Contents

Digital Communication
Other Subjects
Section Progress15%

2 of 13 articles

NRZ Line Coding

Non-return to zero, unipolar, polar, bipolar variants.

Darshan N
Updated: 19 March 2026
9 min read

Line coding is the process of converting binary data into electrical signals suitable for transmission over a channel. Among all line coding schemes, Non-Return to Zero (NRZ) is the most fundamental and widely studied format in digital communications. Understanding NRZ is essential before approaching more complex schemes, and it appears frequently in GATE examinations under baseband transmission topics.

NRZ Line Coding VariantsData Bits:101100NRZ-L (Unipolar):+V0VNRZ-L (Polar):+V0-VNRZ-I (Invert on 1):+V-VT2T3T4T5T6TTimeUnipolar: 0V and +V levels onlyPolar: +V and -V levels used
Figure 1: NRZ variants plotted for bit sequence 1 0 1 1 0 0 showing level transitions and encoding logic

Core Concept of NRZ Encoding

In NRZ encoding, the signal level does not return to zero between two consecutive bits of the same type. This is the defining characteristic that separates NRZ from RZ coding. The signal stays at its assigned voltage for the entire bit period, which means the waveform is flat within each bit slot. This simplicity makes NRZ easy to generate and detect, but it also introduces certain drawbacks related to synchronization and DC content.

There are three important variants of NRZ. NRZ-Unipolar uses two voltage levels, typically 0 V and +V. Binary 1 is represented by +V and binary 0 by 0 V. This scheme is simple but has a non-zero DC component and no self-clocking property. NRZ-L (Level) in its polar form uses +V for bit 1 and -V for bit 0, reducing the average DC component when data is balanced. NRZ-I (Inverted) encodes information using transitions rather than absolute levels. A transition occurs at the beginning of a bit period if the bit is 1, and no transition occurs for bit 0.

NRZ-I has an important practical advantage. Because it uses transition detection rather than absolute voltage comparison, it is immune to polarity inversion on the channel. This property is exploited in USB full-speed signaling. However, long runs of zeros in NRZ-I still cause synchronization problems since no transitions occur, making clock recovery difficult at the receiver.

Mathematical Expression and Bandwidth

The power spectral density (PSD) of an NRZ signal for random binary data determines the bandwidth requirement. For polar NRZ with equal probability of 0s and 1s, the PSD is given by a sinc-squared function. The main lobe of this spectrum extends from 0 Hz to the bit rate frequency. Formally, the minimum bandwidth required for NRZ transmission is:

B_NRZ = R_b where R_b is the bit rate in bits per second. This is the null-to-null bandwidth, defined as the first null of the sinc-squared spectrum. Unlike RZ, NRZ requires less bandwidth because the signal energy is concentrated in a narrower spectral region. The bit rate to bandwidth ratio for NRZ is 1 bps/Hz, making it spectrally more efficient than RZ.

The DC component in unipolar NRZ is non-zero because the average value of the waveform is +V/2 when 1s and 0s are equally likely. This DC component cannot pass through AC-coupled systems such as transformers or capacitor-coupled lines, which limits the use of unipolar NRZ in many transmission systems. Polar NRZ reduces but does not eliminate DC content unless the data stream is perfectly balanced.

Practical Understanding of NRZ

One of the key practical limitations of NRZ is the baseline wander problem. When a long string of identical bits is transmitted, the signal stays at a constant level for an extended duration. If the channel has AC coupling, the capacitors begin to charge or discharge, causing the signal baseline to drift. This makes threshold detection unreliable at the receiver and increases the bit error rate.

Synchronization is another challenge. Since there are no guaranteed transitions in the signal, the receiver clock cannot be recovered from the data stream itself. External clock lines or additional synchronization mechanisms are required. This limitation is overcome in Manchester encoding and other self-clocking schemes studied in subsequent sections.

Despite these drawbacks, NRZ coding is used in short-distance communication links, on-chip data buses, and memory interfaces where the clock is supplied separately. SATA, PCIe internal signaling, and many parallel bus standards use NRZ-based encoding with added scrambling to manage long runs of identical bits.

Numerical Example

The null-to-null bandwidth of an NRZ signal equals the bit rate. Given a known bit rate, the minimum transmission bandwidth can be directly calculated. This relationship is foundational and is tested directly in GATE numerical questions. For polar NRZ, since both +V and -V are used, the signal power equals V squared regardless of data pattern.

Example
Given:
Bit rate R_b = 2 Mbps
Voltage levels: +2V and -2V (Polar NRZ-L)

Why this formula applies:
NRZ null-to-null bandwidth = R_b (first null of sinc-squared PSD)

Formula:
B_NRZ = R_b
Power (per symbol) = V^2 / R_L (across load resistance R_L)

Substitution:
B_NRZ = 2 x 10^6 Hz = 2 MHz
If R_L = 50 ohm, Power = (2)^2 / 50 = 4 / 50

Calculation:
B_NRZ = 2 MHz
Power = 0.08 W = 80 mW

Final Answer:
Minimum bandwidth required = 2 MHz
Signal power across 50-ohm load = 80 mW
Exam Tip: In GATE, NRZ bandwidth is always equal to the bit rate R_b. RZ requires twice the bandwidth of NRZ. If a question asks for minimum bandwidth of NRZ, the answer is simply R_b, not R_b/2 or 2R_b. Do not confuse NRZ-L with NRZ-I in polarity inversion questions.

NRZ Mechanism and Signal Transitions

The waveform behavior of each NRZ variant follows a distinct transition rule. Understanding these rules at the waveform level helps in decoding exam questions that give a waveform and ask for the encoding type.

  • NRZ-Unipolar: Signal is +V for bit 1 and 0V for bit 0. No transitions within same consecutive bit runs. DC component exists always.
  • NRZ-L Polar: Signal is +V for bit 1 and -V for bit 0. DC component is zero only when equal number of 1s and 0s are present.
  • NRZ-I: Signal transitions at the start of a bit period only when bit is 1. For bit 0, signal continues at previous level. Starting level is arbitrary.
  • Long runs of 0s in NRZ-I produce no transitions, making clock recovery difficult for runs exceeding 5 to 8 bits.
  • Bandwidth of all NRZ variants is equal to R_b Hz (null-to-null), which is half the bandwidth of RZ for the same bit rate.
  • NRZ has no inherent error detection capability. A corrupted bit cannot be detected from the signal pattern alone.

Loading lab...

Quick Revision

  • NRZ signal does not return to zero between consecutive same-value bits. Signal stays at assigned level for entire bit period.
  • NRZ-Unipolar uses 0V and +V. NRZ-L Polar uses +V and -V. NRZ-I uses transitions to encode 1s.
  • Bandwidth formula: B_NRZ = R_b (null-to-null). This is half the bandwidth of RZ for the same bit rate.
  • NRZ-I is immune to polarity inversion but suffers from synchronization loss during long zero runs.
  • DC component is non-zero in unipolar NRZ and in polar NRZ with unbalanced data. AC-coupled channels cannot support unipolar NRZ.
  • Exam trap: Do not say NRZ bandwidth is R_b/2. The minimum null-to-null bandwidth equals R_b exactly.
  • NRZ has no self-clocking property. External clock or separate synchronization mechanism is mandatory.

NRZ Line Coding Quiz

Test your knowledge of NRZ variants, their DC components, and synchronization limitations.

Question 1 of 3

Q1.Which NRZ variant has zero DC component for equally likely 1s and 0s?