Ultrasonic Welding Horn: Design, Materials, Amplitude & Failure Fixes

A battery pack line stops mid-shift. Every stack of copper foils comes out fine until, suddenly, welds start slipping and a hairline crack appears across the horn face. The machine still hums, the settings never changed – the only thing worn out is the ultrasonic welding horn, a $2,000 steel tool vibrating 20,000 times a second until fatigue wins. Understanding that tool is the difference between a stable process and a very expensive guessing game.

What Is an Ultrasonic Welding Horn (Sonotrode) and Why Does It Matter?

An ultrasonic welding horn — technically called a sonotrode — is the shaped metal tool that presses your parts together and vibrates them tens of thousands of times per second until they bond. It is the only component of an ultrasonic welder that physically touches the workpiece. Everything upstream generates, converts and shapes energy; the horn delivers that energy into the joint, precisely and repeatably.

The useful answer is that the horn is where a generic ultrasonic process becomes a specific one. Same generator, same pressure, same weld time, different horn — and a clean battery tab weld becomes a cracked, overheating mess. It is not an accessory bolted to the end of the stack; it is the working tool of the process.

Where the Horn Sits in the Ultrasonic Stack

Ultrasonic metal welding equipment is conventionally a five-part assembly, following the Branson horn catalog:

1. Power supply (generator). Converts line electricity into a high-frequency signal, commonly at 15, 20, 30 or 40 kHz, and tracks resonance as the horn heats during a weld.
2. Actuator / press. The frame and cylinder that hold the stack and push it onto the part with controlled force. Force matters as much as amplitude: a perfect horn under inconsistent pressure still makes bad welds.
3. Converter (transducer). Piezoelectric ceramic discs convert the signal into mechanical longitudinal vibration — motion along the horn’s axis, not side to side.
4. Booster. A tuned intermediate mass that steps amplitude up or down before it reaches the horn and provides the clamping shoulder.
5. Horn (sonotrode). The final tuned element, sized to a fraction of the acoustic wavelength, with a working face that contacts the part and transmits vibration.

Directly beneath the horn sits the anvil, the stationary support that holds the lower part and reflects energy back into the joint. Horn and anvil are a matched pair: a horn ground for a flat anvil behaves differently against a contoured one.

One-Half Wavelength: Why the Length Is Never Arbitrary

A horn is designed to be roughly one-half wavelength long at its resonant frequency. At that length the working faces sit at antinodes of maximum displacement while a stress node sits near the center, where longitudinal motion is close to zero — exactly where you clamp it through a slot or stud, since bending stress there is minimal. Get the length wrong and the horn misses the generator’s frequency, amplitude collapses, and the machine hunts for resonance it never finds.

The converter’s raw output also varies with frequency, per the same Branson data:

Resonant frequency	Typical converter output amplitude
15 kHz	≈ 30 µm
20 kHz	≈ 20 µm
30 kHz	≈ 15 µm
40 kHz	≈ 8 µm

Notice the trade: lower frequency gives bigger strokes but coarser motion and a heavier horn; higher frequency gives finer, more localized energy with less visible travel. Most industrial metal-welding systems sit between 20 and 35 kHz, spanning roughly 16 to 75 kHz by application. Frequency then drives horn length, horn mass, available amplitude and how far energy travels from the weld spot.

Why Horn Design Decides Weld Quality

Every weld-quality variable you care about is downstream of three horn decisions:

• Amplitude at the face. Too little and oxides never break, so no solid-state bond forms. Too much and you over-work soft metals, smear the knurl or crack the part. Amplitude is the most influential setting in ultrasonic metal welding, and the horn sets the ceiling.
• Amplitude uniformity. On a long weld — a busbar or harness ribbon — amplitude must hold steady end to end. A horn stronger at the center than the tips gives a weld that holds in the middle and lifts at the edges.
• Stress distribution. Poorly profiled horns concentrate cyclic stress at the transition, and fatigue cracks nucleate there. A horn can look perfect on day one and fail at hour 400 because the geometry was wrong from the start.

This is also why horn design is joint-specific. A knurled flat face built for copper foil stacks is useless on a wrapped wire termination, and a horn sized for thin foils underperforms on thicker aluminum. Surface condition, joint prep and energy delivery have to agree whether you run ultrasonics or follow a routine like Aluminum Cleaner for Welding: How to Prep Aluminum Right. Clean metal helps, but a horn that cannot put the right amplitude into the interface will not bond.

In practice, the horn is the component you specify first and replace most often. It defines what your machine can do on your parts. Everything that follows in this guide — profiles, materials, amplitude selection, knurl geometry, failure diagnosis — builds on one definition: the horn is a tuned acoustic tool whose design, not the generator behind it, determines the weld.

How an Ultrasonic Horn Works: Wavelength, Resonance and Nodes

To the eye, an ultrasonic welding horn is just a machined block of metal with a stud hole in one end and a knurled face on the other. Yet once the stack is switched on, that block is bending, stretching and compressing twenty thousand times a second, and every millimeter of its geometry is doing acoustic work. Three ideas explain how: wavelength, resonance and nodes.

A Horn Is a Half-Wave Resonator

The converter turns electrical energy into rapid longitudinal vibration that travels along the axis of the stack. The booster hands it to the horn, whose job is to receive that motion, amplify it and deliver it to the workpiece. It can only do that if its length matches the wave passing through it.

A sound wave covers a fixed distance in one cycle, called its wavelength: the speed of sound in the material divided by the frequency. A horn is normally cut to a half-wavelength so that waves reflecting off the free end add to the incoming waves instead of canceling them. At 20 kHz the half-wavelength is roughly 8

How an Ultrasonic Horn Works: Wavelength, Resonance and Nodes

To the eye, an ultrasonic welding horn is just a machined block of metal with a stud hole in one end and a knurled face on the other. Yet once the stack is switched on, that block is bending, stretching and compressing twenty thousand times a second, and every millimeter of its geometry is doing acoustic work. Three ideas explain how: wavelength, resonance and nodes.

A Horn Is a Half-Wave Resonator

The converter turns electrical energy into rapid longitudinal vibration that travels along the axis of the stack. The booster hands it to the horn, whose job is to receive that motion, amplify it and deliver it to the workpiece. It can only do that if its length matches the wave passing through it.

A sound wave covers a fixed distance in one cycle, called its wavelength: the speed of sound in the material divided by the frequency. A horn is normally cut to a half-wavelength so that waves reflecting off the free end add to the incoming waves instead of canceling them. At 20 kHz the half-wavelength is roughly 80 mm and a full wavelength about 160 mm, depending on the material, as described in patent US12583053B2. This is why a titanium horn and an aluminum horn for the same machine are never the same length: sound travels faster through some metals than others, so resonant length changes with the alloy and with the horn’s cross-section profile.

Frequency sets the scale of everything. Lower-frequency horns, typically 15 to 20 kHz, are larger and can drive higher amplitudes into big parts and long seams. Higher-frequency horns at 35 or 40 kHz are smaller, more precise, and suited to fine work such as battery tabs, medical devices and tiny electronic assemblies.

Opposite Ends, Opposite Motion, and a Quiet Center

When a horn resonates, its two ends move in opposite directions, alternately lengthening and shortening it, according to the Branson horn catalog. The working face swings through the widest arc; that maximum back-and-forth travel is the amplitude, and the point of greatest displacement is called an antinode.

Halfway along a half-wave horn sits the stress node. Here, and only here, there is essentially no longitudinal motion. Paradoxically, this is also where internal stress is highest: the material at the node absorbs the strain of everything moving on either side of it. Maximum displacement lives at the faces, maximum stress at the center.

Because the node sits mid-span rather than at an attachment, the horn can vibrate freely while the stack holds it rigidly. If wear, damage or a wrong part load shifts that node, the horn no longer runs at resonance: amplitude drops, the generator hunts for frequency and the nodal region heats up. That is how most horn failures begin.

Gain: How a Horn Makes Motion Bigger

Amplitude is the peak-to-peak displacement of the horn face, so a 60 µm reading means the face travels 60 µm from one extreme to the other. Gain is output amplitude divided by input amplitude: the booster contributes one ratio, the horn’s taper another. A bar horn barely amplifies, a catenoidal horn distributes strain evenly, and a stepped horn delivers the most gain, at the cost of a stress concentration where the section changes.

For metal welding, working amplitudes typically fall between 40 and 80 µm peak-to-peak (US20080054051A1). Below that range the interface never breaks its oxide film and welds come out cold or inconsistent. Push much above it and you trade bond quality for knurl wear, part marking, accelerated horn fatigue and, frequently, stacked material that extrudes instead of bonds.

Amplitude setting	What you see at the weld
Too low	Shiny, unbonded surfaces; inconsistent pull strength; long weld times
In range (40–80 µm p-p)	Uniform smear pattern, stable strength, minimal part marking
Too high	Knurl flattening, edge extrusion, cracked thin sections, short horn life

How Vibration Actually Makes a Weld

With the parts clamped under pressure, the vibrating face forces microscopic slip at the interface. That slip abrades away contamination and oxide, flattens surface asperities and brings virgin metal into intimate contact. Once clean surfaces touch under pressure at this scale, metallic bonding occurs without melting, so the joint forms in the solid state in a fraction of a second. That is why ultrasonic welding suits heat-sensitive materials.

Aluminum shows why surface condition matters. Its oxide layer is tough and refractory, melting around 2,070 °C while the aluminum beneath melts at about 660 °C, so the vibration must fracture the oxide and eject it from the joint zone before bonding can start. Pre-weld cleaning and correct aluminum cleaning for welding make that job far easier, and studies of the interface such as this MATEC conference research describe exactly this sequence of oxide disruption, plastic deformation and bonding.

The practical takeaway: the horn does not heat the parts into a weld, it scrubs them into one. Everything that follows in this guide, from profile choice to amplitude setting, exists to control that scrub, and profile determines how evenly motion spreads across the face. That is where we pick up next, with comparisons of aluminum joining methods in mind when choosing between processes.

Horn Shapes and Profiles: Stepped, Catenoidal, Exponential and Bar Horns

Two horns can run at the same 20 kHz, cut from the same titanium billet, and still behave nothing alike on the press. The difference is geometry. The profile of an ultrasonic welding horn — how its cross-section changes from the mounting stud down to the working face — decides three things: how much amplitude you get, how evenly it spreads across the face, and where peak stress builds inside the tool. Profile choice is where stalled ultrasonic applications come back to life or quietly burn through budgets.

The Four Standard Profiles

Commercial horn catalogs, including the Branson 15–40 kHz sonotrode catalog, group standard sonotrodes into four basic shapes: stepped, exponential, catenoidal and cylindrical (bar) designs. All four solve the same acoustic problem — resizing a vibrating bar while keeping it resonant — but they distribute the resulting stress very differently.

Profile	Amplitude gain	Face uniformity	Peak stress	Typical fit
Stepped	Highest (roughly 1:5 and up, depending on the diameter ratio)	Good on a flat face	Concentrated at the step	Small parts needing max amplitude
Exponential	Moderate	Fair to good	Gradual, lower peaks	General-purpose plastic and light-metal work
Catenoidal (conical-catenoidal)	Moderate to high	Good	Best distributed of the tapered designs	High-duty metal welding
Cylindrical / bar (blade)	None (1:1)	Excellent along the length	Low and spread out	Long seams, blades, multi-tip arrays

Stepped Horns: Maximum Amplitude, Maximum Stress

A stepped horn drops to a smaller diameter over one abrupt shoulder. That shoulder is why it delivers the highest gain of any profile for a given length, making it the default when you need aggressive amplitude in a compact tool. The catch is the step itself: it behaves like a notch, concentrating cyclic stress where cracks like to start. A generous fillet radius, careful surface finish, and a shoulder placed near a stress node go a long way toward extending life. Stepped horns suit small plastic assemblies, riveting, and short welds on thin foils.

Exponential and Catenoidal: Smoother Stress Flow

An exponential horn tapers gradually along an exponential curve. With no abrupt shoulder, peak stress drops significantly compared with a stepped design and the amplitude transition is smoother, though you give up some gain for the same overall length. It is the workhorse profile for general-purpose ultrasonic welding.

The catenoidal (conical-catenoidal) horn is a hybrid: a conical section blended into a catenary curve. The result is a more uniform stress distribution along the body, which is what you want when the tool runs thousands of cycles per shift on metal parts. If amplitude is the priority, pick stepped. If duty cycle and fatigue life matter more, catenoidal is usually the better trade.

Bar and Blade Horns: Trading Gain for Weld Length

When the weld is longer than the horn face is wide, a round horn cannot deliver energy evenly across the seam. Rectangular bar horns, often called blades, solve that. A cylindrical bar horn has no gain at all, but it spreads motion uniformly along its working edge, which is what a long lap seam or a continuous textile seal needs.

Length has a hard acoustic limit. A 2022 finite element study in Trends in Sciences showed a rectangular sonotrode with three slots reaching a 400 mm weld length at 15 kHz — near the practical ceiling for a single blade, and only because the slots suppress lateral modes that would otherwise tear the amplitude field apart.

Multi-Tip Horns and the Battery Tab Problem

Battery pack assembly pushed multi-tip horn design into the mainstream. One-wavelength four-tip horns are widely used for lithium-ion battery tab welding, where a single cycle bonds four or more tabs at once. The known failure mode is tip-to-tip amplitude non-uniformity: some tabs weld fully while neighbors stay cold or get crushed. A MATEC Web of Conferences study of a 40 kHz multi-tip horn found that non-longitudinal modes sitting close to the working mode degrade welding performance, so FEA must isolate the intended longitudinal mode rather than merely confirming resonance exists. A horn that “sounds right” on an impedance analyzer can still be running a bending or torsional mode that wastes energy into the stack.

Multi-tip designs also make surface condition non-negotiable. Aluminum cathode tabs, like every aluminum surface that needs proper prep, must be free of oxide and oil before the horn touches them, because the tip array amplifies any inconsistency in the stack.

Picking the Right Profile in Practice

The decision usually collapses into four questions: how much amplitude do you need, how long is the weld, what material are you joining, and what duty cycle will the horn survive? High amplitude on a short joint points to stepped. Long continuous seams point to a slotted blade. High-volume metal welding points to catenoidal. Multi-part battery or connector arrays point to a tuned multi-tip design. The patent literature, including US20080054051A1, shows slot and tip arrangements still being refined — a reminder that profile selection is engineering, not catalog shopping. For structural aluminum assemblies where fusion joining still rules, our Aluminum MIG Welding guide covers the other side of the same joining decision.

Horn Materials: Titanium, Aluminum, Steel and Specialty Alloys

A horn machined from the wrong metal fails even when the design is flawless. Material sets how much amplitude the tool survives before fatigue cracks nucleate, how fast the knurl wears flat, how much heat builds at the node, whether the horn marks soft parts, and what the eventual replacement costs. No single metal is best here, only the best trade-off between acoustic efficiency, fatigue strength, wear resistance, thermal behavior, and price for your specific parts.

Titanium: The Production Default

Titanium alloys dominate professional horn building, and Ti-6Al-4V, known as 6-4, is the familiar workhorse. It pairs a high strength-to-weight ratio with excellent fatigue resistance and low internal damping, so little vibrational energy is wasted as heat inside the tool and more of it reaches the joint. Among the high-strength alloys, titanium offers the best acoustical properties overall, and it is hard enough that it rarely imprints soft workpieces.

Branson goes a step further, building its catalog horns from Titanium 7-4, roughly 15% stronger than Ti 6-4, according to the Branson horn catalog. That extra strength lifts the fatigue ceiling, so the horn either runs safely at higher amplitude or simply lasts longer at the amplitude you already use.

Strength is not the whole story, though. Titanium’s quiet advantage is heat. An infrared thermography study found that titanium horns showed the lowest node temperature rise of the materials tested and could run at amplitudes up to 77 µm without cyclic failure (Roopa Rani et al., Ultrasonics, 2015), summarized in this PubMed record. Because node heating is where fatigue cracks get started, that finding translates directly into service life.

The costs are real: expensive bar stock, slower machining, and strict bolt-torque discipline during stack assembly. Titanium also still wears when you weld glass-filled resins, films, or abrasive media, so the knurl needs inspection on the usual schedule. Its other natural home is aerospace, where the same grade discipline applies; our Aerospace Welding: Processes, Standards, and Careers Guide gives useful context on how titanium is handled across industries.

Aluminum: Cheap, Efficient, and Easy to Wear Out

Aluminum alloys, commonly 7075-T6, have excellent acoustical properties and the best heat dissipation of the group thanks to high thermal conductivity. They are light, fast to machine, and inexpensive, which makes aluminum the standard choice for short-run work, process trials, and any situation where replacing a horn is cheaper than rebuilding one.

The trade-off is softness. Aluminum knurls flatten faster than titanium, the working face wears, and the horn can emboss or score soft plastics and thin foils, which on visible surfaces marks the part and can scrap product. Two surface treatments slow this down: hard chrome plating and clear anodizing. Both raise surface hardness and extend knurl life, and both should be applied before final facing and frequency tuning, since any layer added to the face changes the mass and shifts resonance.

Aluminum horn users are usually dealing with aluminum parts as well, and the fusion side of that metal has its own rulebook, covered in Aluminum MIG Welding: The Complete Practical Guide.

Steel: Wear Resistance Where It Counts

Hardened tool steels show up in horns, inserts, and tips for abrasive duty: cutting, scoring, and welding filled or fibrous materials that would destroy a softer alloy. Steel resists wear well and raw stock is inexpensive, but its internal damping is higher than titanium’s, so more vibrational energy converts to heat inside the horn. The practical result is a tool that survives abrasion yet runs hotter with a thinner fatigue margin at high amplitude. That is why steel is often the smarter play for anvils, inserts, and replaceable tips rather than the entire horn. Reading alloy grades against service conditions is the same discipline we apply in Alloy Steel Welding Electrodes: Selection, Use & Safety.

Nickel Alloys and Specialty Options

When duty cycle and stress climb past what titanium will comfortably take, horn builders turn to nickel-chromium alloys. Nimonic 80A is the common example: approximately 70% nickel with 18-21% chromium plus titanium and aluminum, precipitation-hardened so it holds its strength as the horn warms during continuous running. It is specified for high-stress horns, meaning long welds, high amplitude, tight cycles, and lines that run most of the day.

Nickel alloys buy fatigue strength and temperature resistance, but they cost more, machine poorly, and sit slightly behind titanium acoustically, so the design has to stay conservative. Further reading is available in this MATEC conference paper.

Material	Acoustics	Wear resistance	Heat behavior	Relative cost	Best for
Titanium 6-4 / 7-4	Excellent	Good	Lowest node rise	High	Production horns, long life at high amplitude
Aluminum (7075-T6)	Excellent	Poor	Best dissipation	Low	Short runs, trials, low-cost replacement
Hardened tool steel	Fair	Excellent	Runs hotter	Low to medium	Abrasive fills, cutting, anvils and tips
Nimonic 80A	Very good	Very good	Holds strength hot	Highest	High-stress, continuous-duty horns

If you take one rule from this section, make it this: match the metal to the failure you can least afford. Can’t afford cracked horns? Titanium, or a nickel alloy for extreme duty. Can’t afford marked parts? Titanium or a coated aluminum horn. Can’t afford the tooling budget? Aluminum, with a replacement interval built into the plan. Amplitude and knurl choices come next, and they only work if the material underneath them is right.

Amplitude, Frequency and Knurl Design: Setting the Working Face

Two horns can be cut from the same bar of titanium, tuned to the same 20 kHz, and run on the same press. One produces consistent, overlapping welds; the other dents the part, shreds the knurl and cracks within a week. The difference usually comes down to three controllable settings: how far the face moves (amplitude), how the booster scales that movement, and what pattern the horn and anvil press into the metal (knurl design). This is where a horn stops being a generic tool and becomes a process-specific one.

Amplitude: The Parameter That Dominates Weld Quality

Amplitude is measured peak-to-peak at the horn working face, in micrometres (µm) or thousandths of an inch. It is the single most influential adjustable variable in ultrasonic metal welding. A 2026 study on multi-layer copper foils for battery tabs found amplitude dominated both weld quality and indentation depth, outweighing weld time within the tested range (multi-layer Cu foil study). Too little and the foil layers never break their oxide films, so the bond stays weak. Too much and the horn sinks into softened metal, thinning conductors and deforming the part.

Practical starting ranges look roughly like this:

• Copper and brass foils: moderate amplitude, typically 20–40 µm peak-to-peak, with weld time doing much of the work.
• Aluminium wire and tabs: lower amplitude and shorter welds, because aluminium softens quickly under frictional heat.
• Nickel-plated battery tabs: watch the plating, since excessive amplitude strips nickel and exposes the substrate.
• Plastics: higher amplitude is normal, because the mechanism is melting rather than solid-state bonding.

Amplitude is not set on the generator. It is the converter’s output multiplied by the booster ratio, then by any gain machined into the horn. Change one of those three and the face velocity changes.

The Aluminium Ceiling: Staying Below ~55 µm

Aluminium punishes amplitude mistakes fastest. Above roughly 50–60 µm peak-to-peak the horn penetrates the softened metal instead of scrubbing it, producing deep indentation, edge extrusion and weakened joints. Patent US20080054051A1 describes holding horn amplitude below about 55 µm to prevent exactly that penetration, which makes a useful concrete ceiling when dialling in a new aluminium job.

Amplitude also only works on clean metal. Oxides, drawing lubricants and rolling oils absorb energy and tempt you to raise amplitude, which is the wrong fix. Degreasing and oxide removal before the press cycle is the cheaper correction, the same surface-prep discipline covered in Aluminum Cleaner for Welding: How to Prep Aluminum Right.

Booster Gain Ratios and Their Colour Codes

Boosters are colour-banded at the factory so a technician can read gain at a glance without measuring. The standard Branson scheme is:

Booster colour	Gain ratio	Typical use
Purple	1:0.6	Reduce amplitude for soft or thin materials
Green	1:1	Neutral, no change
Gold	1:1.5	Moderate increase, common default
Silver	1:2.0	Higher amplitude for firmer metals
Black	1:2.5	High-gain applications
Natural	1:3.0	Maximum gain, use with care

These ratios are documented in the Branson horn catalog. The practical lesson: swapping boosters changes amplitude dramatically without re-cutting a sonotrode. Frequency is not flexible. A 20 kHz booster will not run on a 15 kHz stack, and mismatching them throws the stack out of resonance immediately.

Knurl Design: Fine, Medium and Coarse

The knurl is the pattern machined into the horn face, mirrored or complemented on the anvil, that grips the parts, transmits shear and confines plastic flow. Knurled surfaces are generally classed as fine, medium and coarse, with male knurl dimensions of approximately 0.030″, 0.040″ and 0.050″ respectively.

• Fine knurl: thin foils, delicate tabs and small weld areas where indentation must stay minimal.
• Medium knurl: the general-purpose choice for wire, terminals and moderate foil stacks.
• Coarse knurl: thick straps, heavy multi-layer stacks and jobs needing aggressive grip and material movement.

Newer designs go further. Recent sonotrode and anvil patents use tapered, domed protrusions staggered against the anvil peaks so plastic flow concentrates into controlled zones instead of spreading randomly across the joint (US12583053B2). The payoff is stronger interfacial bonding at lower amplitude, which is exactly what battery and electronics lines want.

Matching Tip Geometry to the Joint

Finally, decide how the horn contacts the part. A flat face suits large flat overlapping surfaces. A stepped or profiled tip concentrates force onto a narrow weld line, giving high pressure without high total amplitude. Contoured tips follow curved parts such as terminals and busbars, while multi-tip horns weld several spots in one cycle. Whatever the shape, horn and anvil knurls must be compatible. A coarse horn over a fine anvil lets material escape sideways, and a mismatched pitch lets the parts slide instead of shearing.

Horn Failure Modes: Cracks, Wear, Fretting and Loose Studs

Horn failures rarely announce themselves. They start as drift: a frequency reading that wanders a tenth of a kilohertz, a stack that suddenly needs more trigger pressure than it did last month, a weld that looks perfect until you cut a cross-section. Because the horn is the one component in the ultrasonic stack that actually touches the work, it absorbs nearly all of the abuse. Learn to name the failure mode and a scrapped $2,000 tool becomes a fixable, predictable problem.

Node Cracking: Where Fatigue Starts

A horn performs roughly 20,000 cycles of expansion and contraction every second, and the stress is not spread evenly. The center of the tool, the displacement node, carries the highest load while contributing nothing to the weld itself. A finite element study of a stepped sonotrode confirmed what experienced welders suspect: the displacement node is the most vulnerable part of the tool, with peak stress concentrating at the node and at sharp transitions such as slot roots and step radii (FEA study on stepped sonotrode stress). Cracks then propagate from those stress raisers until the horn detunes or splits outright.

Over-driven welds are a frequent trigger. In one analysis of 60-layer copper foil stacks, excessive energy input produced cracks and reduced effective thickness on the horn side of the stack, meaning the tool was being asked to deliver more energy than the joint could absorb (copper foil stack welding study). If horns keep cracking in the same spot, suspect amplitude and weld time before you suspect the steel.

Knurl Flattening and Face Wear

The knurl is a wearing part by design. Its job is to bite into the surface and transmit shear, so it erodes. Aluminum horn tips wear away after just several weld cycles in abrasive service, causing downtime and defective parts, which is why aluminum is a poor choice for high-count production tips (horn tip wear patent). Symptoms are easy to spot: the part slips instead of welding, collapse distance increases, and the weld nugget shrinks or disappears.

Practical fixes, in order of cost:

• Re-knurl or re-tip the existing horn if the body is sound and the frequency is still in range.
• Move to a hardened steel or titanium working face for glass-filled resins, copper, abrasive powders, and any joint that scours.
• Clean the face regularly. Aluminum faces build a stubborn oxide film that cushions the knurl; proper surface prep guidance in Aluminum Cleaner for Welding: How to Prep Aluminum Right applies to horn faces as much as to workpieces.

Stud Thread Wear and Loose Bolts

The central stud is the stack’s spine, and threads are a classic fatigue location. Repeated assembly, over-torquing, galling, or a dropped horn can round or stretch the threads. The symptom is usually a loose feel, a rattle at start-up, heat building at the joint, or a frequency that will not hold. Check the stud with a calibrated torque wrench against the manufacturer’s figure, inspect for galling or necking, and replace rather than chase a suspect stud. Never run a horn on a stud you have “tightened a little extra” to stop a rattle; that over-stress shortens the tool’s life measurably.

Booster Interface Fretting

Two flat metal faces pressed together still move microscopically under 20 kHz loading. That micro-motion produces fretting: fine oxide debris, polished or blued patches, localized heating, and eventually a joint that no longer transmits energy cleanly. Common causes are a scratched or nicked mating face, dust or swarf trapped between surfaces, non-parallel faces, and insufficient bolt load. When stud-area failure or friction and heat at the interface is a known recurring issue, manufacturers turn to electron-beam-welded horn-to-booster assemblies, which eliminate the joint entirely (Branson horn catalog). For everything else, meticulous cleaning, flatness checks, and correct torque are cheaper insurance.

Frequency Drift and Thermo-elastic Heating

Horns heat thermoelastically as they work, and metal stiffens as it cools, so a warm horn reads lower in frequency than a cold one. Real drift, however, comes from somewhere else: a cracked horn, worn knurl mass, a loose stud, a detuned manufacturing tolerance, or an added fixture. Once the resonant frequency wanders far enough from the generator’s tracking range, amplitude collapses and the weld goes soft even though the display looks normal. Thermo-elastic heating also feeds itself, because higher amplitude raises nodal temperature, which raises stress, which accelerates fatigue.

Symptom	Most likely cause	First check
Hairline crack at center	Node fatigue	Amplitude, weld time, FEA stress concentration
Slipping, shallow welds	Knurl flattening	Face condition, tip material
Rattle or frequency instability	Loose or worn stud	Torque spec, thread condition
Heat and bluing at joint	Booster fretting	Face flatness, debris, bolt load
Frequency walks down during run	Overheating or crack growth	Nodal temperature, tuning check

Diagnose in this order: read the frequency, feel for heat, look at the face, then inspect the node under magnification. Most “mystery” horn failures reveal themselves in the first minute.

Selecting, Tuning and Maintaining a Horn for Your Application

You cannot buy an ultrasonic welding horn the way you buy a replacement filter. A horn is a tuned acoustic component, custom-cut for one job, and the shops that get the longest life and the most stable welds are the ones that answer a short list of questions before a single chip of titanium is cut. This section walks through that selection workflow, then covers the tuning checks and maintenance habits that keep the horn alive.

Start With the Part, Not the Horn

The most common ordering mistake is calling a supplier and asking for “a 20 kHz horn.” That is like calling a machine shop for “a part.” Branson’s horn catalog guidance is explicit about what an order actually needs to specify, and the list is short but non-negotiable:

• Contact area — the exact footprint of the working face that touches your part, including knurl pattern requirements.
• Part material — thermoplastics, elastomers and metals each demand different amplitude, tip geometry and horn material.
• Expected annual volume — a horn running 50,000 welds a year and one running 5 million are designed very differently.
• Weld requirement — the joint type (butt, lap, energy director), required strength, and any cosmetic constraints on the visible side.

You will find that guidance in the Branson ultrasonic horn catalog, still one of the best free references for standard horn profiles, sizes and ordering conventions.

Weld length drives horn width, and part geometry drives horn shape. A long seam on a plastic housing pushes you toward a wide bar or blade horn, which quickly runs into the amplitude-uniformity problem we covered earlier: the tips of a wide bar horn move more than the center. Slots, tuned contours or a multi-tip design solve that, but only if the designer knows the weld length up front. And if you are joining aluminum components ultrasonically instead of fusion welding them, read our aluminum MIG welding guide first to confirm that solid-state joining is actually the right process for the joint.

Choosing the Frequency

Frequency is a trade between horn size and horn stress. Lower frequency means longer wavelengths, so 15 and 20 kHz horns can be physically large and deliver high amplitude at reasonable stress levels. Move up to 30 and 40 kHz and the horns shrink, which lets you weld tiny medical parts and miniature electronics, but the allowable amplitude drops and tolerance sensitivity climbs sharply. A useful rule of thumb: pick the highest frequency that still lets you build a horn large enough for the weld area and stiff enough to hold amplitude without excessive stress.

FEA Validation: Never Cut Metal Blind

Finite element analysis is now standard practice for any custom horn, and two analyses matter. Modal analysis confirms the horn resonates at the target frequency in the desired longitudinal mode, with the stress node landing where the mounting hardware expects it. Harmonic analysis maps stress concentration and amplitude distribution across the working face, catching hot spots at radii, fillets and knurl transitions before they become crack initiators.

Academic work on horn design repeatedly shows why this matters: stress peaks at profile transitions are the dominant cause of premature fatigue, and FEA reliably predicts them. See, for example, this MATEC conference study on ultrasonic horn design and stress analysis.

Tuning Checks After Installation

A new horn is not ready the moment it is bolted on. Run these checks at commissioning and repeat them on a schedule:

1. Free resonance test. With the stack assembled and unloaded, measure the resonant frequency with an impedance analyzer or the generator’s tuning function. It should sit inside the specified band.
2. Amplitude verification. Use a horn gauge or laser vibrometer to confirm working-face amplitude matches the design value before running production parts.
3. Weld energy baseline. Record stable weld energy, collapse distance and peak power on known-good parts. These become your drift baseline.
4. Watch frequency during welding. Modern control systems continuously monitor resonant frequency drift through the weld cycle, flagging deviations from nominal. This is not just a data log — it is an early warning system for loose tooling, stud wear or a developing crack. The approach is documented in recent patent work, including US12365147B2 on ultrasonic process monitoring.

Storage, Re-Tipping and a Maintenance Schedule

Horns fail from neglect as often as from design error. A simple schedule prevents most of it:

Interval	Action
Every shift	Inspect knurl for flattening, wipe the working face, check stud torque.
Weekly	Verify free resonance; compare weld energy trend to baseline.
Monthly	Inspect booster-horn interface for fretting; clean and re-torque.
Per production run	Confirm part surfaces are clean — contamination accelerates horn wear. Our aluminum welding prep guide covers compatible cleaning approaches.
On wear detection	Send out for re-knurling, re-studding or re-tipping.

Re-knurling restores tip geometry when the knurl wears flat. Re-studding replaces a damaged threaded stud. Re-tipping machines or welds a fresh layer of material onto a worn working face. All three are routine repairs costing a fraction of a new horn, and a growing repair network makes them practical even for custom titanium tools.

Done consistently, this workflow turns horn selection from guesswork into engineering: specify the job precisely, validate the design before cutting metal, verify tuning at installation, and monitor drift so you replace or repair the tool on your schedule, not in the middle of a shift.

Safety, Noise and Common Mistakes with Ultrasonic Horns

Ultrasonic welding looks deceptively gentle. There is no open arc, no glowing puddle, no plume of smoke, so it is easy for a plant to treat the machine as low-risk and skip the same discipline it would apply to a stick or MIG station. That assumption costs people fingers and hearing. An ultrasonic horn is a steel or titanium tool storing real mechanical energy, running next to a rigid anvil, and the failure modes are fast enough that nobody reacts in time. The good news: almost every serious incident traces back to a short list of preventable causes.

Noise: Why an “Ultrasonic” Machine Is Not Silent

The welding frequency itself, typically 20 kHz, sits right at the edge of human hearing. But that is only the fundamental. Real machines also emit subharmonics, broadband noise from cavitation at the weld interface, and plain structural rattle from the booster, fixtures and stack fasteners. The result is that ultrasonic welders are commonly heard as a sharp, high-pitched squeal rather than silence, and operators standing eight hours a day at the same station absorb it continuously.

Regulatory guidance is blunt about this. OSHA’s noise guidance sets 115 dBA as an ceiling that applies regardless of how short the exposure is — no amount of “it was only a second” makes exceeding that level acceptable. The NIOSH Criteria Document 88-110 on welding hazards places noise alongside fumes, radiation, vibration, heat and electrical exposure as hazards that demand controls, not just awareness.

Stored Energy, Pinch Points and Horn Ejection

A horn at working amplitude carries enough stored mechanical energy that a sudden detune or a failing stud can release it violently. A horn that unthreads mid-cycle can be ejected from the stack, which is why stack fasteners are torqued to specification with a calibrated wrench and why a horn should never be pointed toward an operator or a walkway during setup.

The second hazard is the obvious one: the pinch zone between horn face and anvil. Hands near the weld zone during a cycle are the most common serious-injury pathway. Interlocked guarding, two-hand anti-tie-down controls, and fixtures that require the operator’s hands outside the zone are non-negotiable. Note also that a horn stays mechanically live for a moment after power-down; the stored vibration decays, so wait before reaching in.

The Most Common Horn Mistakes, and the Fixes

Mistake	What you see	Fix
Running the horn free-air, without a workpiece contacting the face	Rapid horn heating, premature fatigue cracking, short tool life	Never cycle the generator unloaded; the full amplitude dumps straight into the tool
Wrong amplitude for the material	Burn-through on thin polymer, cold welds on copper, whitening on plastics	Match amplitude to material and thickness; aluminum tooling tops out around 55 µm before fatigue life collapses
New horn paired with a worn anvil	Uneven weld lines, slipping, one side of the joint under-welded	Replace or re-knurl the anvil whenever the horn is replaced; they wear as a matched pair
Over-torqued horn stud	Stud thread wear, cracked node, frequency drift after removal	Torque to the manufacturer’s figure with a torque wrench; never impact-driver a stud on and off
Stacking mismatched booster and horn gains	System runs above the horn’s rated amplitude, cracks appear in days not months	Calculate total gain (booster × horn) and stay inside the horn catalog rating, as set out in the Branson horn catalog

Two more habits quietly destroy horns: hammer-tuning by grinding without re-checking resonance, and stacking mismatched components from different suppliers without a frequency sweep. Both end the same way — a $2,000 tool cracked across the node.

Building Safety Into the Process

Sound horn practice is mostly discipline. Torque specs honored, loads matched, anvils replaced on schedule, and guards respected. If your operation expands into fusion or laser joining, the hazard-assessment mindset transfers directly; see Alpha Laser Welding: Systems, Uses, and Safety Guide for how a different energy source changes the control set. Sourcing correctly rated PPE and consumables is part of the program too, and guides like Airgas Welding Supplies: The Complete Buyer’s Guide keep that procurement honest. For operators moving between ultrasonic and fusion cells, structured training such as an Aluminium Welding Course: What You Learn, Costs & Careers builds the hazard-recognition habit that no machine guard can supply.

Frequently Asked Questions

What does an ultrasonic welding horn actually do?

It is the final acoustic tool in the stack: it resonates at the machine frequency, concentrates and amplifies the converter’s vibration onto the workpiece, and holds the parts under pressure while the joint forms. In short, it delivers the mechanical energy that creates the weld.

How long is a 20 kHz ultrasonic horn?

A horn is normally one-half acoustic wavelength long at its operating frequency. At 20 kHz that works out to roughly 80 mm (about 3.2 inches) depending on the horn material’s speed of sound, which is why titanium and steel 20 kHz horns look similar in length.

Why is my ultrasonic horn cracking?

Most cracks start at the displacement node, where stress concentration is highest and thermo-elastic heating peaks. Typical causes are running amplitude above the material’s fatigue limit, over-tightened or mismatched tooling, and a horn that has drifted off resonance from wear or damage.

Which horn material is best: titanium, aluminum or steel?

Titanium (Ti-7-4 or Ti-6-4) is the best all-round choice: strong, fatigue resistant and dimensionally stable. Aluminum is cheap and acoustically excellent but soft and short-lived; steel is durable and inexpensive for rough duty but has higher acoustic losses and mass.

What amplitude should I run for aluminum?

For aluminum, keep the horn below roughly 55 µm peak-to-peak. Above about 50-60 µm the interface softens and the horn tip can sink into the sheet, leaving deep marks and weakening the parent metal.

Can I machine a horn myself?

Not reliably. A horn must be tuned to its resonant frequency with tight tolerance; an incorrectly tuned horn can damage the converter and power supply. Most shops buy FEA-designed, factory-tuned tooling and limit in-house work to re-knurling or re-studding.

How often should ultrasonic horns be replaced?

There is no fixed interval; it depends on amplitude, duty cycle, abrasive fillers and knurl wear. Track weld energy and frequency drift, inspect knurl height and node area regularly, and plan refurbishment (re-knurl, re-stud) before cracks appear.

What is the difference between a horn and an anvil?

The horn vibrates and delivers ultrasonic energy; the anvil is the static backing tool that supports the part and reacts the clamp force. Their knurl patterns interlock so the top sheet is scrubbed and deformed while the lower part stays supported.

Do ultrasonic welders make harmful noise?

The 20-40 kHz carrier is mostly beyond hearing, but structural rattling and subharmonics produce a sharp audible squeal. Treat it like any industrial noise source: monitor levels and provide hearing protection where OSHA or NIOSH limits may be exceeded.