Plastic Sonic Welding Explained: Ultrasonic Process Guide

There is a sound on a plastics production line that experienced operators can identify from three rooms away: a rising, shrieking tone that lasts half a second and then stops dead. That brief howl is an ultrasonic welder joining two injection-molded housings with no adhesive, no screws and no heat gun. If you have ever wondered how two pieces of hard plastic become a single airtight shell in under a second, this guide explains the process that manufacturing engineers call ultrasonic welding and that the shop floor calls plastic sonic welding.

What Plastic Sonic Welding Actually Is (and Why the Name Confuses People)

Start with the naming. Plastic sonic welding is ultrasonic welding. There is no separate process, no different physics, no different equipment. “Sonic welding” is a shortened trade name, while “ultrasonic welding” is the standard term in process documentation, equipment manuals, test standards and academic literature. A vendor saying sonic welding means an ultrasonic plastic welder.

It is a solid-state joining method: high-frequency mechanical vibration becomes localized frictional heat at the joint, while a holding force squeezes the parts together and the molten surface layer fuses. Herrmann Ultrasonics describes it plainly: the vibration of a tool called a horn is transmitted into the plastic, surface friction at the interface generates heat, and the joint solidifies as one piece. Nothing melts in bulk, nothing evaporates, and no liquid chemical is part of the bond itself.

The three numbers that define the process

Three headline figures explain why the process behaves as it does:

• Frequency: 20–70 kHz. That is Herrmann’s range for plastic ultrasonic welding. Medical device practice narrows it to typically 25–40 kHz with amplitudes of 1–25 µm.
• Amplitude: micrometres, not millimetres. The horn tip moves only a few thousandths of a millimetre, yet delivers enough energy to melt a thin layer of thermoplastic. Herrmann quotes typical amplitudes of 5–50 µm.
• Cycle time: 0.5–2 seconds, usually faster. Fictiv puts typical ultrasonic weld cycles at 0.5 to 2 seconds, and Branson reports most welds complete in under 500 milliseconds.

What it is not: four common mix-ups

Vibration welding is the closest relative and the most common confusion. It runs at much lower frequencies, 100–300 Hz according to Fictiv, so it is slower and less precise. It suits large, flat parts where the whole seam melts at once, which makes ultrasonic welding look extremely quick by comparison.

Spin welding uses rotational friction instead of vertical vibration. One part spins against the other until the surfaces melt, making it excellent for round containers such as bottles and canisters, and far less useful for complicated profiles.

Hot plate welding heats a metal plate that presses against the parts. It is simple and reliable, but it soaks the whole seam with heat: cycle times run from tens of seconds to minutes and parts can deform or sink around the weld. Hot-air and hot-blade plastic repair tools, sold to consumers and used at electronics benches, work the same way in miniature. They are useful for a single cracked housing, but they are not a repeatable production process and carry no hermetic guarantee.

The core promise

The appeal is a short list of concrete benefits. According to Medical Design Briefs, the process needs no ventilation, introduces no contamination, is energy efficient and cleanroom compatible, produces airtight and liquid-tight weld lines, and delivers short cycle times with high weld strength.

Operators also notice what is missing: no consumables to reorder, no adhesive to mix, no cure time to wait out, no solvent vapours to ventilate, and every cycle after the first is nearly identical to the last. A half-second weld can run thousands of times an hour with one horn and a custom nest. That mix of speed, cleanliness and repeatability is why the process ended up inside everything from medical filter housings to mobile phone enclosures.

How it stacks up against other joining methods

No joining method wins everything. Here is the honest comparison, with cycle times as order-of-magnitude figures rather than guarantees:

Joining technology	Typical cycle time	Consumables	Hermetic seal possible?
Ultrasonic (plastic sonic welding)	0.5–2 s, most under 500 ms	None	Yes, reliably
Vibration welding	Several seconds	None	Yes
Spin welding	1–5 s	None	Yes
Hot plate welding	Tens of seconds to minutes	None	Possible, heat-sensitive parts struggle
Laser welding	Sub-second to ~2 s	None	Yes
Solvent cementing	Seconds plus set time	Solvent	Yes
Adhesive bonding	Minutes to hours of cure	Adhesive, primer or mixer	Only with careful joint design
Mechanical fastener	Seconds to minutes	Fastener, plus a gasket if sealed	Only via a separate seal

Ultrasonic welding wins the speed and cleanliness columns and rarely loses the hermetic column. It gives up on capital cost, maximum joint size, and joining two materials that will not weld to each other, which is where adhesives and fasteners still earn their place. The same process-versus-material logic runs in Air Gas Welding Explained: Process, Equipment, Safety, and Alpha Laser Welding: Systems, Uses, and Safety Guide covers the laser option above in more depth.

With the naming straightened out, the rest is mechanical: how a stack of generator, converter, booster and horn turns electrical energy into a half-second bond, which plastics cooperate and which fight it, and why joint shape decides the outcome before anyone touches a settings knob.

How the Physics Works: From Electrical Signal to Solidified Joint

Ultrasonic welding is a solid-state joining process. Nothing burns, no filler is added, and no flame appears in the cycle. A tool shakes two plastic parts against one another at ultrasonic frequency while squeezing them together, and friction at that interface converts vibration into heat until a thin layer of resin melts and fuses. Unlike air gas welding, where heat arrives with a torch, the heat here is generated inside the joint itself, which is why the process is fast and clean.

The Ultrasonic Stack: From Mains Power to a Vibrating Tool

The machine is a chain of components, each feeding the next. The power supply (generator) takes line voltage and turns it into high-frequency electrical energy at the chosen operating frequency, typically 15 to 70 kHz. The converter, or transducer, is where electrical becomes mechanical: piezoelectric ceramics flex slightly under an alternating field, and a coil in the converter head excites them to vibrate. A 20 kHz converter typically delivers about 20 µm of vibration.

The ultrasonic stack, top to bottom

Component	What it actually does
Power supply (generator)	Converts line voltage into high-frequency energy at the set frequency
Converter (transducer)	Piezoelectric ceramics turn electrical energy into vibration
Booster	Scales amplitude up or down for the part and load
Horn (sonotrode)	Tuned titanium tool that concentrates vibration at the joint
Actuator	Applies weld force and drives the horn into the part
Nest / anvil	Rigid fixture locating the lower part and supporting the joint

Each element multiplies or divides amplitude, so the useful number is the product of the gains, not the converter rating alone. A converter at 20 µm feeding a 2:1 booster and a 1.5:1 horn delivers roughly 60 µm at the joint face. Branson design guidance puts typical amplitudes for polycarbonate at 60 to 100 µm and notes many plastics will not weld properly below about 25 µm.

Amplitude Governs the Rest

Amplitude is the peak-to-peak movement of the horn tip, and it stays the dominant process parameter against weld time, force and energy. Research on the minimum amplitude required for ultrasonic welding is blunt: too little vibration and the interface never reaches melt, however long the cycle is held, and extra time does not recover the lost strength. Too much and the melt runs ahead of the joint, flash appears and bond strength drops. Modern welders therefore trim amplitude electronically instead of only changing the physical stack. Fictiv’s overview of the process describes the same chain in practical terms, and most production machines hold amplitude, force, downspeed and time under closed-loop control.

Friction Is the Heat Source

Heat appears in two places at once. At the interface the rough surface of one part rubs against the other, converting vibration into thermal energy. Inside the material, intermolecular friction adds more heat as the chains begin moving against each other. Amplitude and force together set the rate.

Once the temperature crosses the melt point, the resin liquefies. Polymer chains from the two parts diffuse into each other, interpenetrate and entangle. When power is removed and the parts cool, those chains lock into a single continuous network. The result is not a glued or mechanically fastened joint but a genuine molecular weld.

Because energy goes into the joint rather than the surrounding part, the tooling can measure what happened. Weld energy, in joules, is the average power dissipated in the joint multiplied by weld time, one of the most useful numbers for comparing a good cycle against a bad one.

The Four Phases of a Weld

The cycle unfolds in four stages, as Herrmann Ultrasonics describes in detail.

1. Contact and initial melt. The horn descends, the energy director contacts and collapses, and displacement rises as the first material melts.
2. Parts melting together. A developing melt layer forms across the joint and the two parts begin fusing as one body.
3. Steady melt layer. The melt reaches a roughly homogeneous temperature and the collapse rate settles.
4. Power off, hold, cool. Vibration stops, force is maintained, the melt solidifies and the bond forms.

That fourth phase is why hold time matters as much as weld time. Research published through Plastics Decorating found optimum weld times of roughly 0.8 s for PMMA and 1.0 s for ABS, and that pushing well past those points reduced strength or at best failed to improve it. Weld longer than the material needs and you are degrading the interface.

Near Field and Far Field Welding

One more piece of physics decides whether a joint is practical. If the horn tip sits within about 6 mm of the joint, the weld is in the near field. Beyond 6 mm it becomes far field welding, where the vibration travels further through the material and loses energy on the way.

Near-field welding gives the operator a wider processing window and fewer marks where the horn touches the part. Far-field welding is limited to small joints and forgiving materials. In practice, far-field welding is only practical for semi-crystalline plastics under narrow conditions, because those resins need a concentrated, well-controlled energy input to reach their sharp melting point.

Which Plastics Weld Well — and Which Ones Will Fight You

The rule that eliminates most of the guesswork: only thermoplastics can be joined by ultrasonic (sonic) welding. As Herrmann Ultrasonics notes, apart from PTFE, all thermoplastics can be welded this way. Elastomers and thermosets cannot.

The reason is molecular structure. A thermoplastic is built from linear chains held together by weak secondary bonding. Heat them and the chains slide past one another, soften, and can be persuaded to interlock with a neighbouring part. An elastomer has a cross-linked network that springs straight back the moment pressure is released, so it will not hold a reshaped weld. A thermoset is so heavily cross-linked that heating makes the network disintegrate rather than melt: it chars and degrades instead of fusing. Phenolic, epoxy and melamine are off the table entirely.

PTFE (Teflon) is the odd exception among thermoplastics. Its chains are so stiff and its melting behaviour so unusual that it cannot be melted at ordinary processing conditions, so it cannot be ultrasonically welded either.

Amorphous versus semi-crystalline

Amorphous resins such as ABS, polycarbonate, polystyrene and PMMA have no long-range crystalline order. They soften gradually over a broad temperature band rather than melting at a sharp point, and the melt stays fluid long enough to wet out the joint. That is why they weld readily with modest energy and forgiving designs.

Semi-crystalline materials such as nylon, polypropylene, polyethylene and acetal contain ordered crystalline regions mixed with amorphous ones. They have a sharp melting point, high shrinkage as they solidify, and a narrower processing window. They need more energy and punish sloppy joint geometry. As Plastics Decorating has documented, the base material alone can swing weld quality dramatically even when every machine setting stays the same.

The typical welding windows reported by Herrmann:

Material	Welding temperature window	Watch out for
ABS	270–310 °C	Amorphous, forgiving; generous energy window
PS	270–310 °C	Amorphous, but brittle; not for impact parts
PC	270–310 °C	Amorphous; prone to crazing from uneven stress
PP	230–280 °C	Semi-crystalline; burns fast, minimal flash tolerance
PE	220–280 °C	Semi-crystalline, low stiffness, weak energy director
PVC	250–280 °C	Burns fast; releases hydrochloric acid as it degrades
PMMA	120–180 °C	Low window; sensitive to internal stress and crazing

The four properties that decide how a plastic welds

Once the resin family is chosen, four properties explain most of what you see on the machine:

• Elastic modulus (shear modulus). Governs how efficiently mechanical energy travels through the part and down to the joint. Stiff, well-supported sections conduct the vibration where you want it; soft sections absorb it before it arrives.
• Damping, or mechanical loss factor. Sets how much of the vibration is converted into heat at the interface. Damped or heavily filled grades need different energy settings than clean, stiff grades.
• Melting point. Sets the heat demand. A high-melting polymer needs more energy per weld to reach a molten interface; a low-melting one can be burned through before it flows.
• Melt viscosity. Controls how the melt spreads and how even the finished joint will be. Too viscous and it will not fill the joint; too fluid and it squeezes out as flash.

Fillers and additives quietly change the rules

Two findings from Herrmann are worth memorising. First, flame-protection additives make plastics harder to process with ultrasonics, because the additive package changes melt behaviour and damping characteristics. Second, glass fibres and glass beads improve sound conductivity, which helps in semi-crystalline plastics that would otherwise dissipate the vibration before it reaches the joint.

Both effects cost something. Every filler displaces polymer chain in the weld zone, so there is less material available to interlock. A housing may carry 30 percent glass and still hold a perfectly good weld, but the bond comes from the remaining resin, and the fibre is abrasive against the horn face. The same controlled trial-and-error approach you would apply to a metal program, such as our aluminum MIG welding guide, pays off here too. Aerospace is a good example; our aerospace welding guide covers the metal side of that industry.

The moisture trap nobody warns you about

Polyamides are hygroscopic, absorbing water from the air until moisture becomes a manufacturing variable. In a case documented by Madison Group, nylon parts that sat in inventory for a week absorbed roughly 2.5 percent water by part weight. Welded, that water turned to steam at the interface and produced porosity, voids and a porous weld line that looked acceptable until leak tested or pull tested.

The fix is unglamorous but effective: weld polyamide parts dry, as molded. Control drying per the resin supplier’s guidance, keep parts in sealed bags or dry cabinets between molding and welding, and log moisture status in the traveler next to the machine settings. Our overview of ultrasonic weld design makes the same point about geometry: material condition is an unaccounted variable that quietly ruins good parts.

I’ll analyze the fragment length first, then rewrite it to fit the640–960 word target while keeping the heading, facts, and links intact.

Joint Design: The Geometry That Makes or Breaks a Weld

The machine is the easy part. Guidance across the plastics welding industry keeps landing on the same conclusion: roughly 70% of ultrasonic welding problems trace back to part design rather than machine settings. Dial in amplitude, force and time until the numbers look perfect and you will still produce weak, leaky or flashed parts if the geometry is wrong. The ultrasonic welding design guides from equipment makers are blunt about it. Get the joint right and welding is almost boring; get it wrong and no machine setting rescues you.

Energy directors: the little ridge that starts everything

An energy director is a small ridge molded onto one of the two mating parts, usually triangular or trapezoidal in cross-section. Its job is to concentrate the horn’s vibration into a tiny area. Without it, energy spreads across the joint face, the resin warms unevenly, and the weld never fuses or fuses patchily. With it, the sharp apex hits melt point almost instantly, the ridge collapses, and the surrounding material softens around it to form the bond. That phase is covered in the Herrmann guide to ultrasonic welding of plastics.

Published recommendations give a workable window:

• Fictiv recommends a director height of 0.2 to 0.5 mm (0.008 to 0.020 in) with an apex angle of 45° to 60°, using the lower end of that range for thin-walled parts and higher-frequency equipment.
• Röhm gives a height range of 0.30 to 0.80 mm and advises placing the director symmetrically in the middle of the joining area.
• Branson documentation illustrates a director of roughly 0.015 in (about 0.38 mm) as a typical starting geometry.

The principle is volume. Size the director so that when it melts, its volume fills as much of the joint cross-section as possible without squeezing out as flash. Too small, it heats slowly and leaves the joint partly unfused. Too big, it collapses violently, pushes resin out as flash and steals collapse travel from the seam.

One rule matters when you join two different materials: when joining plastics with different stiffness, put the energy director on the less stiff plastic. The softer, lower-modulus part deforms first and collapses around the stiffer part, which acts as a stable anvil. Put it on the rigid part and you are melting a ridge into something with nowhere to go.

Directors suit amorphous polymers such as ABS, PC and PS, which soften gradually and forgive small errors. Semi-crystalline resins like nylon, PP and POM melt more abruptly, so they often need a larger director, a double director, or a melt trap, a small reservoir that catches squeeze-out instead of letting it flash onto the cosmetic surface.

The joint geometry catalog

Eight joint families cover nearly every molded plastic assembly. Choose by asking: does it need to hold peel load, does it need to be hermetic, and can the mold make it?

Joint	How it works	Strengths	Watch-outs
Butt	Flat face to flat face	Simple to mold, easy to inspect	Low peel strength, visible flash
Step (rebate)	Lip seats into a step	Self-aligning, hides flash	Needs wall support around the step
Tongue-and-groove	Tongue into a matching groove	Highest strength, weld hidden, needs no alignment fixtures	Tight tolerances, hard to mold
Shear	Angled land to vertical wall	Best for hermetic seals, favored for semi-crystalline plastics	Careful geometry, stable nest
Mash	Interlocking jagged faces	Fast, tolerates misalignment	Not a sealing joint, moderate strength
Double-V	A V-groove on both parts	Symmetrical melting, good on thin walls	Complex tooling, flash on both sides
Spot	Small localized welds	No visible joint line	No energy director needed, not structural
Staking & swaging	Post deforms over a metal insert	Joins plastic to metal or other materials	No seal, insert geometry critical

For staking and swaging, the insert is a metal part in a plastics process, so the metal side needs the same care. If that insert is aluminium, clean it first, the way our guide to aluminum cleaner for welding describes. Shops joining the bracket itself to structure are covered in the aluminum MIG welding guide.

Design rules that prevent most failures

• Round every edge and corner in the join area. Sharp internal corners act as stress concentrators and craze first. A generous radius costs nothing in strength.
• Keep the join area perpendicular to the horn axis, in a single plane and parallel to the horn face. A tilted or non-planar joint makes horn contact uneven, so half the surface never sees the force.
• Keep the horn close to the joint. Every millimeter of part between the horn and the seam absorbs energy that never reaches the melt zone.
• Provide adequate horn contact area and part stiffness. Thin, unsupported walls deform under the horn before the joint welds.
• Size the joint so collapse stays inside the intended collapse height. Standardized test specimens are defined around this dimension in AWS G1.2M, the specification for standardized ultrasonic welding test samples for thermoplastics.
• Add flow relief or a melt trap for semi-crystalline resins so squeeze-out has somewhere to go that is not the cosmetic surface.

The Machine and Its Parameters: Amplitude, Force, Time, Energy, Distance

Every ultrasonic weld you have ever heard came down to a handful of numbers. The machine has no opinion about how strong the joint should be. It repeats, faithfully, whatever the operator told it to repeat. That means a weak joint is usually traceable to a setting, a tooling problem, or a part design, and all three can be found.

The six parameters and what each one physically does

Amplitude is the headline setting. It is the peak-to-peak movement of the horn face, measured in micrometers, and the most influential parameter in the process because it determines how fast the energy director reaches melt. Amplitude is not a number you type into the control box. It is the product of gains in the converter, the booster and the horn, and on modern equipment it is trimmed electronically, with no spanner on the stack. Herrmann’s overview of the ultrasonic welding process describes how amplitude is set at the stack level and then fine-tuned at the generator.

Weld force squeezes the energy director against the horn face, and that pressure converts vibration into friction heat. Branson puts the typical starting point at 30 to 50 psi: lower for small, thin-walled parts, higher for larger parts with thicker walls. Force does not make the joint strong on its own; it maintains contact so the vibration can do its work.

Downspeed is how fast the horn descends onto the part, and Branson notes that most setups run at 1.5 to 2.5 in/sec. It matters more than it sounds, because downspeed sets the rate at which force builds during the weld. Drop too slowly and the force ramps gently. Drop fast and the energy director is slammed into its mating surface before it can melt evenly.

Weld time is how long the horn stays energized once force is reached. Weld energy, in joules, is the total energy delivered over that time. Energy is an output rather than an input, which is why two parts that look identical can need different joule counts: joint design, wall thickness and material all change the demand. Hold time and hold pressure follow power-off, keeping the parts pressed together while the molten interface solidifies. Skipping the hold is a classic cause of parts that pop apart on the cooling rack.

Parameter	What it physically does	Typical starting point
Amplitude	Sets the speed of melt generation	Set by stack gains; trimmed electronically
Weld force / pressure	Maintains contact so vibration becomes heat	30–50 psi
Downspeed	Sets how quickly force builds	1.5–2.5 in/sec
Weld time	How long the horn is energized	0.15–0.5 sec for most joints
Weld energy	Total energy delivered, in joules	Monitor it, do not chase it blindly
Hold time / hold pressure	Compresses the joint while it solidifies	Matched to resin cooling rate

Control modes: how the machine decides to stop

Mode	How the weld ends	Where it earns its keep
Time mode	Power stops after a set weld time	The honest starting point
Energy mode	Weld time varies to hit a joule target	Multi-cavity tools with varying demand
Collapse mode	A linear encoder stops the weld at a set collapse distance	A 0.015 in director collapses exactly 0.015 in
Absolute mode	Fixed actuator travel holds overall height	Final part height is a specification
Peak power mode	Weld stops at a power spike	Textile plunge welding; prevents blow-through

Pre-trigger and the encoder that makes it repeatable

Pre-trigger means the horn moves down a set distance before the generator fires. It is essential in shear joint, staking and insertion work, where the horn would otherwise crush or deform the part before contact. Fictiv’s guide to ultrasonic welding for assembly describes how pre-trigger combined with a linear encoder of roughly 0.0001 in resolution is what makes shear joint and staking repeatable rather than a matter of luck.

Every cycle follows the same sequence: approach (horn descends with power off), trigger (force reaches the trigger point and the generator energizes), weld (vibration and pressure do their work), hold (power off, pressure maintained while the joint solidifies). The trigger point can come from a mechanical switch or, better, from a trigger force / load cell that fires at a set measured force. Because it reacts to actual force rather than a hard mechanical stop, it tolerates the surface and height variations that come off a molded part and still lands the weld consistently.

The physics caveat you cannot tune your way around

Amplitude and force are not simply “more is better” dials. Work documented by Roehm shows that amplitudes that are too high cause molecular alignment, heavy flash and reduced weld strength, while amplitudes that are too low produce nonuniform melt initiation and premature solidification, which shows up as parts sticking in the tool. Pressure behaves the same way: too little gives poor energy transmission into the joint, too much causes molecular alignment and again lowers strength. Research on determining minimum amplitude for ultrasonic welding reaches a similar conclusion, that usable amplitude has a floor and a ceiling rather than one ideal value.

The practical lesson carries into any process you tune. Watching heat input, controlling force and verifying the result is not unique to plastics. If you have tuned a metal process, the habits transfer directly. Set your parameters, then verify with a peel or pull test rather than assuming the machine agreed with you.

From First Weld to Validated Production: A Practical Workflow

Most ultrasonic welding problems are solved long before the machine is switched on. The weld takes a fraction of a second, but pass or fail is decided over days: in the drawing, in the tooling, and in the parameter data you collect while scrapped parts are still cheap. This sequence holds up on a real production floor.

Step 1: Define the requirements before you touch a machine

Write it down first: material grade and supplier, joint geometry, wall-thickness consistency along the joint line, whether thick sections create voids, whether the seal must be hermetic, the structural load the joint carries in service, whether weld marks are acceptable on a visible face, and how many weld sites the part has. A housing with a lid, a port and a sensor may need three tools and three weld operations.

Be precise about which requirement you are buying. A hermetic seal, a structural joint and a cosmetic exterior are three different jobs, and a part needing all three is often better split into two assemblies than forced into a single weld.

Step 2: Build the tool

The nest locates and supports the lower part, holds alignment rigid so the joint cannot shift under load, and gives the horn a flat, parallel surface to press against. Leave room for loading and unloading too: a fixture that is awkward to load will eventually be loaded badly.

The horn is sized to the joint area, not the part. Too large and you get marking and flash; too small and you lose energy into the components instead of the joint. Assemble and torque the stack to specification so stack amplitude matches the material you run. An under-torqued stack is a common reason a machine seems unable to hold settings from one run to the next.

Size the joint early and build it up incrementally. Madison Group points out that shrinking an oversized weld joint later means welding steel back into the mold and machining it down again, one of the costliest mistakes in the process.

Step 3: Start in time mode and let the parts teach you

Fancier control modes feel more advanced, which is where many people go wrong. Branson recommends starting in time mode, refining the process manually, and only then running a design of experiments on amplitude, pressure and downspeed before selecting a control mode.

Time mode is the honest starting point because it tells you what the parts are actually doing. An energy or collapse mode will hold a stable number on a joint that is melting unevenly, masking the problem rather than fixing it. Set the trigger, weld parts by hand, inspect them, then change one variable at a time and write down what changed.

Step 4: Run a design of experiments

Guessing gets you close; a DOE gets you to the optimum. Vary amplitude, force and downspeed across a small matrix. Branson suggests a minimum batch of 50 welded and labelled parts, with accuracy improving as more cavity combinations are included, so build the matrix around what the tool actually produces.

For every part, capture energy, peak power, collapse and absolute values from the machine, then test them:

Validation method	What it tells you
Peel test on a fixture	Shear strength along the joint line; most common pass/fail criterion
Pull or tensile test	Whether the joint or the surrounding wall fails first
Burst or leak test	Hermetic integrity for fluid-handling parts
Destructive sectioning	Internal voids, incomplete collapse, flash intrusion
Visual cosmetic check	Flash, sink marks, horn marking, visible flow lines

Then compare the good parts against the bad ones. A failing part usually stands out in one data channel, telling you whether you have a power, collapse or force problem.

Step 5: Choose the control mode, then profile amplitude

Once the population is stable, pick the control mode with the least scatter, not the one that produced the single best weld. Energy mode often proves more consistent on multi-cavity tools, where each cavity sees slightly different conditions. Then run a second DOE on amplitude profiling: high amplitude early to reach melt quickly, reduced amplitude mid-cycle for consistency and flash control.

Step 6: Move to production with monitoring and records

Monitor energy or collapse on every part and set limits and alarms so a drifting process stops the line instead of filling a bin with rejects. Record machine settings, tool IDs, part revisions, material lot numbers and drying status for hygroscopic materials in the traveler or work instruction, not in someone’s notebook.

Herrmann highlights another benefit of precise control: reproducible results with no tool warm-up or cool-down wait. On a short-cycle line, a 30-second wait per tool change is lost capacity you never get back.

Troubleshooting: Weak Welds, Flash, Warping, Bubbles and Noisy Failures

Troubleshooting is where plastic sonic welding proves itself as a process you can actually learn. Once you can describe what a joint is doing, you can usually trace the cause to a parameter, a tool or the part design. Read the evidence in order, from the part in your hand to the machine on the bench.

1. The Weld Is Weak or Separates at the Seam

The most common complaint, and usually a design issue rather than a machine issue. A joint that pulls apart cleanly at the seam almost always means the energy director never collapsed. That happens when amplitude is too low, when the horn presses on a large flat area so the energy spreads out, or when the part has no director at all. Herrmann describes the ideal mechanism: the director melts under vibration and, under additional pressure, creates a tight, visually flawless and strong bond. No collapse, no bond.

Other culprits: parts misaligned when the horn lands, weld force too low to push the molten director into the joint, and semi-crystalline resins given too little energy to reach a uniform melt. The Roehm work on base materials makes the pressure balance explicit: too little weld pressure gives poor energy transmission into the joint, while too much pressure aligns the molecules and reduces weld strength. The window is narrower on semi-crystalline grades.

2. Flash and Excessive Squeeze-Out

Flash is melt escaping the joint line, the visible face of too much energy or too much force: amplitude too high, an aggressive energy target, a director with more volume than the joint can swallow, or a semi-crystalline resin that needs a melt trap to hold the squeeze-out. Insufficient hold time is the other half of the equation, because the joint must be pinned while it solidifies. Flash on one side only usually means a tilted horn or parts that are not seated flat.

3. Warping and Sink Marks

Warping happens when heat soaks beyond the joint into a thick section, and sink marks appear when the horn presses on an area with no stiffness behind it. Uneven wall thickness worsens heat sinking, and hold pressure applied off-centre twists the part. Fix the geometry before the settings: add a rib or boss under the horn contact, keep wall thickness consistent and centre the hold pressure.

4. Bubbles, Sizzle and Porous Joints

A sizzle during welding or a porous line in a cut joint points at moisture. Madison Group traced porosity along the weld interface to moisture in conditioned polyamide parts undermining the intended design, matching the wider material data on hygroscopic nylons. Wet tooling and condensation on a humid day do the same thing. Dry polyamide parts as molded and keep the nest dry.

5. Strength Varies From Cavity to Cavity

With a multi-cavity tool, one cavity can weld beautifully while the next runs weak. The causes are energy variation across cavities and a trigger that fires inconsistently. Moving to energy or collapse control smooths this out. Branson also notes that a pre-trigger turns ultrasonics on before the horn contacts the part, and that on machines without distance capability, starting ultrasonics on leaving the home position can create excessive noise, horn failures and inconsistent welds. Weld and label each cavity separately, then compare energy traces.

6. Machine-Level Symptoms

Dukane lists increased wattage draw and a change in the sound of the weld as signs of a problem with the welder or the horn. In practice that means inspecting for a worn or scored horn face, a cracked booster, loose stack bolts or contamination on the horn tip. Retorque the stack, clean the face and check measured amplitude at the horn before touching the weld schedule.

7. Horn Marking or Surface Damage

A witness mark on the part is not always a defect, but dents and raised edges are. They come from a horn too large for the joint, a misaligned horn, a part without enough stiffness under the contact face, or a frequency wrong for the joint size. Shrink the horn, realign it, add reinforcement and check that the frequency suits the joint mass.

Fast Diagnostic Table

Symptom	Likely physical cause	First adjustment	Confirmatory test
Seam separates by hand	Energy director not collapsing	Raise amplitude, shrink horn contact area	Peel test, section the seam
Joint intact but weak	Low force, under-energized semi-crystalline resin	Raise force and hold time	Peel test, check for voids
Flash at the seam	Amplitude or energy too high, oversized director, no melt trap	Reduce energy, add melt trap	Inspect flash edge and collapse
Warp or sink marks	Heat sinking into a thick, unreinforced area	Add ribs, lower hold pressure	Flatness check against an unwelded part
Bubbles or sizzle	Moisture in polyamide, wet tooling, condensation	Dry parts as molded, dry the nest	Cut the joint, look for voids
Strength varies by cavity	Energy variation, inconsistent trigger	Switch to energy or collapse mode	Test each cavity separately
Wattage up, weld tone changed	Worn horn, cracked booster, loose stack bolts	Inspect and retorque the stack	Re-measure amplitude, re-weld and listen
Marked or dented part	Oversized or misaligned horn, part not stiff enough	Resize and align the horn	Compare mark to horn face

For those who work with metal rather than polymers, the parallel is worth reading in Aluminum MIG Welding: The Complete Practical Guide: in both cases the joint geometry, the heat input and the tooling decide the outcome, and the controls only fine-tune what the design allows.

Plastic Sonic Welding Compared With Other Joining Methods

Ultrasonic welding is fast, clean, and right for many plastic assemblies. It is not right for all of them. An engineer who reaches for it first, without asking how the parts come apart, what the joint must seal, or who pays for the tooling, will spend more time fixing joints than saving them.

The Head-to-Head Comparison

Joining Method	Cycle Time	Consumables	Hermetic Seal	Strength	Cosmetic Result	Dissimilar Materials	Rework / Disassembly
Ultrasonic (plastic sonic welding)	0.5–2 s	None	Excellent	High in shear, moderate in peel	Clean seam; slight witness ring	Limited (same or similar family)	Difficult; not designed for disassembly
Adhesive bonding	Seconds to apply; hours to cure	Adhesive, applicator, primer	Good if designed for it	Moderate; joint dependent	Often invisible	Excellent	Usually not reversible
Solvent cementing	Minutes plus set time	Solvent cement	Good on compatible plastics	Moderate	Visible seam, usually clean	Poor; same resin family only	Not reversible
Mechanical fasteners	Seconds per fastener	Screws, nuts, bosses	Not by default; needs a seal	Very high	Holes and fastener heads	Excellent	Fully removable
Hot plate welding	Seconds to minutes plus soak	None	Good	Good	Visible seam, some flash	Limited	Not reversible
Hot-air / hot-blade repair	10 s to minutes, manual	None, sometimes filler rod	Not reliable	Moderate to poor	Melted, uneven surface	Poor	Not reversible
Vibration welding	Several seconds	None	Good	High	Good seam	Limited	Not reversible
Spin welding	1–3 s	None	Excellent	High in torsion	Clean on round parts	Limited	Not reversible
Laser welding	1–5 s	None	Good	High	Very clean, no contact	Moderate with coatings	Not reversible

Where Ultrasonic Pulls Ahead

The headline advantage is speed and cost per part at volume. Fictiv notes that ultrasonic welds complete in roughly 0.5 to 2 seconds with no consumables, where an adhesive joint may need minutes to hours to cure. That is the gap between a cell making 60 parts a minute and one making four, and it is why ultrasonic dominates in medical devices, electronics housings, and automotive interiors.

The seal is the second headline. Ultrasonic welds are genuinely hermetic, which is why Fictiv highlights their use for airtight and watertight seals in filtration housings, diagnostic test kits, and waterproof electronics. The joint is also clean: no adhesive squeeze-out, no solvent fumes, no fastener holes, no print left on the part.

Where Ultrasonic Falls Short

Herrmann identifies the two structural limits clearly: the size of the joint interface you can weld, and the requirement that the joint be specially designed for the process. You cannot put a horn on a flat, unsupported plate and expect a good weld. Every ultrasonic joint needs a designed energy director or equivalent geometry.

Capital cost is the other honest weakness. An ultrasonic welder plus custom horn and nest tooling for one joint costs real money, while a hot-air repair gun costs a small fraction of that. For a single cracked enclosure, the hand tool wins on economics even though it wins on nothing else.

The Alternatives in Detail

• Adhesive bonding. Joins dissimilar materials and leaves an invisible joint, but adds cure time, consumables, mess, and possible outgassing in a sealed product.
• Solvent cementing. Cheap and clean, but the vapor is a hazard, only compatible plastics can be joined, and stress can cause cracking later.
• Mechanical fasteners. Removable and strong, but they add weight and part count, and need drilling, tapping, or insert molding as extra steps.
• Hot plate welding. Simple and well suited to large flat seams, but heat soak lengthens cycles and can deform parts.
• Hot-air and hot-blade tools. The default for a single repair; they will not deliver repeatable production joints of consistent strength.
• Vibration welding. Operates at 100 to 300 Hz, far below ultrasonic frequencies, which suits larger parts but with longer cycles.
• Spin welding. Ideal for round containers such as bottles and canisters, with leak-tight seams, but limited to rotational joints.
• Laser welding. Clean and contactless, but capital cost is high and reflectivity or absorption can be troublesome.

Metal-to-Plastic: Staking and Swaging

When a plastic part must lock onto a metal insert, ultrasonic staking and swaging are the answer. The plastic carries a molded pin, and the horn deforms that pin over the metal feature, trapping it permanently. No consumables, a fast cycle, and a joint strong in shear and pull. The trade-off is permanence: the assembly cannot be disassembled without destroying the plastic.

Making the Call

Choose ultrasonic when the joint is small to medium, the geometry is designed for it, the seal matters, and you are running volume. Choose adhesives or fasteners when you need dissimilar materials or field serviceability. Choose spin or vibration for large round or bulky parts, and a hot-air tool when you are fixing one part rather than building a line.

For context on how the same energy-based thinking plays out in metalwork, see our guide to aluminum flux core welding and the practical overview of aluminum MIG welding.

Where Plastic Sonic Welding Is Used, and What It Costs

Ultrasonic welding is one of those processes you rarely see but almost never notice. It holds together the two halves of a blender jug, the cap on a washer reservoir, the shell of a diagnostic test kit, and many housings inside cars. The flash is minimal and the seam often sits in a shadow line, so the process is hard to judge from a finished part. The economics show up on the line, in cycle time, and in what it costs to run a part that never leaves the station.

Sectors Where the Process Earns Its Keep

Herrmann, one of the larger ultrasonic assembly specialists, groups its applications into automotive, consumer, electronics and medical, and reports running more than 1,500 customized projects per year in its automotive branch alone. Every one of those projects is a different joint, which tells you two things: demand is enormous, and every application needs its own purpose-built tooling.

• Medical devices: syringes, catheters, filtration housings, diagnostic test kits, reservoir and collection chambers.
• Consumer electronics: mobile phone housings, wearable housings, remote controls, sensor pods.
• Automotive interior and fluid handling: fluid reservoirs, filters, sensors, trim and vent assemblies.
• Appliances and water treatment: water filter housings, dispensers, pump chambers.
• Everyday goods: toys, printer and toner cartridges, packaging blisters and clamshells.

Why Medical and Electronics Buyers Care More

In medical work, the argument is contamination. Roehm notes that ultrasonic welding prevents the introduction of contaminants or degradation sources at the weld, which supports biocompatibility and makes the process cleanroom compatible. There is no solvent to evaporate near a critical component and no adhesive outgassing. The joint is simply the same polymer, melted and fused.

Electronics makers care about a second property: no threaded holes, no metal in a housing, and a hermetic seal that keeps dust and moisture out of an enclosure containing a battery. Where no tooling contact is wanted at all, laser welding is the alternative, though it carries far higher capital cost and its own material constraints.

A Worked Production Example

Take a two-cavity ABS enclosure for a small appliance accessory. The joint is a step joint with a molded energy director, welded at 30 to 40 kHz. The weld takes under a second, but the cell adds handling, so the full cycle lands near three seconds. At three seconds per cycle, one station on an eight-hour shift produces roughly 9,600 assemblies, or about 2.4 million parts a year at 250 working days.

An in-process collapse monitor watches the joint geometry on every part and flags a drift before the part moves downstream. That moves scrap detection from the end of the line to the cheapest possible point on it.

What Actually Drives the Cost

Cost element	Type	What happens as volume rises
Welder and power supply	One-time capital	Amortizes fast; a fraction of a cent per part at volume
Horn, nest and fixture tooling	One-time, joint-specific	Fixed regardless of volume, so per-part cost falls sharply
Trials, DOE and validation	One-time	Largest early cost, and the one most underestimated
Fasteners, adhesives, solvents	Per part	Zero. Nothing to buy, stock, cure or expire
Cycle time and labour	Per part	Dominates at low volume, less significant at high volume
Scrap and rework	Per part	Falls with in-process energy or collapse monitoring
Downstream failure cost	Per failure	The number that should drive process investment

The Costs Engineers Underestimate

Budgets usually allocate money to the machine and the tooling. They rarely allocate enough to the time between them. Getting from a first weld to a validated production process means running trials, measuring results, adjusting the joint, rebuilding the horn and repeating. Each iteration is a small engineering project, and the schedule slip that follows is usually the real cost.

None of this is unique to plastics welding, and none of it is hidden by designing the joint with the joining process in mind. A design that anticipates the weld comes out of the tooling phase faster and cheaper than one that does not.

So the honest answer on cost: ultrasonic welding is expensive to start and cheap to run in volume. For a handful of repair joints, a hot-air tool beats it. For a hermetic medical housing running hundreds of thousands of units a year, an ultrasonic cell with custom tooling is one of the cheapest joining methods available, because nothing is consumed and the joint takes less time to make than it does to load the part.

Safety, Standards and the Realistic DIY Picture

Call it ultrasonic welding or plastic sonic welding, the gap between a job that ships and one that sits in a drawer comes down to the sound of the joint and the paperwork behind it.

Noise is the dominant operator exposure

Ultrasonic welders are loud, and hearing conservation is routine on any line running one. OSHA requires a hearing conservation program in general industry at or above 85 dBA as an eight-hour TWA, and sets the permissible exposure limit at 90 dBA over eight hours with a 5 dBA exchange rate. NIOSH recommends keeping worker noise below 85 dBA over eight hours with a 3 dBA exchange rate. CDC estimates 22 million US workers are exposed to potentially damaging noise at work each year.

Reference	Level	Exchange rate	Practical effect
OSHA permissible exposure limit	90 dBA, 8-hour TWA	5 dBA	Legal ceiling
OSHA hearing conservation trigger	85 dBA, 8-hour TWA	5 dBA	Program required
NIOSH recommended limit	85 dBA, 8-hour TWA	3 dBA	More protective

Use the hierarchy of controls. Engineering controls first: an acoustic enclosure, isolation from the shop, and distance from the horn, since a 40 kHz horn is far less of a problem three metres away than at arm’s length. Administrative controls, like time limits near a running machine, come next. Hearing protection is the last layer, not the only one.

Guarding and the rest of the shop safety list

Ultrasonic welders have few exposed moving parts, but there is a crush zone between horn and nest, and the tooling can hurt: a cracked horn can fragment, and a loose stack bolt changes amplitude and sound. OSHA’s machine guarding standard, 29 CFR 1910.212, requires guarding where employees are exposed to injury, and requires fixed machinery to be anchored. In practice: a guard or interlock over the horn stroke, a bolted-down base, and lockout before changing a horn, fixture or nest. General shop habits, as in this welding equipment and safety guide, apply here.

What the fumes actually are

In normal use there is no smoke, because the joint melts locally and stays in the weld line. An over-driven joint cooks the resin instead of fusing it. Herrmann’s plastics guidance is blunt about the worst case: PVC burns very quickly and forms hydrochloric acid, so temperature and heating duration need close attention. Avoid smoke from any overheated polymer with local exhaust, a safety data sheet review for each resin, and no welding of unknown or contaminated parts.

Standards worth pointing at

AWS publishes two documents for this process: G1.1M, Guide for the Ultrasonic Assembly of Thermoplastics, and G1.2M, Specification for Standardized Ultrasonic Welding Test Samples for Thermoplastics. G1.2M defines standard sample geometries, so you can compare weld strength between labs or machines. ISO 17663 covers ultrasonic welding of plastic parts and assemblies.

One warning: D1.1 is the structural welding code for steel, not a plastics code. It governs metal arc processes such as stick welding and says nothing about thermoplastics.

The DIY reality check

Most handheld plastic repair tools sold to consumers are hot-air or hot-blade welders, not ultrasonic systems, and that is not a criticism. For a cracked ABS or PC enclosure, a hot-air gun with a proper tip and a holding fixture will do a decent job for a fraction of an ultrasonic unit’s price.

Genuine ultrasonic repair needs an energy director or a prepared surface, support under the joint, and a tool that reaches it with the horn face flat and parallel. No frequency and amplitude can rescue a flat, poorly supported joint. And the material must be weldable: thermoplastics fuse, elastomers spring back, thermosets disintegrate under heat, exactly as Herrmann describes.

A small shop can repair a cracked housing with hot air and use a small ultrasonic spot welder for locating features, clips or plastic bosses. Hermetic fluid-handling joints, thin-walled filters, tight leak requirements and production volumes still need industrial equipment.

Buy a bench welder, or send the work out?

A bench ultrasonic welder makes sense with a repeatable part, a designed joint and steady volume, because the tooling is specific to that joint. For one-off jobs or prototypes, sending work to a plastics welding service is usually cheaper and faster than specifying a machine, building a horn and nest, and tuning them.

Treat joint design, material choice and the noise plan as part of the job, not an afterthought. Those three decisions separate a repeatable part from a story about the one that failed.

Frequently Asked Questions

Is plastic sonic welding the same thing as ultrasonic welding?

Yes. ‘Sonic welding’ is a trade-floor nickname for ultrasonic welding; there is no separate industrial process called sonic welding for plastics. Ultrasonic plastic welding runs at roughly 20–70 kHz with micrometre-scale vibration, while vibration welding runs at a much lower 100–300 Hz. Searching for ‘ultrasonic plastic welding’ alongside the primary term will bring up far deeper technical documentation.

Which plastics can be welded with ultrasonic energy?

Thermoplastics can be welded, with PTFE as the notable exception; elastomers and thermosets cannot, because they either spring back or disintegrate when heated. ABS, polystyrene and polycarbonate are relatively forgiving, while polypropylene, PVC and nylon are more demanding because of sharp melting behaviour and moisture sensitivity. Elastomers and thermosets are simply not candidates for this joining method.

How long does one plastic sonic weld take?

Most production welds complete in well under a second — typically 0.5 to 2 seconds for the whole cycle, and Branson reports most welds occurring in less than 500 milliseconds. There is no warm-up or cool-down wait for the tooling, which is a major advantage over thermal joining methods.

Why is my ultrasonic weld weak or pulling apart?

The most common causes are an energy director that is not collapsing (amplitude too low or joint area too large), misalignment in the nest, insufficient force or energy, and moisture in hygroscopic polyamides creating voids along the interface. A study on PMMA and ABS also found that weld times past the optimum point stopped improving strength. If parameter changes do not fix the defect, the joint geometry itself is usually the problem.

Can you weld two different plastics together?

It is possible, but it is one of the harder cases because the two resins must share a workable melting window. As a rule, the less stiff plastic should carry the energy director so it melts first and starts the joint. Mismatched melting ranges, differing shrinkage and differing damping behaviour all make the process window narrow, so trials are essential.

What causes flash or squeeze-out around an ultrasonic weld?

Flash usually means too much molten material for the joint to contain: excessive amplitude, energy or force, an oversized energy director, or a semi-crystalline resin that needs a melt trap to capture excess melt. High amplitude also causes molecular alignment, which lowers weld strength while producing heavy flash, so the fix is reducing energy input rather than adding more.

Can I weld plastic at home with a handheld tool?

Most consumer plastic repair tools are hot-air or hot-blade welders rather than true ultrasonic systems, and they will not reproduce a repeatable production joint. A genuine ultrasonic repair needs a designed energy director or a prepared joint feature, plus controlled amplitude, force and dwell, so the joint still has to be designed rather than simply melted. For one-off repairs, hot-air or hot-blade methods remain the practical choice.

Is ultrasonic plastic welding safe for operators?

The main exposure is noise. OSHA requires a hearing conservation program at 85 dBA eight-hour time-weighted average in general industry, with a 90 dBA permissible limit and a 5 dBA exchange rate, while NIOSH recommends controlling exposure below 85 dBA. Engineering controls such as enclosure and isolation come first, followed by administrative controls and hearing protection, and point-of-operation guarding is required under 29 CFR 1910.212.

Which standards cover ultrasonic welding of thermoplastics?

The American Welding Society publishes G1.1M, a guide for the ultrasonic assembly of thermoplastics, and G1.2M, a specification for standardized ultrasonic welding test samples for thermoplastics. European and national bodies also publish plastics welding and testing standards. AWS D1.1 covers structural welding of metals and does not apply to plastics ultrasonic welding, so specify the plastics-oriented documents instead.

How much does ultrasonic plastic welding cost compared with glue or screws?

The welder and its custom horn and nest are a real capital investment, so the comparison depends on volume. Per part, ultrasonic welding is cheap because there are no fasteners, no adhesive and no solvents to buy, and no cure or set-up time; at low volume that capital cost dominates. Process development and validation trials are usually the largest early expense and are the part engineers most often underestimate.