join the I2C bus as either a master or slave||Wire.begin();||Joins the I2C bus as a master. If an address is specified as a parameter to the method, the Arduino joins the bus as a slave with that address.|
|beginTransmission||For Arduino configured as a I2C master: initiate transmission with the slave that has the given address||Wire.beginTransmission(0x68);||Begins transmission with the slave that has the hexadecimal address 0x68|
|write||Write the byte data over the I2C bus||Wire.write(0x6B);||Writes byte data of value 0x6B over the I2C bus|
|requestFrom||Request a specified number of bytes from the slave with the given address; Has the option of releasing the I2C line or holding onto it for further communication
The requested bytes are put into a buffer to be subsequently read through Wire.read() calls.
|Wire.requestFrom(0x68, 6, true);||Requests 6 bytes from the slave that has the address 0x68. Releases the I2C line after the request is complete.
If the last parameter is false, the Arduino will hold onto the I2C line for further communication, not allowing other devices to communicate over it.
|read||Read bytes put into the buffer after transmissions from the slave device.
This method is called after a call to Wire.requestFrom
|Wire.read();||Reads one byte from the buffer (after a call to requestFrom) has been made. To read two bytes from the buffer, this method must be called twice:
Wire.requestFrom(0x68, 2, true);
|endTransmission||Ends the current transmission; Has the option of releasing the I2C line or holding onto it for further communication||Wire.endTransmission(true);||End transmission and releases the I2C line.|
Multiple masters can be connected over an I2C bus just by connecting their SDA and SCL lines to the bus’s lines. However, only one master can communicate with slaves at a time, because having multiple devices attempting to communicate with each other would lead to bus contention. Similarly, communication cannot occur bidirectionally from master to slave and from slave to master at the same time because that would also lead to bus contention. This makes I2C half-duplex, just like UART!
Finally, multiple master devices cannot communicate with each other over that same I2C bus. In applications where multiple masters are connected to slaves, masters might communicate with each other over a separate bus or via a separate communication protocol.
If you made it through this tutorial, you should have all the tools you need to work with the UART, SPI, and I2C communication protocols! Check out some of our Arduino projects for further examples on using these protocols!
Interested in hearing more about how to make the most of your Arduino projects? Explore some of our comprehensive Aduino guides:
Bitwise Operations in C: http://www.cprogramming.com/tutorial/bitwise_operators.html
Shift Registers: https://learn.sparkfun.com/tutorials/shift-registers