Thompson Friction Welding: Process, Machines & Uses

A haulage axle maker once had to choose between a slow, filler-hungry arc weld that warped every tube and a process that joined steel to steel in about twenty seconds with no wire, no gas and no grinding. That second option is direct-drive rotary friction welding, the technology Thompson Friction Welding has been building in England’s Black Country since 1963.

What Thompson Friction Welding Actually Is

A haulage axle maker once faced a familiar fork in the road. On one side sat a slow, filler-hungry arc weld that warped every tube it touched. On the other sat a process that joined steel to steel in roughly twenty seconds, with no wire, no gas and no grinding. That second option is direct-drive rotary friction welding, the technology Thompson Friction Welding has been building in England’s Black Country since 1963.

But what does “Thompson friction welding” actually mean? Is it a brand, a machine, or a process? The short answer is all three, and separating them is the fastest way to understand the subject.

Thompson: A Brand Traceable to 1834

Thompson is a British machine-building name with roots going back to 1834, when William Thompson established a boiler works in the industrial Black Country near Birmingham. By the mid-twentieth century the wider Thompson group employed some 6,000 people and built heavy vehicle components, including drive and trailer axles for heavy goods vehicles. In 1963 the company set up a dedicated friction welding business precisely because conventional arc welding could not make those axles reliably. That business is the origin of what people now call Thompson Friction Welding. Today the machines are manufactured by KUKA in Halesowen, in the Black Country, with a supporting office network in Michigan, USA, serving North American manufacturers (Thompson Friction Welding, About Us).

The Process: Direct-Drive Rotary Friction Welding

Friction welding itself is a family of solid-state joining processes. Thompson’s variant is direct-drive rotary friction welding (RFW): two components, at least one of them round in cross-section, are clamped in axial alignment. One is spun under motor drive while both are pressed together. The rubbing at the contact face generates heat, the metal softens and plastically deforms, and then rotation is actively braked to a stop while a heavy forge force is applied. The two parts fuse into a single solid piece without ever reaching a molten state.

The key phrase there is solid-state. Under normal conditions the faying surfaces do not melt. Because nothing liquefies, the process needs no filler wire, no flux and no shielding gas (AWS C6.1, Recommended Practices for Friction Welding). That is a world away from fusion processes such as MIG welding, where a consumable electrode melts into a molten weld pool that later solidifies, with all the shrinkage, porosity and distortion risks that entails. If you have ever fought a warp-prone aluminium job, you will appreciate the contrast with techniques covered in our aluminium MIG welding guide.

Why the Distinction Matters

Because no melt pool forms, the joint does not suffer from solidification cracking, and the heat input stays comparatively low. That means less distortion, fewer defects, and the ability to weld combinations that fusion welding handles badly, including dissimilar metals. It also explains why the process appeals to mission-critical sectors such as oil and gas, rail and power generation, where a failed joint is not an inconvenience but a hazard.

What This Article Covers

If your search is informational — what Thompson friction welding is, how the cycle works, and where it is used — you are in the right place. Over the next sections we will cover the heritage story behind the company, a step-by-step walkthrough of the direct-drive cycle, the Thompson machine range from 15-tonne to 400-tonne capacity, real applications from truck axles to API drill pipe, the benefits and honest limitations, the standards and inspection framework, and how Thompson’s approach compares with inertia friction welding, linear friction welding and arc welding. We will close with practical scoping advice if you are considering friction welding for your own production line.

A Short History of Thompson and the Black Country

The Black Country got its name from the smoke, soot and coal dust that hung over the West Midlands through the industrial century, and for generations its defining trade was turning iron and steel into objects that had to hold pressure without leaking. It is no accident that the firm now known for friction welders came out of that trade. Thompson Friction Welding is, in a real sense, a boiler maker’s descendant, still working out of Halesowen, a few miles from where the family business began.

1834: William Thompson Opens a Boiler Works

William Thompson set up his boiler works in 1834, just as Britain’s railways, collieries and shipyards were creating almost insatiable demand for pressure vessels. The company grew with the region’s heavy engineering base, and its reputation rested on weld integrity rather than sheer output. Two milestones from that era still define the name: in 1930 Thompson pioneered the union-melt submerged arc welding process, and it became the only company approved by Lloyd’s to produce forge-welded pressure vessels. By 1947 the organisation employed around 6,000 people, a number that shows how central it was to the local economy and how much welding skill was concentrated in one place.

Boiler fabrication never became obsolete; it simply narrowed into a specialist field with its own codes and shop practices, a trade we touch on in our look at Alamo Welding & Boiler Works.

1963: When a Process Became a Product

Thompson Friction Welding was established in 1963 to build machines and parts around direct-drive rotary friction welding, and the timing mattered, because the science was barely a decade old in the West. The first published rotary friction welding experiments came from A. I. Chudikov in the Soviet Union in 1956, the process reached the United States in 1960, and TWI began the first UK studies in 1961. Within a few years the technique had moved from laboratory curiosity to shop-floor production: in 1966 KUKA and Thompson brought friction welding into European manufacturing, according to the company’s own company history.

1969, 1993 and 2023: New Owners, Same Craft

In 1969 Thompson became part of the NEI Group (Northern Engineering Industries), a move that put the friction welding business inside one of Britain’s largest engineering conglomerates. The next change came in 1993, when the KUKA Group acquired the business. Rather than absorb the name, KUKA kept it: the machines are still marketed as Thompson, with KUKA’s automation and robot systems folded into the same packages. That continuity paid off in 2023, when the brand celebrated 60 years of friction welding innovation, six decades in which the fundamental process changed far less than the equipment around it.

The Wider Global Story

Thompson’s story sits inside a longer timeline worth knowing, because it explains why rotary friction welding is now treated as mature, well-documented technology rather than an exotic specialty. Chudikov’s 1956 publications were followed by US adoption in 1960, European production through KUKA and Thompson in the mid-1960s, and decades of standardisation work. TWI later went on to invent friction stir welding in 1991, a solid-state cousin that replaced rotating pins with a non-consumable tool, and that same institute remains a reference point for process data summarised in overviews such as the friction welding process summary. Aerospace programmes were among the earliest adopters of the newer variants, a point we return to in our feature on aerospace welding processes and standards.

Why the Heritage Still Shows Up in the Product

Heritage matters here for practical reasons. A century and a half of pressure-vessel work is why the company treats metallurgy, not just machine building, as core competence; the Halesowen site still runs a metallurgical laboratory alongside its subcontract welding service, where customer components are welded, sectioned and examined before a production line is ever tooled. The Black Country also supplies something harder to buy: an established base of toolmakers, machinists and weld technicians who understand axial alignment, burn-off control and flash removal. When a haulage axle maker or a drill pipe producer specifies a joint, they are buying into that accumulated judgement, not simply a machine with a tonnage label attached.

Which brings us to the machine itself, and to the sequence of clamping, spinning, pressing and braking that turns two separate forgings into a single continuous grain structure in a matter of seconds.

How the Direct-Drive Rotary Friction Welding Cycle Works

Friction welding sounds exotic, but the cycle itself is refreshingly simple to describe. One part is held still, the other is spun, both are pushed together, and the heat generated by rubbing turns a mechanical movement into a finished joint. No arc, no filler wire, no shielding gas, and no external heat source at all. As Thompson Friction Welding puts it, the process uses no electrical energy: mechanical energy is converted into heat purely at the joint interface. The whole event is over in roughly the time it takes to read this paragraph aloud.

The sequence below follows the standard direct-drive rotary friction welding (RFW) cycle as described by both Thompson Friction Welding’s own process guide and AWS C6.1, Recommended Practices for Friction Welding.

The Cycle, Step by Step

1. Clamping and axial alignment. Both components are chucked into the machine: one in a rotating headstock, the other in a fixed tailstock or hydraulic ram. Everything in friction welding is axial, so alignment matters more than in most joining processes. Even a small angular misalignment will show up as uneven flash and a wandering interface.

2. Rotation begins. In a direct-drive machine the headstock motor spins one component at a precisely controlled, programmable speed. This is a genuine variable, not a side effect, and it matters later when we talk about why direct drive differs from inertia welding.

3. Friction (bond) force is applied. The ram brings the rotating part into contact with the stationary one under a controlled axial load. Contact is made while rotation continues, and the two faying surfaces begin rubbing against each other.

4. Frictional heating and plasticisation. Friction at the interface rapidly generates heat, softening the material locally to a plasticised, putty-like state. Critically, this happens below the melting point: the metal never becomes a liquid, which is exactly what makes this a solid-state process. AWS C6.1 summarises this phase simply: rubbing heats the workpieces locally at the joint.

5. Upsetting begins and rotation brakes. Once sufficient plasticised material has formed, the machine applies axial displacement to start the upset. Simultaneously, the direct-drive motor applies a controlled brake, bringing rotation to a precise stop at a programmed point. The ability to control how and when rotation stops is one of the strongest arguments for direct drive.

6. Forge force and bond formation. With rotation stopped, a heavier forge force is applied and held. The two plasticised faces are forced together, and the metallurgical bond forms on cessation of motion. Because the material never melted, there is no solidification step and therefore no risk of the solidification cracking that can plague fusion welds.

7. Flash formation and removal. The displaced plasticised material is squeezed outward as a collar of flash (also called the upset). That flash is actually useful during the cycle, a visual confirmation that full plasticisation and forging occurred, but it has to be machined off afterwards. Many Thompson machines offer in-cycle ID and OD flash removal, which saves a separate machining operation.

The Three Principal Parameters

Anyone specifying a friction weld soon learns that the entire metallurgy of the joint can be dialled in through three numbers:

• Rotational speed controls how much heat is generated per second.
• Applied load (friction force, then forge force) controls how that heat is distributed and how hard the plasticised faces are consolidated.
• Axial displacement (burn-off length) tells you exactly how much material has been consumed, and the machine monitors it in real time as a quality indicator.

Because all three are programmable and measurable, friction welds are unusually controllable. Every machine cycle can be logged, checked against limits, and tied to a serial number.

Why Direct Drive Differs From Inertia Friction Welding

In an inertia friction welding machine, a flywheel spins up to speed and is then simply declutched, letting stored energy dissipate at the joint until everything stops. It is a beautifully simple, purely mechanical approach, but the weld profile is governed by the flywheel’s inertia rather than by a machine that can change its mind mid-cycle.

A direct-drive Thompson machine behaves more like a CNC system than a flywheel. Speed, friction force, forge force, displacement and the braking profile are each independently programmable. The machine can hold speed under load, brake at a chosen point, apply a different forge pressure than friction pressure, and stop exactly where the programme says. For demanding applications such as API-specification drill pipe joints or dissimilar-metal combinations, that programmability is what enables consistent results across every weld on a production run.

It is worth contrasting this once with fusion processes. An arc weld on a comparable steel joint needs consumables, gas coverage, heat control, and usually a grinding pass afterwards. Compare that to reading our overview of how MIG welding works or flux welding processes, and the difference in setup complexity is obvious. Friction welding’s cycle is mechanical, repeatable, and closed-loop by nature, which is precisely why it found its way into safety-critical production lines decades ago and has never left.

The Thompson Machine Range: From 15T to 400T

Once you understand the direct-drive cycle, the next question is practical: which machine actually does the job? Thompson’s answer is a deliberately graded family of machines rather than a one-size-fits-all press. The range runs from the smallest 15-ton platform up to a 400-ton flagship, spread across nine numbered machines, Thompson 1 through Thompson 9. Each is built around the same direct-drive principle, so a procedure proven on a mid-range machine transfers logically up or down the family. What changes is forge capacity, bed length, chuck configuration and how much automation sits around the welding head. You can see the full line-up on the official Thompson friction welding machines page.

Nine Platforms, One Family

The tonnages are the specification buyers actually shop by, because forge force dictates the cross-section you can upset in a sensible cycle time. The nine platforms break down like this:

Platform	Forge capacity	Typical fit
Thompson 1	15T	Small rods, valves, sensors, small-diameter joins
Thompson 2	50T	Piston rods, small shafts, connectors
Thompson 3	60T	Light automotive and hydraulic components
Thompson 4	100T	Drive shafts, pump shafts, prop shafts
Thompson 5	125T	Heavier shafts and bi-metallic joins
Thompson 6	150T	Axles and larger hydraulic rods
Thompson 7	250T	Truck and bus axle components
Thompson 8	300T	Heavy axle and drill rod work
Thompson 9	400T	Drill pipe, large cross-sections, heavy OEM jobs

Treat that table as an indicative map rather than a fixed rule: material pairing, joint diameter and target cycle time all shift the recommendation. The important point is that the platforms share controls, tooling philosophy and software, so a business can start smaller and move up without retraining operators from scratch.

Single-Ended and Double-Ended Designs

Every platform is offered in two基本 configurations. A single-ended machine loads, welds and unloads one joint per cycle, which suits long or awkward parts and lower volumes. A double-ended machine welds two joints in the same cycle, effectively doubling output per operator hour on symmetrical components such as axle beams or piston rod assemblies. Double-ended designs are where the automation argument becomes strongest: one robot or gantry can load both ends, weld simultaneously and stack finished parts while the operator does something more useful than feeding a machine.

Tooling, Flash Removal and Measurement

Three features separate a modern Thompson from a 1970s friction welder. First, in-cycle flash removal: ID and OD upset can be turned off inside the machine before the part is unloaded, so the焊接 flash never reaches a separate machining station. Second, quick-change tooling, which cuts changeover time when a line runs mixed variants. Third, integrated measurement and marking, so dimensional checks and traceability marks are applied as part of the cycle rather than bolted on afterwards. Together these turn the welder from a standalone press into a cell in a production flow.

Thompson Welder Software and Traceability

Thompson Welder software records the parameters that define each joint: rotational speed, friction and forge pressures, burn-off distance and displacement trace. Every weld can carry a data record and a mark, which is what auditors in automotive, oil and gas and rail actually ask for. Robot load/unload is supported across the range, and KUKA’s own configurator shows more than 800 Thompson machines in production worldwide, with over 1,200 rotary friction welding machines installed across more than 44 countries.

Subcontract Welding at Halesowen

Not every buyer needs a machine. Thompson’s UK site at Halesowen runs a subcontract friction welding service under KUKA’s ISO 9001 and ISO 14001 accredited UK operation, complete with an on-premise metallurgical laboratory for sectioning, hardness checks and procedure trials. That makes it the sensible first stop for proving a joint, running a pilot batch or handling overflow, before anyone signs a capital order. For smaller aluminium fabrication where friction welding simply is not the right tool, arc processes remain the fallback, as our Aluminum MIG Welding: The Complete Practical Guide explains.

Where Thompson Friction Welding Is Used

Few joining processes cover such an extreme range of sizes as friction welding. On one end of the scale, Thompson machines fuse components small enough to sit in the palm of your hand, such as engine valves and airbag initiators. On the other, they weld structures as long as 15 metres. What links those two extremes is geometry: a circular cross-section, two parts on the same axis, and a joint that has to be as strong as the parent metal. That combination appears far more often in industry than most people expect, which is why the company’s machine platforms are named after the parts they were originally built to weld. You can see the full line-up on the Thompson friction welding machines page.

Application Series: PR, AX, DP, DR, PS and TR

Thompson groups its machines into application series, and the two-letter prefixes are a quick map of where the technology earns its keep:

• PR (piston rod) – hydraulic cylinder rods and cylinder caps for excavators, loaders and agricultural equipment, the machines commonly called “yellow and green goods”.
• AX (axle) – truck and bus drive axles and beam axles, where a full-penetration, defect-free joint along the centreline is non-negotiable.
• DP (drill pipe) and DR (drill rod) – API-specification pipe and rods for oil and gas exploration, water well drilling and sheet piling.
• PS (prop shaft) – drive shafts, pump shafts and CV joint components that run at high speed and must stay balanced.
• TR (track roller) – undercarriage rollers and similar heavy-duty tracked-equipment parts.

The same physics that suits a haulage axle also suits a hydraulic rod: a solid-state bond with no filler, no flux and no grinding back of excess weld metal. For a wider look at the process itself, see our guide to friction welding basics.

Energy, Exploration and Piling

Drill pipe and drill rods are perhaps the clearest example of a mission-critical application. A weld that fails downhole does not simply scrap a part; it costs a rig crew days of tripping pipe, and in the worst case the string is lost in the borehole. That is why these joints are made to API specifications and validated with full-strength, full-penetration welds rather than the softer acceptance criteria applied to decorative or low-load work. The same logic applies in power generation and oil and gas, where a leak path or a brittle joint is unacceptable and where friction welding’s consistent, monitorable cycle gives engineers the evidence they need.

Automotive, Safety and Small Components

Down the size scale, friction welding turns up inside parts most drivers never see. Engine valves are a classic case: a head made from a heat-resistant alloy joined to a stem of a different steel, a bi-metallic combination that would be extremely awkward to fusion weld without distortion or cracking. Pistons, drive shafts, pump shafts and constant-velocity joint components use the process for the same reasons: low heat input means little distortion, so downstream machining allowances stay tight.

Safety components are even less forgiving. Airbag initiators rely on a precise, hermetic joint between tiny dissimilar parts, where repeatability matters more than operator skill. Arc processes in this territory bring consumables, spatter and a heat-affected zone that has to be managed; if you are comparing that trade-off for aluminium parts in particular, our aluminium MIG welding guide explains what fusion joining actually demands.

Electrical Connections and Transition Joints

One of friction welding’s neatest trick is joining metals that arc welding struggles with. Copper-to-aluminium electrical connectors and pantograph rail contacts are produced this way, giving a low-resistance joint without the brittle intermetallic mess that fusion welding tends to create. Beyond that sits the specialist world of stainless steel, aluminium and titanium transition joints for cryogenic vessels, scientific instruments and nuclear plant, where a dissimilar-metal boundary must remain tough at extreme temperatures.

That range explains why Thompson keeps a subcontract welding service at Halesowen alongside its machine business, backed by a metallurgical laboratory for joint qualification. For context on how the technique has spread across sectors, friction welding applications now reach aerospace, medical and rail manufacturing worldwide.

The common thread across every application above is integrity. Where a joint is inspected, loaded, cycled or simply trusted with someone’s safety, manufacturers choose a process that produces the same metallurgical result every cycle, records the data to prove it, and never runs out of wire.

Benefits: Why Manufacturers Choose Thompson RFW

Friction welding tends to sell itself on the shop floor rather than in a brochure. After a few weeks on a Thompson direct-drive machine, the reasons for leaving arc welding collapse into the same handful: the weld gets cheaper, the part comes off straighter, and quality stops depending on who showed up for the night shift.

No Consumables, Less Dependency on Operator Skill

Thompson states it plainly: compared with conventional welding techniques, friction welding requires no consumables such as shielding gases and wire, and it does not rely on a skilled operator to produce a sound joint (Thompson Friction Welding, The Process). Arc processes carry a running cost that never appears in the machine purchase price: spool after spool of wire, cylinder after cylinder of gas, flux, grinding discs. Shops can price a year of it through a distributor such as Airgas welding supplies. Friction welding removes that line item: the parts being joined are the only thing consumed.

Operator skill is the second quiet cost. Even cheap, portable arc processes still burn through electrodes and lean heavily on the welder’s hand, as covered in Advantages of Stick Welding. A MIG or TIG welder takes years to build the control needed for consistent penetration, and that skill walks out the door with them. A Thompson machine instead runs a programmable cycle, logging force, displacement and rotation, so a new operator loading blanks gets the same joint as a veteran.

A Solid-State Bond Free of Solidification Cracking

Because the joint forms below the melting point, the bond is produced without solidification, so solidification cracking is avoided altogether (TWI technical insight on friction welding). Cracking-prone materials behave very differently in a fusion pool than in a forge bond. Anyone who has fought hot cracking in aluminium knows how narrow that fusion window is. Friction welding sidesteps it: nothing ever freezes, the material upsets in a plastic state and bonds atom to atom.

Low Heat Input, Low Distortion

Heat is concentrated at the faying surface and the cycle lasts seconds rather than minutes. The result is minimal distortion and a narrow heat-affected zone. Tube and shaft assemblies come off straight instead of bowed, so less straightening, less rework and fewer scrap bins. For stainless steels, the short cycle also limits time in the sensitising range where chromium carbides precipitate along grain boundaries and corrosion resistance drops. Low distortion also removes the fixturing and restraint usually required when arc welding long or thin-walled parts.

Dissimilar-Metal Joints Made Routine

Perhaps the most strategically valuable benefit: friction welding joins material pairs that arc processes usually refuse. Steel to stainless, copper to aluminium, mild steel to titanium, cast iron to forged steel. These bi-metallic joints are everyday work on a Thompson machine, because the solid-state mechanism limits the growth of brittle intermetallic layers that would make a fusion weld fail. That is how copper-to-aluminium electrical connectors, rail current collectors, cryogenic transition joints and bi-metallic drill pipe collars get made. Corrosion-alloy assemblies such as Alloy 20 welding also benefit, since filler-metal matching stops being the constraint it is with fusion welding.

Seconds-Long Cycles, Full Automation, Full Traceability

Welding time is measured in seconds, not minutes, letting one friction welding cell replace several arc stations. Robot load and unload, quick-change tooling and in-cycle flash removal push labour content down further, and because the machine controls every phase, the hundredth weld matches the first. Each weld can be recorded against applied load, axial displacement and burn-off, with Thompson Welder software storing the data against a component identity. In mission-critical sectors, that record is nearly as valuable as the joint itself.

Material Savings and Less Post-Weld Work

Friction welding joins near-net-shape parts, so designers can put the expensive alloy only where it earns its keep: a stainless tip on a plain steel shaft, an aluminium body on a steel thread. It also eliminates the filler metal arc welding would have added, and because the bond is sound as forged, post-weld heat treatment is often unnecessary, shortening lead times.

The Bottom Line

None of these benefits is unique in isolation, but the combination is. A process that consumes nothing but electricity and the parts themselves, produces a crack-free solid-state bond in seconds, handles dissimilar metals and documents every weld automatically is hard to match at volume. That is why Thompson machines sit in axle shops, hydraulic cylinder plants, oilfield supply chains and cryogenic research facilities.

Limitations, Challenges and Common Mistakes

While Thompson friction welding can produce incredibly strong, repeatable joints in a matter of seconds, it is not a process that can be forced onto every part. Before investing in a machine or outsourcing to a subcontractor, it is worth knowing the very real constraints that come with it.

Design and Geometric Restrictions

The biggest hurdle is shape. For conventional direct-drive rotary friction welding, the faying surfaces must be able to rotate relative to each other. Because of this, production is almost entirely limited to circular cross-sections, such as bars, tubes, and shafts. As stated in AWS C6.1, friction welding in production is essentially an automatic process for circular components. Non-round profiles require alternate variants such as orbital, linear, or friction stir welding to achieve a bond. Even with round parts, extreme length-to-diameter ratios can introduce issues with axial rigidity and part alignment, both of which are vital for keeping the joint concentric under load.

Flash, Machining and Machine Costs

Every sound rotary friction weld creates upset, more commonly known as flash. This ring of plasticised material is a by-product of the bond formation, and in the vast majority of applications, it must be removed by a secondary machining operation. For high volume parts, this can be automated in-cycle, however for many others, it is an added cost that must be factored into the total price per component.

On top of this, the machines themselves are a major investment. Tonnages from 15T all the way up to 400T are built for longevity, however the justification has to come from the part count, cycle time and labour saved. For lower volumes, sending parts to a subcontract service is often the far more realistic route.

Dissimilar Metals, Surfaces and the Heat-Affected Zone

One of Thompson’s greatest strengths is its ability to join dissimilar metals. However, this comes with strict control. When welding two different alloys, brittle intermetallic compound layers can form at the interface. The advantage of friction welding is that its peak temperatures are far lower than fusion processes, massively reducing this risk, but it does not remove it entirely. Careful selection of compatible pairs and tightly controlled burn-off are needed to keep these layers thin enough to remain ductile.

Surface condition is just as unforgiving. Oxides, oils, cutting fluids and even fingerprints act as contaminants. These must be expelled from the joint interface, which is why interface contamination is expelled as flash. For this reason, adequate burn-off and the correct surface preparation are critical to a sound joint. For parts that are more prone to tenacious oxides, such as aluminium, following the same discipline as a job that would utilise aluminum cleaner for welding can go a long way. In comparison to many arc processes, such as those covered by aluminum flux core welding, far less filler is involved, however cleanliness is non-negotiable all the same.

Finally, whilst distortion is minimal in comparison to fusion welding, a heat-affected zone (HAZ) is still present. Depending on the material, this can result in hardness changes, most notably in quenchable steels. For some of these alloys, this is manageable, however for others, it may require consideration against the requirements of a code. This is especially true for higher strength grades, similar to those specified for alloy steel welding electrodes, where mechanical properties must be maintained.

Troubleshooting: Common Mistakes and Fixes

With so many variables, small errors can quickly ruin a weld. The majority of defects come down to parameters or preparation:

• Weak welds: Usually caused by insufficient forge force, too little burn-off, or incorrect speed. These fail to fully consolidate the plasticised material and must be corrected through validated procedure settings.
• Uneven flash: Typically down to poor axial alignment or uneven face preparation. Even a slight taper or lack of concentricity will cause the material to extrude disproportionately around the circumference.
• Lack of bonding: Almost always linked to contamination or inadequate friction time. If oxides aren’t expelled, a true solid-state bond cannot form across the entire interface.
• Excessive burn-off: Over-applying friction time or load can waste material, create an oversized upset to machine off, and in extreme cases, overheat the joint. This is wasteful and can increase the size of the HAZ unnecessarily.

As outlined by TWI’s technical insight into friction welding, these variables are highly interdependent. With Thompson’s direct-drive system, each phase is programmable and logged for traceability, however that advantage is only realised if a weld procedure has been correctly qualified. At the end of the day, the process rewards precision. Cut corners on cleanliness, alignment or parameters, and the biggest benefit of friction welding, its consistency, is the first thing to be lost.

Standards, Quality Control and Inspection

A friction weld is only as good as the evidence behind it. In a mission-critical application, “it looked fine when it came off the machine” is not a quality strategy. What makes Thompson’s direct-drive process attractive to engineers is that the joint is built from controlled, measurable inputs, and every one of those inputs can be recorded, checked and audited. This section walks through the standards, monitoring, testing and traceability framework that surrounds a production friction weld.

The Standards That Govern Friction Welding

The starting point in North America is AWS C6.1, Recommended Practices for Friction Welding. It is worth quoting because it captures the whole philosophy of the process: AWS defines friction welding as a solid-state process in which coalescence is produced by the heat developed from mechanically induced rubbing between moving parts under compressive force, without filler metal, flux or shielding gas. Nothing melts into a pool, so there is no molten weld metal to sample and no atmospheric protection to worry about. Quality therefore comes down to force, motion, displacement and the resulting bond.

In Europe, the relevant standard cited for Thompson subcontract work is BS EN 15620:2019, Friction welding of metallic materials, which KUKA UK references directly for its Thompson rotary friction welding machines and related services. Alongside the process standards sit the familiar management system certifications: ISO 9001 for quality management and ISO 14001 for environmental management, which most automotive, energy and oilfield customers treat as a baseline requirement for approving a supplier. Where components go into exploration or drilling service, additional product specifications apply, most notably API’s drill pipe requirements, which govern everything from material grade to the weld region of the pipe body.

If you are new to how these frameworks differ from those covering fusion processes, our Aerospace Welding: Processes, Standards, and Careers Guide breaks down how qualification works in regulated industries.

In-Process Monitoring: Catching Bad Welds Before They Exist

The biggest quality advantage of an automated direct-drive cycle is that it is fully instrumented. A Thompson machine continuously tracks a handful of variables, and the ones that matter most are:

• Rotational speed during the friction phase and how cleanly it brakes to zero.
• Applied load (force) through both the friction/bond stage and the forge stage.
• Axial displacement, usually described as burn-off or upset — the exact distance the parts shorten as material plasticises and is expelled.
• Timings: how long friction lasted, when braking started, how long forge pressure was held.

Because each parameter has a permitted band, the machine can compare the real-time curve against a qualified reference and flag any weld that falls outside it. A short burn-off, a slow brake or a force spike is rejected automatically rather than discovered three weeks later in the field. Consistency like this is why aluminum MIG welding and other arc processes, which depend on operator technique and consumable condition, struggle to match friction welding’s repeatability at volume.

Inspection: Non-Destructive and Destructive Testing

Non-destructive testing checks production parts without damaging them. Common methods include ultrasonic testing, which can probe the bond line for any lack of consolidation, and magnetic particle inspection (MPI) for surface and near-surface discontinuities in ferromagnetic components. Visual checks of the flash are also diagnostically useful: uneven or asymmetric upset often points to misalignment or inconsistent heating before any instrument is applied.

Destructive testing is used to qualify the procedure rather than to inspect every part. Sample welds are sectioned for macro and micro examination of the bond line, then subjected to tensile, bend and hardness testing. This is where you confirm that the heat-affected zone is sound, that no brittle intermetallic layers have grown in a dissimilar-metal joint, and that the failure location is in the parent material rather than along the weld. Surface condition is part of that qualification too: oil, oxide and coatings must be controlled before joining, and aluminium work in particular demands disciplined cleaning, as covered in Aluminum Cleaner for Welding: How to Prep Aluminum Right.

Traceability and Procedure Qualification

Thompson machines can integrate component measurement, in-cycle flash removal, part marking and Thompson Welder software, giving what the manufacturer describes as total traceability across the machine range. Every weld’s parameters are logged against a serial number, so a field return can be traced back to its exact force and displacement record. Procedure validation follows the familiar logic of welding procedure qualification: parameters are proven on representative material, documented as the approved procedure, and requalified whenever material grade, geometry or the parameter set changes.

Note: The practical test when choosing a supplier is simple — ask how they would retrieve the weld data for one specific failed part, and how long a procedure requalification would take after an engineering change. If either answer is vague, treat that as a risk.

Thompson RFW Compared With Other Joining Processes

No joining process gets chosen in a vacuum. An engineer specifying a hydraulic piston rod, a drive axle or a length of drill pipe has to weigh rotary friction welding against everything else on the shortlist, and the answer depends on geometry, alloy, production volume and what the joint must survive. Here is how Thompson’s direct-drive rotary friction welding (RFW) compares with its closest relatives and with the more familiar fusion and filler-based alternatives.

Direct-Drive RFW vs Inertia Friction Welding

These two are siblings, and the most common source of confusion. In inertia friction welding, energy is stored in a heavy flywheel spun up to speed before the parts ever touch. The workpiece is then released against the stationary component, and the stored kinetic energy is spent in a single coast-down. Once the flywheel is engaged, the weld is largely committed: speed, torque decay and burn-off follow from how much energy was stored and how the interface behaves as it slows.

Thompson’s machines take the other route. A motor-driven spindle delivers direct drive, so rotational speed, friction (bond) force, forge force, the braking point and axial displacement are each controlled independently and can be programmed as distinct phases of the cycle. That flexibility matters when a small-diameter rod meets a large flange, when material pairings change, or when you want to shape the heat-softened zone precisely. It also makes the process highly repeatable in an automated cell: same recipe, same burn-off, same flash, part after part. As Thompson describes the cycle, the bond forms on cessation of motion under forge force, and the machine, not a flywheel, decides exactly when that happens.

Linear Friction Welding: Oscillation Instead of Rotation

Linear friction welding swaps rotation for back-and-forth sliding. One component is held stationary while the other oscillates against it under pressure; friction heat plasticises the interface, and the parts are forged together the instant oscillation stops. The concept was patented as far back as 1924, but it found its killer application decades later in aerospace: titanium aeroengine blisks (bladed discs), where blades are joined into a disc slot with no filler metal, no casting defects and almost no machining allowance. LFW handles joint shapes rotary friction cannot reach, such as rectangular roots and airfoils. The trade-off is that the machines are specialised and costly, and clamping must be exceptional. For a shaft, axle, tube or drill pipe, rotary friction welding remains the practical workhorse.

Friction Stir Welding: TWI’s 1991 Breakthrough

Friction stir welding (FSW) is the related process invented at TWI in 1991. Rather than joining two workpieces directly to each other, a non-consumable rotating tool with a shoulder and a probe is driven along the joint line, mechanically stirring the material into a solid-state bond without reaching the melting point. FSW excels on aluminium alloys and on plate, sheet and extrusion joints — aircraft structures, ship decks, heat exchanger panels — where fusion welding would distort the metal or degrade its properties. It is not a substitute for rotary friction welding on round components, but it shares the same appeal: no filler, low distortion, and a fine-grained joint.

Against Arc Welding and Brazing

MIG, TIG and stick are fusion processes: they melt the base metal and rely on consumable wire or electrodes plus shielding gas or flux, with an operator’s hand controlling the puddle. That skill dependence, heat input and resulting distortion are exactly why the haulage axle maker in our opening example walked away from arc welding. Where arc welding is still the right answer, our guides to aluminum MIG welding and aerospace TIG welding cover the technique properly, and the case for stick welding explains when SMAW still wins.

Brazing sits further down the strength scale: a filler alloy melts and wets the parts by capillary action while the base metals stay solid. It joins dissimilar materials cheaply with little distortion, as air-acetylene brazing shows, but the joint is only as strong as the filler, it demands clean, fluxed surfaces, and it never achieves the parent-metal continuity of a friction weld.

Side-by-Side: Motion, Joint, Consumables, Cost and Cycle Time

Process	Motion	Joint geometry	Consumables	Typical cycle	Capital cost
Thompson direct-drive RFW	Rotation, then controlled braking	Circular sections: rods, tubes, shafts, axles	None	Seconds to about a minute	High
Inertia friction welding	Flywheel spin-up and coast-down	Circular sections	None	Seconds	High
Linear friction welding	Linear oscillation	Rectangular and airfoil roots, blisks	None	Seconds to minutes	Very high
Friction stir welding	Rotating probe traversing a seam	Plates, extrusions, closed seams	Non-consumable tool only	Minutes per metre	Very high
MIG / TIG arc welding	Electric arc	Nearly any geometry	Wire or electrode, gas, flux	Operator-paced, continuous	Low to moderate
Brazing	Torch or induction heat	Overlaps and close-fit joints	Filler alloy, flux	Seconds to minutes	Low

What the Comparison Means for a Buyer

If the part is round, volumes are serious, and the joint must match or beat the parent metal, direct-drive rotary friction welding is usually the fastest route with the lowest total cost per joint, because consumables, skilled labour and secondary straightening largely disappear. Inertia welding earns its place where one fixed energy setting suits the part and simplicity is prized. Linear friction welding is the specialist for aerospace blisk work, friction stir welding owns the aluminium plate and extrusion seam, and arc welding and brazing remain the flexible, low-capital options for low volumes and complex geometries. Thompson’s platform, now within the KUKA Group’s machine family, is built squarely around that high-volume, round-part, zero-consumable sweet spot.

Practical Tips, Buying Guidance and Conclusion

With a solid understanding of Thompson friction welding, the task shifts from theory to execution. The decisions you make before the first trial weld will have the biggest impact on both cost and quality, so it’s worth getting them right from the outset.

Scope Your Application First

Rather than starting with a machine quote, work through the following points in order:

1. Material pairing. Dissimilar joints, such as steel to aluminium, copper to aluminium, or titanium transitions, are where this solid-state process excels. However, intermetallic growth must be controlled by speed and forge load. If fusion is still a consideration for a tricky alloy, it’s worth reviewing the demands of a procedure like those for Alloy 20 Welding: Procedures, Filler Metals, and Best Practices.
2. Joint cross-section. Conventional rotary friction welding works best for circular or near-axisymmetric sections that can be gripped and pushed axially. Irregular profiles or thin sheet tend to fall outside its practical range.
3. Tonnage class. Thompson platforms range from 15T up to 400T to suit everything from small parts to heavy axles and drill pipe. Size your machine to the required forge load for your section, not the largest part you might one day weld. The current lineup can be viewed on the Thompson friction welding machines page.
4. Cycle time and throughput. For many shaft applications, a direct-drive cycle is only seconds long. Be realistic with your target throughput, as this alone can determine whether ownership is viable.
5. Flash removal. All rotary welds produce upset material. Factor in in-cycle removal, secondary machining, fixtures and spindle time when costing the assembly.
6. Automation. Only add features like part marking, in-process measurement, or data logging if they are needed for traceability or volume. Avoid over-specifying for a prototype run.

Buy or Subcontract?

For high, repeatable volumes with a stable design, a dedicated machine can pay off. For everything else, subcontracting to an accredited facility is usually the smarter route. It allows you to qualify a procedure, receive test reports and prove the joint without tying up capital.

This near-net-shape approach is well documented by TWI’s research on linear friction welding, which found that near-net-shape parts saved over 50% in material, production and overall part cost. That same principle applies to rotary work: weld closer to finished size and machine less away.

Questions to Ask and Mistakes to Avoid

Before committing to any supplier, get clear answers to the following:

• Qualification: Which standard will the procedure be qualified to, who signs off the coupons, and will testing be in-house?
• Traceability: Is burn-off, force and displacement logged per part, and how are records retained?
• Scalability: How much tonnage headroom remains if your section or material pairing changes?
• True cost: What is the breakdown for flash removal, both in-cycle and secondary?

To protect your budget, watch out for these common pitfalls:

• Guessing tonnage: Undersized machines won’t reach the required forge load, whilst oversized ones inflate capital and floor space.
• Skipping qualification: A clean-looking weld isn’t proof of a bond. Follow AWS C6.1 (Recommended Practices for Friction Welding) where applicable for your industry requirements.
• Forgetting flash: Budget for its removal from day one, not at the end of quoting.
• Poor surface prep: Oil, oxides or coatings will prevent bonding. Proper cleaning is just as critical as it is for aluminium, as outlined in Aluminum Cleaner for Welding: How to Prep Aluminum Right.
• Blind parameter setting: Rotational speed, load and axial displacement must be validated, not set by feel.

Final Takeaway

Thompson friction welding is at its most cost-effective when approached with discipline. Scope honestly, prove the joint through metallurgical qualification, and only price the assembly with flash removal accounted for. If you’re weighing it against fusion routes to build a complete cost comparison, our Aluminum MIG Welding: The Complete Practical Guide can help put consumables, heat input and operator skill into perspective. Get those fundamentals right, and this solid-state process can deliver its full value on your parts.

Frequently Asked Questions

What is Thompson friction welding?

Thompson friction welding is the direct-drive rotary friction welding process and machine range originally developed by Thompson Friction Welding in the UK and now built by KUKA. One component is rotated against another under axial pressure, generating frictional heat that plasticises the interface, after which rotation stops and a forge force creates a solid-state joint without melting.

How does Thompson’s direct-drive process differ from inertia friction welding?

Inertia friction welding stores energy in a spinning flywheel and lets it coast down, so speed and force fall together. Thompson’s direct drive uses a powered spindle that controls rotational speed, friction force and forge force independently, giving programmable, repeatable weld cycles.

What machines does Thompson offer?

The Thompson range covers nine platforms from 15-tonne to 400-tonne forge capacity, in single- and double-ended configurations, with application-specific series for piston rods, axles, drill pipes, drill rods, prop shafts and track rollers. Options include in-cycle internal and external flash removal, quick-change tooling, measurement, marking and robotic load/unload.

Which materials and joints can Thompson friction welding join?

It joins both similar and dissimilar metals in bar-to-bar, tube-to-bar, bar-to-plate, tube-to-plate and tube-to-tube geometries. Common pairings include carbon and stainless steels, alloy steels, copper-aluminium electrical connections and stainless-to-aluminium or titanium transition joints, because the solid-state bond limits brittle intermetallic formation.

How long does a friction weld cycle take?

Cycle times can be as short as a few seconds depending on material combination, section size and required burn-off. The whole weld — spin-up, friction phase, braking and forge — typically completes in seconds, followed by flash removal and any finishing operations.

Does friction welding need filler metal or shielding gas?

No. Friction welding is autogenous: no filler metal, flux or shielding gas is required because the parts never melt. That removes consumable cost, spatter and fume issues, and eliminates solidification cracking associated with fusion welding.

What standards govern friction welding quality?

Key references include AWS C6.1 Recommended Practices for Friction Welding and BS EN 15620:2019 Friction welding of metallic materials, alongside quality management standards such as ISO 9001. Drill pipe work also has to satisfy oil-industry specifications such as API requirements.

What are the main limitations of Thompson rotary friction welding?

Conventional rotary friction welding is essentially limited to circular cross-sections, and every weld produces flash (upset) that must be machined away. Machines are also significant capital investments, and dissimilar-metal joints need metallurgical qualification to control intermetallic layers.

Can I use subcontract friction welding instead of buying a machine?

Yes. KUKA’s Halesowen facility runs subcontract friction welding on Thompson machines with an in-house metallurgical laboratory, pre- and post-weld heat treatment and part preparation. This suits irregular or low-volume runs, or companies that want an end-to-end service before committing to capital equipment.