How to Make an Ultrasound? The Physics, Parts, and Build Steps

To understand how ultrasound works, picture a short burst of high-frequency sound leaving a probe, traveling into tissue, and bouncing back from every internal boundary it meets. A piezoelectric crystal sends the pulse and receives the echo, a high-voltage pulser fires the transmit pulse, a low-noise amplifier catches returning signals measured in microvolts, and software converts echo timing into a grayscale cross-section. Building even a basic scanner means wiring that crystal to a pulser-receiver, sampling echoes at tens of megahertz, and sweeping the probe over acoustic gel across a target.

What follows covers the physics, the components, the build sequence, and the safety limits any hobbyist should respect before pointing a probe at something living.

The Pulse-Echo Principle Behind Every Ultrasound Image

A diagnostic probe fires a sound burst lasting roughly half a microsecond, then waits in silence for echoes to drift back, and that round trip is the entire image. Each echo arrives at a different time because sound moves through tissue at about 1,540 meters per second, so a reflector one centimeter deep returns its echo in roughly 13 microseconds.

Why Echoes Form at Tissue Boundaries

Sound waves reflect whenever they cross into a material with a different acoustic impedance, the product of density and sound speed. Fat, muscle, bone, blood, and air each have distinct impedance values, and the larger the mismatch at a boundary, the stronger the echo. Bone reflects heavily, soft tissue reflects moderately, and air reflects nearly all incoming energy, which is why gel becomes mandatory the moment your target has any surface roughness.

Frequency, Depth, and Resolution

Higher frequencies resolve finer detail but penetrate less deeply. A 2 MHz probe can image structures 30 cm below the surface, while an 18 MHz probe resolves sub-millimeter detail but only reaches a few millimeters into tissue. Most clinical scans land between 2 and 18 MHz, depending on whether the goal is a deep abdominal view or a high-resolution look at a tendon.

Resolution also depends on round-trip timing accuracy, because the system must distinguish echoes that arrive only microseconds apart. Shorter wavelengths improve axial resolution, while the geometry of the transducer aperture governs lateral resolution.

Achieving that resolution, however, depends entirely on the hardware that generates, focuses, and records each returning echo.

Core Components of an Ultrasound System and What Each One Does

Every scanner, from a $200,000 cart-based system to a $40 hobby kit, contains the same six functional blocks. Understanding them makes the difference between a pile of parts and a working device.

• ▸Transducer element: A piezoelectric disc, commonly lead zirconate titanate (PZT), that converts voltage pulses into pressure waves and reconverts returning pressure waves into voltage.
• ▸Pulser circuit: A high-voltage switch, often built around fast MOSFETs, that dumps 30 to 100 volts across the crystal for a fraction of a microsecond.
• ▸Receiver amplifier: A low-noise front end that boosts returning echoes from microvolts up to a level the digitizer can resolve above the noise floor.
• ▸Digitizer and timing controller: A microcontroller, FPGA, or DSP that triggers the pulse, samples the returning signal at tens of megahertz, and timestamps each echo.
• ▸Coupling medium: Acoustic gel that displaces air between the probe and the target. Without it, more than 99 percent of the acoustic energy reflects at the surface.
• ▸Display pipeline: Software that converts a stream of samples into a 2D cross-sectional scan, applying beamforming, time-gain compensation, and grayscale mapping.

Why Each Block Matters

The pulser-receiver board sits at the heart of the system, because the same crystal transmits and receives. A protection circuit must switch the sensitive amplifier in within microseconds of the pulse firing, or the high-voltage transmit pulse will destroy the receiver. Beamforming, meanwhile, only works if you have multiple receive channels with precisely matched delays.

Sourcing and Fabricating the Transducer Element

The transducer is the only part you cannot improvise. A bare piezo disc is cheap, but turning it into a working probe requires acoustic matching layers, a damping backing block, electrical contacts, and a mechanical housing.

New Parts Versus Salvaged Probes

A bare 10 mm PZT disc from a ceramics supplier costs a few dollars and works for ranging experiments. For imaging, the crystal needs a quarter-wavelength matching layer on the front and a dense backing material on the rear to shorten the pulse and absorb rearward radiation. Salvaging an element from a retired imaging probe gives you matched layers for free, but the elements are fragile, and de-soldering them without cracking the crystal takes a steady hand.

Single-Element Versus Array Transducers

One scan line per pulse limits single-element probes, forcing a mechanical sweep across the target to assemble any 2D image. Phased arrays and linear arrays use dozens of elements fired in sequence, with small delays between them, to steer and focus the beam electronically. Cutting a PZT disc into strips and wiring each strip to its own channel creates a crude linear array suitable for B-mode imaging without commercial parts.

Once the element itself is ready, the real engineering challenge becomes shaping and interpreting the electrical signals driving it.

“Any delamination between the crystal and its electrodes kills both transmit and receive sensitivity. Keep the bonding surfaces clean and use a thin, uniform layer of conductive epoxy.

Building the Driver, Receiver, and Signal-Processing Chain

The electronics chain has to fire a sharp pulse, switch to receive within microseconds, amplify a faint echo, digitize it at high speed, and hand the samples to a processor for image reconstruction.

The Pulser and T/R Switch

A simple pulser uses a MOSFET or a high-speed transistor to discharge a capacitor across the crystal. Pulses of 30 to 100 volts lasting 50 to 200 nanoseconds work well for frequencies in the 2 to 10 MHz range. A transmit/receive switch, often a pair of back-to-back diodes, isolates the sensitive amplifier during the high-voltage pulse and opens the receive path within a few hundred nanoseconds.

Amplification and Digitization

Returning echoes arrive at microvolt levels, so a low-noise preamp with 40 to 60 dB of gain sits right at the probe. Time-gain compensation applies more amplification to echoes from deeper structures, flattening brightness across the image. Digitization must sample at least twice the highest frequency of interest, so a 10 MHz signal needs a 50 megasample-per-second ADC to capture its full bandwidth.

Beamforming and Image Reconstruction

For a phased array, the beamformer applies precise time delays to each channel before summing, steering the beam and focusing it at a chosen depth. The summed signal becomes one scanline. Sweeping the delays across a sequence produces a sector scan, and stacking scanlines builds a B-mode frame. Open-source projects have demonstrated working beamformers on Raspberry Pi and FPGA platforms.

Assembling a Basic DIY Scanner and Running a First Test

A minimal working setup needs only a transducer, a pulser-receiver, an ADC, and a laptop running plotting software. The goal at this stage is not medical imaging but verifying that echoes return and timing scales match known depths.

Wiring the Probe to the Electronics

Connect the crystal to a commercial pulser-receiver board or build a homebrew MOSFET switching circuit around an IRF510 or similar fast transistor. Coaxial cable with characteristic impedance around 50 ohms keeps ringing under control. The receive output feeds an ADC, often a USB oscilloscope or a dedicated digitizer, which streams samples to a PC.

Streaming Echoes to a Computer

Python with NumPy and Matplotlib, MATLAB, or LabVIEW can all plot a single A-line, the one-dimensional echo profile along the beam axis. Use a water tank with a flat reflector at a known depth, such as a metal plate 5 cm below the surface, as a calibration target. The echo should appear at roughly 65 microseconds, matching the 7.7 cm round-trip distance at 1,540 m/s.

Sweeping for a 2D Image

Mount the transducer on a motorized slide or a manually scanned carriage so each A-line is taken at a known lateral position. Stack the A-lines into a 2D array and map amplitude to brightness. Even a slow manual sweep can produce a recognizable B-mode image of a layered phantom, a piece of foam with embedded voids, or a fruit with seeds inside.

Seeing a recognizable image is one thing; understanding what a hobbyist rig can responsibly be pointed at is another.

“Apply acoustic gel liberally. Even a hairline of air between the probe and the target ruins signal transmission and produces a blank or noisy image.

Safety Limits, Legal Boundaries, and What a DIY Build Cannot Do

Diagnostic ultrasound is regulated because high-intensity sound can heat tissue and, in extreme cases, trigger inertial cavitation. Manufacturers must demonstrate that their devices stay within safe output limits, and the AIUM publishes guidelines that define safe clinical practice.

Output Limits and ALARA

Commercial machines display Mechanical Index and Thermal Index in real time, letting sonographers track cavitation risk and heating as they scan. Output stays capped under the ALARA principle, As Low As Reasonably Achievable, to keep exposure minimal. Hobby builds have none of these safeguards, no calibrated power meter, no exposure tracking, and no thermal model.

What a DIY Device Can and Cannot Replace

A homebrew scanner is a powerful teaching tool. It demonstrates pulse-echo physics, beamforming math, and signal-processing pipelines in a way no textbook can. It cannot replace a regulated diagnostic device for medical imaging. Diagnosing a gallbladder obstruction, measuring fetal growth, or guiding a biopsy requires calibrated output, validated image quality, and the regulatory clearance that only certified equipment carries.

The piezoelectric effect, discovered by the Curie brothers in 1880, is the foundation of every probe, hobby or clinical. Pairing that physics with modern low-cost analog front ends and open-source beamforming software is steadily closing the capability gap between garage-built prototypes and clinical scanners, but the regulatory gap remains wide.

Bottom Line

Ultrasound is sound waves, a piezoelectric crystal, and clever timing. Building a working scanner means wiring those pieces together, calibrating with a known target, and respecting the safety limits that keep sound from becoming a hazard.

FAQ