In electronics, a signal is a time-varying electrical voltage or current that carries information from a source to a destination. Whether it is a 5V square wave toggling a microcontroller pin or a 4-20mA current loop reporting tank pressure, the signal is the physical medium through which data moves.
What a Signal Actually Changes in a Circuit
When we talk about the definition of signal in electronics, it is critical to understand what that signal physically alters in a downstream circuit. A signal changes the state, bias, or operating point of receiving components. It does not necessarily provide the energy to do the heavy lifting; instead, it provides the instructions.
For example, applying a 3.3V logic HIGH signal to the gate of an IRLZ44N MOSFET changes its drain-source resistance from >1 megaohm (off state) to roughly 0.022 ohms (fully enhanced on state). The signal itself only sources a few milliamps to charge the gate capacitance, but that tiny change in voltage allows tens of amps of separate power current to flow through the drain to spin a motor.
Beginners frequently confuse power delivery with signal transmission. A 12V power rail feeding a DC motor is delivering energy (watts). A 5V PWM (Pulse Width Modulation) square wave feeding the motor driver's enable pin is delivering a signal (information). The power rail makes the motor move; the signal tells it how fast to move.
Analog vs. Digital Signals: The Core Distinction
Every electrical signal falls into one of two primary categories based on how it represents information. Understanding this distinction is foundational to reading schematics and writing firmware.
| Criteria | Analog Signal | Digital Signal |
|---|---|---|
| Voltage/Current Levels | Continuous, infinite values within a range (e.g., 0.00V to 5.00V) | Discrete, quantized levels (typically two: Logic LOW and Logic HIGH) |
| Information Encoding | Proportional magnitude (amplitude, frequency, or phase) | Timing, pulse width, or binary bitstreams (1s and 0s) |
| Noise Immunity | Low; any added noise directly corrupts the data value | High; noise must cross a specific threshold to flip a bit |
| Processing Hardware | Op-amps, filters, analog multipliers, ADCs | Microcontrollers, FPGAs, logic gates, shift registers |
| Common Examples | Audio from a microphone, thermocouple millivolt output | I2C/SPI data lines, UART serial, PWM motor control |
To visualize this, think of water pressure in a pipe. An analog signal is like smoothly varying the water pressure to communicate an exact value—for instance, 14.7 PSI means exactly 100 degrees Fahrenheit. A digital signal is like a valve that is either fully open (1) or fully closed (0), sending a code based entirely on the timing of the pulses rather than the exact pressure.
Where You Meet Signals in Practice
You will encounter signal conditioning and conversion constantly when bridging the physical world (analog) with microcontrollers (digital). Let us look at a concrete numeric example involving a common temperature sensor and a popular development board.
The Scenario: You are using an ESP32-WROOM-32 to read ambient temperature via a TMP36 analog sensor.
- The Sensor Signal: The TMP36 outputs an analog voltage signal with a 500mV (0.5V) offset at 0°C, scaling at 10mV per degree Celsius.
- The Math: At a room temperature of 25°C, the sensor outputs:
0.5V + (25 * 0.010V) = 0.75V. - The Receiver: The ESP32 features a 12-bit Analog-to-Digital Converter (ADC), meaning it maps the input voltage to an integer range of 0 to 4095.
Calculating the ADC Reading:
Assuming a perfect 3.3V reference, the digital value the ESP32 registers for a 0.75V signal is:
Digital Value = (Input Voltage / Reference Voltage) * Max ADC Value
Digital Value = (0.75V / 3.3V) * 4095 = 930
When your Arduino/ESP32 code calls analogRead(34), it returns 930. Your firmware then reverses the math to calculate the temperature: (930 / 4095) * 3.3 = 0.75V, then (0.75V - 0.5V) / 0.010 = 25°C.
The math above assumes an ideal ADC. In reality, the ESP32's internal ADC (specifically ADC1 on GPIO32-39) is notoriously non-linear and saturates around 3.1V, meaning a 3.3V input will not yield 4095, and readings in the middle of the curve can drift by ±50mV. If your project requires precision signal measurement (like a medical or scientific instrument), bypass the internal ADC and use an external 16-bit I2C ADC like the Texas Instruments ADS1115. For hobbyist room thermostats, the internal ADC with software oversampling is usually sufficient.
Frequently Asked Questions
What is the exact definition of signal in electronics compared to power?
Power is the rate at which electrical energy is transferred by a circuit, measured in watts (Voltage × Current). Its primary purpose is to perform work, such as heating a coil, illuminating an LED, or turning a motor shaft. A signal, conversely, is a specific variation in voltage or current designed to convey data or control instructions. While a power rail is typically held as steady as possible (e.g., a flat 12V DC), a signal is intentionally varied (e.g., a 0-5V sine wave) because the variation itself is the payload. As noted in standard Analog Devices data conversion literature, signal integrity focuses on preserving the shape of that variation, whereas power integrity focuses on minimizing voltage droop under heavy load.
How does noise affect the definition of a signal in high-impedance circuits?
In high-impedance circuits (e.g., a piezoelectric sensor or a photodiode with a 10-megaohm feedback resistor), the signal current is incredibly small—often in the nanoamp or picoamp range. Because Ohm's Law dictates that V = I × R, even microscopic stray currents induced by electromagnetic interference (EMI) or capacitive coupling from nearby mains wiring will generate massive voltage errors across that high resistance. In these environments, the 'signal' becomes indistinguishable from the 'noise' unless you use guarded traces, shielded coaxial cables, and transimpedance amplifiers placed as physically close to the sensor as possible.
Why is a 4-20mA current signal preferred over a 0-10V voltage signal in industrial wiring?
Industrial environments often require running sensor signals over hundreds of feet of copper wire. If you use a 0-10V voltage signal, the resistance of the long wire creates a voltage drop. If the sensor outputs 5.0V, the PLC at the other end might only read 4.6V due to the wire's resistance, resulting in a false measurement. A 4-20mA current loop solves this. Because current in a series circuit remains constant regardless of wire resistance (within the compliance voltage limits of the transmitter), the PLC reads exactly 12mA whether the wire is 10 feet long or 1,000 feet long. Additionally, a 4mA 'live zero' allows the system to instantly detect a broken wire (which would read 0mA) versus a valid minimum sensor reading.
Can a ground wire be considered a signal return path?
Yes, in single-ended signaling, the ground (GND) wire is absolutely part of the signal path. A voltage is always a potential difference between two points. When an op-amp outputs a 2.5V signal, that 2.5V is measured relative to the ground plane. If the ground wire has high impedance or carries noisy return currents from a switching power supply, the ground potential at the receiver will bounce, effectively adding noise directly to the signal. This is why high-speed or high-precision differential signals (like RS-485, USB, or LVDS) use dedicated twisted-pair return wires for the signal rather than relying on a shared chassis or PCB ground plane.
