The Costas loop is a closed-loop feedback system that recovers the suppressed carrier from double-sideband suppressed-carrier (DSB-SC) and binary phase-shift keying (BPSK) signals. Unlike a standard phase-locked loop (PLL), it multiplies the incoming signal with two quadrature local oscillator outputs—an in-phase (I) arm and a quadrature (Q) arm—and multiplies the resulting baseband products to generate a phase error signal that is independent of the modulating data's polarity.
Glossary
Costas Loop

What is Costas Loop?
A Costas loop is a phase-locked loop architecture designed for carrier recovery and coherent demodulation of suppressed-carrier modulation schemes, using an error detector that is inherently insensitive to phase ambiguity.
Invented by John P. Costas at General Electric in 1956, the architecture solves the fundamental problem of synchronizing a local oscillator to a signal that lacks a discrete carrier component. The I-arm performs coherent demodulation once locked, while the Q-arm output remains near zero. The loop's squaring effect in the error detector inherently resolves the 180-degree phase ambiguity of BPSK, making it a foundational building block in software-defined radio (SDR) and modern digital receiver chains.
Key Features of a Costas Loop
A Costas Loop is a phase-locked loop (PLL) architecture specifically designed for carrier recovery of suppressed-carrier modulation schemes, such as BPSK and QPSK. Its defining characteristic is an error detector that is insensitive to the 180-degree phase ambiguity inherent in these modulations.
Suppressed-Carrier Recovery
Unlike standard PLLs that lock onto a discrete carrier tone, the Costas Loop recovers the carrier from a suppressed-carrier signal where the carrier is absent from the transmitted spectrum. It does this by using the modulation itself to generate a phase error signal. The loop effectively squares or multiplies the incoming signal to create a discrete frequency component at the carrier frequency, which the PLL can then lock onto. This is essential for demodulating BPSK and QPSK signals, where transmitting a carrier would waste significant power.
Phase Ambiguity Resolution
A critical feature of the Costas Loop is its inherent tolerance to phase ambiguity. In a BPSK Costas Loop, the error detector output is proportional to the sine of twice the phase error (sin(2θ)), not the phase error itself. This means the loop has stable lock points at both 0° and 180°. While the loop locks, there is a 180-degree phase ambiguity in the recovered carrier, which is resolved by using differential encoding of the data at the transmitter. This prevents catastrophic data inversion.
I/Q Arm Architecture
The classic Costas Loop uses two parallel arms: the In-Phase (I) arm and the Quadrature (Q) arm. The local Voltage-Controlled Oscillator (VCO) feeds the I-arm directly and the Q-arm through a 90-degree phase shifter. Both arms mix the incoming signal down to baseband. The outputs of the I and Q arms are then multiplied together in the phase detector. The product of these two baseband signals yields a DC term proportional to the phase error, which is filtered and fed back to control the VCO, closing the loop.
Hardware vs. Software Implementation
Costas Loops can be implemented in analog hardware or digitally in an FPGA or DSP. A digital Costas Loop offers superior precision and flexibility. Key digital components include:
- Numerically Controlled Oscillator (NCO): Replaces the analog VCO.
- Direct Digital Synthesizer (DDS): Generates the precise sine and cosine mixing signals.
- Loop Filter: A digital proportional-integral (PI) controller that defines the loop's dynamic behavior, including its damping factor and natural frequency. Digital loops avoid DC offset and component drift issues common in analog designs.
Loop Filter Dynamics
The loop filter is the control system's brain, defining the trade-off between acquisition speed and phase noise rejection. It is typically a second-order filter. Key design parameters include:
- Noise Bandwidth (BL): A narrow bandwidth filters more noise but slows acquisition. A wider bandwidth locks faster but results in a noisier recovered carrier.
- Damping Factor (ζ): Usually set near 0.707 for a critically damped response, balancing fast settling with minimal overshoot. The filter integrates the phase error to drive the steady-state phase error to zero, ensuring coherent demodulation.
Modified Costas Loop for QPSK
For QPSK and higher-order QAM, the basic Costas Loop is modified. A QPSK Costas Loop requires additional logic to handle the four-phase ambiguity (0°, 90°, 180°, 270°). This is often achieved using a hard-limited Costas Loop or a digital decision-directed variant. The error detector uses the sign of the I and Q data decisions to strip the modulation, generating an error signal proportional to sin(4θ). This ensures stable lock points at multiples of 90 degrees, which are again resolved through differential encoding.
Frequently Asked Questions
Explore the core mechanisms, design trade-offs, and practical applications of the Costas loop, a critical carrier recovery circuit for suppressed-carrier modulation schemes.
A Costas loop is a phase-locked loop (PLL) architecture specifically designed for carrier recovery of suppressed-carrier modulation schemes, such as Binary Phase-Shift Keying (BPSK) and Quadrature Phase-Shift Keying (QPSK). Unlike a standard PLL, it uses an error detector that is insensitive to the 180-degree phase ambiguity inherent in these modulations. The loop works by splitting the input signal into two parallel branches: an in-phase (I) arm and a quadrature (Q) arm, driven by a voltage-controlled oscillator (VCO) with a 90-degree phase shift between them. The product of the low-pass filtered outputs from these two arms generates an error voltage proportional to the phase difference. This error signal drives the VCO to lock onto the suppressed carrier frequency, enabling coherent demodulation without a discrete pilot tone.
Enabling Efficiency, Speed & Accuracy
Intelligent Analysis, Decision & Execution
We build AI systems for teams that need search across company data, workflow automation across tools, or AI features inside products and internal software.
Talk to Us
Search across company data
Give teams answers from docs, tickets, runbooks, and product data with sources and permissions.
Useful when people spend too long searching or get different answers from different systems.

Automate internal workflows
Use AI to route work, draft outputs, trigger actions, and keep approvals and logs in place.
Useful when repetitive work moves across multiple tools and teams.

Add AI to products and internal tools
Build assistants, guided actions, or decision support into the software your team or customers already use.
Useful when AI needs to be part of the product, not a separate tool.
Related Terms
The Costas Loop is a foundational architecture for coherent demodulation. These related concepts define the impairments it corrects, the signals it processes, and the performance metrics used to evaluate its effectiveness.
Carrier Frequency Offset (CFO)
The difference in frequency between the transmitter and receiver local oscillators. In a baseband signal, CFO manifests as a rotating phase error that causes the IQ constellation diagram to spin at a constant rate. The Costas Loop's primary function is to lock onto and nullify this offset, preventing symbol errors in coherent detection schemes like QPSK.
IQ Constellation Diagram
A two-dimensional scatter plot mapping the in-phase (I) component on the x-axis against the quadrature (Q) component on the y-axis. The Costas Loop stabilizes this diagram. Without carrier lock, the constellation rotates; with lock, discrete symbol points become clearly defined. This visualization is the primary tool for diagnosing phase error and loop performance.
Error Vector Magnitude (EVM)
A comprehensive metric quantifying the deviation of measured constellation points from their ideal reference positions. Residual phase noise from an imperfectly locked Costas Loop directly degrades EVM. It captures the combined impact of all signal impairments—including phase error, amplitude imbalance, and noise—in a single decibel value.
Direct Conversion Receiver
Also known as a zero-IF or homodyne architecture, this receiver downconverts the RF signal directly to baseband in a single mixing stage. The Costas Loop is essential here because the local oscillator is at the exact carrier frequency, making the receiver highly sensitive to phase and frequency offsets that must be corrected digitally.
Complex Baseband
A frequency-shifted representation of a bandpass signal centered at zero hertz. The Costas Loop operates entirely within this complex baseband domain, processing the I and Q sample streams to extract phase error. Representing the signal as a complex exponential simplifies the mathematical derivation of the loop's error detector.
IQ Imbalance
A hardware impairment where gain and phase mismatches between the I and Q branches cause mirror-frequency interference. While the Costas Loop corrects carrier phase, it does not inherently fix IQ imbalance. In practice, a separate IQ correction block often precedes the loop to ensure the error detector receives orthogonal signal components.

About the author
Prasad Kumkar
CEO & MD, Inference Systems
Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.
His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.
Partnered with leading AI, data, and software stack.
How We Work
Custom AI workflows for your Business
One-fit-all AI don't work for modern businesses. At Inferensys, we aim to understand your business & custom requirements; which we use to define most efficient agentic workflows, the data, and the tools for your business.
01
Review the use case
We understand the task, the users, and where AI can actually help.
Read more02
Pick the right approach
We define what needs search, automation, or product integration.
Read more03
Build the first useful version
We implement the part that proves the value first.
Read more04
Improve from there
We add the checks and visibility needed to keep it useful.
Read moreThe first call is a practical review of your use case and the right next step.
Talk to Us