Arduino vs Raspberry: Which Board Wins for I2C IoT Sensor Projects?

When building an I2C sensor node with network telemetry, the choice between an Arduino microcontroller and a Raspberry Pi single-board computer dictates your entire software architecture. For raw hardware control, sub-millisecond timing, and low-power 24/7 relay switching, the Arduino Uno R4 WiFi is the definitive pick. For local database storage, complex API integrations, or running a local web dashboard alongside your sensors, the Raspberry Pi Zero 2 W wins.

This guide cuts through the generic comparisons and provides a concrete decision framework, a complete I2C climate controller build targeting the Arduino Uno R4 WiFi, and exact debugging steps for the most common network failures.

The Core Divide: Microcontroller vs. Microprocessor

The fundamental difference lies in the execution environment. The Arduino Uno R4 WiFi utilizes a Renesas RA4M1 ARM Cortex-M4 microcontroller running at 48 MHz. It executes bare-metal C++ (or an RTOS), meaning your I2C read commands and GPIO relay toggles happen with deterministic, microsecond-level precision. There is no operating system to interrupt your code for background tasks.

The Raspberry Pi Zero 2 W uses a Broadcom BCM2710A1 quad-core Cortex-A53 running at 1GHz, booting a full Linux OS. While vastly more powerful for floating-point math and multitasking, Linux introduces jitter. A Python script reading an I2C BME280 sensor might be delayed by milliseconds if the OS decides to handle a network interrupt or write to the SD card.

Decision Tree: Arduino Uno R4 WiFi or Raspberry Pi Zero 2 W?

Use this matrix to terminate your board selection process. Do not default to the Pi just because you are more comfortable with Python; the hardware overhead is rarely worth it for simple telemetry.

Project Build: I2C Climate Controller with MQTT Telemetry

This build targets the Arduino Uno R4 WiFi. It reads temperature and humidity from an I2C BME280 sensor and controls two 5V relays (heater and exhaust fan), publishing the telemetry to a local MQTT broker.

Parts List

• Microcontroller: Arduino Uno R4 WiFi (ABX00087) - ~$27.50
• Sensor: Adafruit BME280 I2C/SPI Breakout (2652) - ~$14.95
• Actuators: 4-Channel 5V Relay Module with Optocoupler Isolation - ~$7.99
• Wiring: 22 AWG solid core hookup wire, 4.7kΩ pull-up resistors (if BME280 board lacks them)

Pin Mapping Table

Complete Compilable Code

This code requires the ArduinoMqttClient, WiFi, and Adafruit_BME280_Library libraries installed via the Arduino IDE Library Manager.

#include <WiFi.h>
#include <ArduinoMqttClient.h>
#include <Wire.h>
#include <Adafruit_BME280.h>

// --- Pin Definitions ---
#define RELAY_HEATER_PIN 8
#define RELAY_FAN_PIN    9
#define STATUS_LED_PIN   LED_BUILTIN

// --- Thresholds ---
#define TEMP_HIGH_C 26.0
#define TEMP_LOW_C  20.0

// --- Network & MQTT Config ---
const char ssid[] = "YourNetworkSSID";
const char pass[] = "YourNetworkPassword";
const char broker[] = "192.168.1.100";
const int port = 1883;
const char topic[] = "greenhouse/climate";

WiFiClient wifiClient;
MqttClient mqttClient(wifiClient);
Adafruit_BME280 bme;

unsigned long lastPublish = 0;
const long publishInterval = 10000; // 10 seconds

void setup() {
  Serial.begin(115200);
  pinMode(RELAY_HEATER_PIN, OUTPUT);
  pinMode(RELAY_FAN_PIN, OUTPUT);
  pinMode(STATUS_LED_PIN, OUTPUT);
  
  // Relays are Active LOW; set HIGH to turn off initially
  digitalWrite(RELAY_HEATER_PIN, HIGH);
  digitalWrite(RELAY_FAN_PIN, HIGH);

  // Initialize I2C Sensor with error handling
  if (!bme.begin(0x77)) { // Adafruit default is often 0x77
    Serial.println("FATAL: Could not find a valid BME280 sensor, check wiring!");
    while (1) {
      digitalWrite(STATUS_LED_PIN, HIGH); delay(100);
      digitalWrite(STATUS_LED_PIN, LOW); delay(100);
    }
  }

  // Connect to WiFi
  Serial.print("Connecting to WiFi...");
  WiFi.begin(ssid, pass);
  while (WiFi.status() != WL_CONNECTED) {
    delay(500);
    Serial.print(".");
  }
  Serial.println(" connected.");
}

void loop() {
  // Maintain MQTT connection
  if (!mqttClient.connected()) {
    Serial.print("Connecting to MQTT broker...");
    if (!mqttClient.connect(broker, port)) {
      Serial.print("MQTT connection failed, rc=");
      Serial.println(mqttClient.connectError());
      delay(5000); // Wait 5s before retry
      return;
    }
    Serial.println(" connected.");
  }
  mqttClient.poll();

  // Read sensors and control relays
  float tempC = bme.readTemperature();
  float humidity = bme.readHumidity();

  if (tempC < TEMP_LOW_C) {
    digitalWrite(RELAY_HEATER_PIN, LOW); // ON
    digitalWrite(RELAY_FAN_PIN, HIGH);   // OFF
  } else if (tempC > TEMP_HIGH_C) {
    digitalWrite(RELAY_HEATER_PIN, HIGH); // OFF
    digitalWrite(RELAY_FAN_PIN, LOW);     // ON
  } else {
    digitalWrite(RELAY_HEATER_PIN, HIGH); // OFF
    digitalWrite(RELAY_FAN_PIN, HIGH);    // OFF
  }

  // Publish Telemetry
  if (millis() - lastPublish > publishInterval) {
    String payload = "{\"temp\":" + String(tempC, 2) + ",\"hum\":" + String(humidity, 1) + "}";
    mqttClient.beginMessage(topic);
    mqttClient.print(payload);
    mqttClient.endMessage();
    Serial.println("Published: " + payload);
    lastPublish = millis();
  }
}

Debugging: Resolving 'MQTT connection failed, rc=-2'

When the Serial monitor outputs the exact string MQTT connection failed, rc=-2, the ArduinoMqttClient library is failing to establish a TCP socket to the broker. This is a network-layer rejection, not an MQTT protocol rejection (which would yield codes like rc=4 for bad credentials).

First Three Things to Check

1. Verify Broker Reachability: Open a terminal on your PC and run ping 192.168.1.100. If the Pi or server hosting Mosquitto is asleep or on a different VLAN, the socket will immediately timeout.
2. Check Port 1883 Binding: Ensure your MQTT broker (e.g., Mosquitto) is configured to listen on 0.0.0.0:1883 and not just 127.0.0.1. Check the mosquitto.conf file for the listener 1883 directive.
3. Confirm IP Assignment: Check the Serial monitor during setup(). If the WiFi connection silently failed or grabbed an APIPA address (169.254.x.x), the TCP SYN packet will never reach the gateway.

Ranked Causes for rc=-2

Extending and Simplifying the Build

Once the baseline I2C telemetry and relay control is stable, you will inevitably need to adjust the system complexity based on your deployment environment.

How to Extend the Build

To add local visual feedback without consuming more GPIO pins, wire an SSD1306 128x64 I2C OLED display to the exact same SDA and SCL lines (A4 and A5). The BME280 uses I2C address 0x77 (or 0x76), while the SSD1306 uses 0x3C. Because the addresses do not collide, they share the bus perfectly. Add the Adafruit_SSD1306 library and update the display in the loop() right after reading the sensor. Ensure you add 4.7kΩ pull-up resistors to the SDA and SCL lines if you exceed two devices on the bus to maintain signal integrity.

How to Simplify the Build

If setting up a local Mosquitto broker on a Raspberry Pi or NAS is too much infrastructure overhead, strip out the ArduinoMqttClient library entirely. Replace the MQTT block with the native WiFi library's HTTP client capabilities. Use HTTPClient to send a simple POST request containing your JSON payload to a free webhook service like IFTTT, Maker Webhooks, or a basic Node-RED HTTP endpoint. This reduces the code footprint and eliminates the need to maintain a persistent TCP connection, trading real-time bidirectional control for simpler unidirectional logging.