Beamforming in Modern Ultrasound Systems: Physics, Architectures, and Transducers

Introduction

Ultrasound imaging is, at its core, an exercise in wave control. Every ultrasound image you have ever seen is the result of thousands of acoustic waves being launched, steered, focused, reflected, and coherently recombined in real time. The technique that makes this possible is beamforming.

In this article, we take a structured journey through modern ultrasound beamforming:

• Analog vs digital beamforming

• FPGA‑ and GPU‑based architectures

• Why transmit and receive beamforming are handled differently

• The concept of aperture in optics versus ultrasound

• Diffraction‑limited waveforms

• Transducer array types

• And finally, a comparison of PZT and CMUT transducers, including construction details and fabrication challenges

This is written for engineers, researchers, and students who want to understand why ultrasound systems are built the way they are, not just how.

Why Intuitus?

Intuitus brings deep, domain‑specific expertise in ultrasound systems, built through years of focused patent analysis and technology intelligence across beamforming, transducers, signal processing, and system architectures.

Unlike generalist IP analytics providers, Intuitus understands the engineering realities that shape defensible claims in ultrasound — including architectural constraints, hardware–software tradeoffs, and the underlying physics that govern system performance.

By combining rigorous patent analysis with a first‑principles understanding of how ultrasound systems actually work, Intuitus helps organizations identify true white space, assess claim strength and design‑around risk, and make confident R&D, filing, and acquisition decisions in a rapidly evolving ultrasound IP landscape.

Beamforming: Analog and Digital Variants

What Beamforming Does?

Beamforming is the process of controlling when, how strongly, and with what phase each transducer element transmits or receives sound. By doing so, the ultrasound system can:

• Steer the beam electronically

• Focus acoustic energy at a specific depth

• Improve resolution and signal‑to‑noise ratio

Beamforming exists on both the transmit (Tx) and receive (Rx) sides of the system.

Analog Beamforming

Early ultrasound systems relied on analog beamforming:

• Time delays implemented using analog delay lines

• Signals summed before digitization

• Fixed, hardware‑defined behavior

Analog beamforming is inherently deterministic and low‑latency, but it lacks flexibility. Modifying imaging modes or beam characteristics requires hardware changes, making advanced imaging techniques impractical.

Digital Beamforming

Modern ultrasound systems use digital beamforming:

• Each transducer element is digitized independently

• Delays and weights are applied numerically

• Signals are summed in software or programmable hardware

Digital beamforming enables:

• Dynamic focusing

• Synthetic aperture imaging

• Plane‑wave and multi‑angle compounding

• Software‑defined ultrasound pipelines on Receive Side

This transition is what makes a GPU‑accelerated and FPGA‑based ultrasound systems possible.

FPGA and GPU Beamforming Architectures

Why Transmit Beamforming Is Hardware‑Dominated

Transmit beamforming places extreme demands on the system:

• Sub‑microsecond (often nanosecond‑scale) timing accuracy

• Fully deterministic behavior

• Safety‑critical control of acoustic output

• Direct driving of high‑voltage pulsers

These requirements align perfectly with FPGAs or dedicated ASICs.

GPUs, by contrast:

• Operate behind an operating system

• Have non‑deterministic timing

• Cannot directly control high‑voltage hardware

• Are difficult to certify for safety‑critical timing paths

As a result, Transmit beamforming is almost always implemented using FPGAs or ASICs.

Receive Beamforming: Where GPUs Enter

Receive beamforming is computationally heavy but far less timing‑critical. This opens multiple architectural choices.

FPGA‑based receive beamforming

• Ultra‑low latency

• Deterministic timing

• High power efficiency

• Ideal for portable or battery‑powered systems

GPU‑based receive beamforming

• Massive parallelism for delay‑and‑sum operations

• Excellent support for 3D/4D imaging

• Ideal for synthetic aperture and plane‑wave techniques. Synthetic aperture is a signal‑processing technique where data from multiple spatially separated measurements are coherently combined to emulate a single, much larger physical aperture than actually exists. Instead of relying on one large sensor or antenna, the system:

  • Collects echoes from many positions or element activations

  • Applies precise time delays (and phase corrections)

  • Sums them coherently to achieve higher resolution and improved SNR

  • This idea appears across radar, sonar, optics, and ultrasound, and is fundamentally tied to aperture physics and diffraction limits.

• In modern ultrasound, “synthetic aperture” almost always refers to the receive (Rx) side.

  • Rx synthetic aperture is where data from multiple transmit events is stored and later coherently combined during receive beamforming. Enables rapid experimentation and AI‑assisted reconstruction

  • This is feasible because Rx is computationally heavy but not timing‑critical, so it can be done in FPGA/GPU software.

  • This is exactly how plane‑wave imaging, multi‑angle compounding (image stitching), and many research systems work.

  • The concept of synthetic aperture is not theoretically limited to Rx. You can define synthetic transmit aperture (STA) (e.g., one‑element or small‑sub aperture Tx fired sequentially). However, Tx synthetic aperture is rare in clinical systems because:

    • Tx requires nanosecond‑level determinism

    • Is safety‑critical (acoustic output limits)

    • Must directly drive high‑voltage hardware

The Hybrid FPGA–GPU Architecture

Most modern systems combine both worlds:

• FPGA

  • Controls transmit beamforming

  • Handles ADCs and channel preprocessing

  • Packages and streams raw RF data

• GPU

  • Performs full receive beamforming

  • Handles dynamic focusing and compounding

  • Executes 3D reconstruction and image rendering

This hybrid approach is the foundation of software‑defined ultrasound.

Aperture: Optics vs Ultrasound

Aperture in Optics

In optics, the aperture is the physical opening—a slit, pupil, or lens diameter—that limits the wavefront. It is:

• Fixed

• Passive

• Directly responsible for diffraction and resolution

Aperture in Ultrasound

In ultrasound, the aperture is the effective radiating area of the transducer:

• Defined by which elements are active

• Can be changed electronically

• Can grow dynamically with depth

Unlike optics, ultrasound apertures are programmable. This enables:

• Dynamic aperture expansion

• Electronic apodization. Apodization in ultrasound means intentionally tapering the amplitude across the active aperture so that the beam has lower side‑lobes and fewer off‑axis artifacts

Synthetic apertures larger than the physical probe

Your probe might have:

• 128 elements

• 38 mm width

• A fixed physical aperture. This physical aperture sets the diffraction limit θ (beam divergence):

  • θ ≈ λ / D.

Where D is the physical aperture. Physics dictates a diffraction‑limited waveform is the best possible beam a finite aperture can produce and depends on:

• beam divergence which is a function of:

  • is wavelength

  • is aperture size

Key consequences:

• Larger aperture → narrower beam

• Higher frequency → better resolution

• The beam profile is the Fourier transform of the aperture function. The aperture function is the spatial weighting across the probe. The beam pattern is the Fourier transform of the aperture function

  • The aperture function determines the beam pattern

  • The beam pattern determines the main‑lobe width

  • The main‑lobe width gives you the diffraction‑limited divergence

• On receive, each element’s echo signal is multiplied by an apodization weight before summation. The aperture function is continuous in theory, but in a real array you have discrete elements at positions xᵢ.

• Different types aperture functions are:

This principle is shared by ultrasound, optics, radar, and sonar.

Synthetic Aperture = Using Time to Emulate a Larger Space

Instead of firing all elements at once, you fire small sub‑apertures or even single elements, one after another.

For each transmit event:

• You record echoes on all receive channels

• You store the raw RF data

• Subsequently, you integrate all these datasets in a coherent manner by implementing the appropriate delays and phase corrections, allowing the system to function as though you had a significantly larger aperture. This is one of the key advantage Synthetic Aperture.

Transducer Array Types

• 1D Arrays

  • Single row of elements

  • Electronic focusing laterally

  • Fixed elevation focus using an acoustic lens

  • Used in most linear, convex, and phased probes

• 1.5D Arrays

  • Multiple rows with limited control

  • Improved elevation focusing

  • Reduced slice‑thickness artifacts

  • Still no true 3D steering

• 2D (Matrix) Arrays

  • Full row‑column grid

  • Electronic focusing in all directions

  • Enables real‑time 3D and 4D imaging

  • Extremely high channel counts

  • Compute intensive perhaps requires FPGA or FPGA/GPU‑based beamforming

Transducer Technologies: PZT vs CMUT

PZT Transducers

Construction

• Piezoelectric ceramic elements

• Elements separated by kerfs (air or epoxy‑filled cuts)

• Kerfs reduce mechanical coupling and suppress lateral modes

• Matching layers improve acoustic impedance matching

Strengths

• High acoustic output

• Excellent sensitivity

• Mature, reliable manufacturing

• Dominant in deep abdominal, cardiac, and Doppler imaging

CMUT Transducers

Construction

• Micromachined silicon membranes suspended over vacuum cavities

• Electrostatic actuation instead of piezoelectric strain

• Often fabricated as CMUT‑on‑CMOS

• No traditional kerfs:  isolation is achieved through MEMS geometry

Advantages

• Very wide bandwidth

• Low acoustic impedance

• Tight integration with electronics

• Compact probe designs

Where CMUTs Are Used Today

• Handheld ultrasound devices

• Point‑of‑care imaging

• Catheter‑based intravascular ultrasound (IVUS)

• Research platforms

Fabrication Challenges of CMUTs

Despite their promise, CMUTs face real challenges:

• Membrane collapse under bias voltage

• Dielectric charging over time

• Vacuum cavity reliability

• Wafer‑level yield for large arrays

• Packaging stress and long‑term stability

These challenges have limited CMUT adoption in high‑end cart‑based scanners, where PZT still dominates.

Final Thoughts

Modern ultrasound systems are a convergence of physics, transducer engineering, and heterogeneous computing.

• FPGAs dominate transmit beamforming for deterministic, safety‑critical control

• GPUs increasingly dominate receive beamforming for flexibility and computational scale. A deeper patent and non‑patent literature analysis could further illuminate implementation trade‑offs and architectural evolution.

• Aperture physics and diffraction limits unify ultrasound with optics and antenna theory

• PZT remains the workhorse, while CMUTs are reshaping portable imaging