CAD Welding Process: The Complete Guide to Exothermic Welds

A substation crew once tightened every ground lug by torque wrench, logged the values, and still watched a fault event melt the connection a year later. The fix was not a bigger wrench — it was a bond made by fire inside a graphite mold, a molecular joint that cannot loosen because it was never mechanically clamped in the first place.

What Is CAD Welding? Definition, Origin, and Why It Matters

CAD welding is a method of joining electrical conductors with a chemical reaction instead of a mechanical grip. Also called exothermic welding, cad weld, or thermite welding, it works by triggering a self-sustaining reaction inside a reusable graphite mold. That reaction turns a charge of metal powder into a small pour of molten copper, which fuses the conductor ends into one continuous piece of metal. There is no arc, no electrode, and no external heat source melting the joint — the energy is built into the material itself.

The result is what the industry calls a molecular bond. Unlike a bolted lug or a crimp, where two surfaces are pressed together and held by friction, an exothermic connection becomes a single mass of copper with the same or better current-carrying capacity as the conductors it joins. That distinction is the whole reason the process exists, and it is why utilities, railroads, and lightning protection crews keep reaching for it nearly a century after it was invented.

A Brand Name That Became the Industry’s Default Word

CAD welding is really a trademark habit. CADWELD is the registered product name from nVent ERICO for its line of exothermic welding products, and like many dominant brands, it slipped into everyday speech. Technicians now say “cad weld” for almost any exothermic connection, whether the part came from nVent, Harger, TE Connectivity, or another manufacturer. The neutral, accurate term is exothermic welding or exothermic connection — but if you are reading job specs, drawings, or inspection reports, expect to see CADWELD used generically.

The story starts with the railroad. CADWELD was introduced to the rail industry in 1938 as a reliable way to bond rail ends and signal cables, and it became the standard for exothermic welded track connections, a position nVent still documents today. From track, the process migrated into electrical infrastructure, and the scale grew quickly: nVent states that Cadweld exothermic connections have been specified and installed more than 100 million times worldwide.

Where CAD Welding Is Used

The applications share one theme: a connection that must carry fault current, survive decades underground or outdoors, and never be allowed to loosen. The main ones are:

• Grounding and grounding electrode bonding — substation ground grids, building ground rings, and copper-clad ground rods where joint resistance must stay stable for the life of the site.
• Lightning protection — air terminals, down conductors, and structural steel connections in roof-mounted systems, where a high-current pulse arrives with almost no warning.
• Cathodic protection — pipeline and storage tank anode leads, where joints sit in soil or backfill and cannot be conveniently re-torqued later.
• Rail bonding — signal bond wires, weld-bonded rail, and cable-to-rail connections on track.
• Telecommunications — tower grounding, equipment grounding conductors, and shield bonds at cell sites and antenna farms.

The Three Pieces of the System: Mold, Reaction Material, Ignition

People often blur the equipment together, so it helps to separate the three parts. They do different jobs, and each one affects whether the weld succeeds.

Component	What it is	What it does
Graphite mold and clamp	A machined graphite block with a weld cavity and tap hole, held shut by a handle clamp.	Contains the molten metal, shapes the joint, and sets the geometry for a specific conductor combination.
Reaction material	Welding material (copper oxide mixed with aluminum granules) plus a small amount of starting powder.	Provides the fuel for the reaction that yields molten copper and aluminum oxide slag.
Ignition system	Either a flint spark gun or a self-contained electronic ignition cartridge.	Starts the reaction. Once lit, it needs no further input and finishes in seconds.

The reaction chemistry itself is an aluminum reduction, the same basic family of aluminothermic behavior you can read about in our guide to aluminum flux welding — though CAD welding uses it to produce copper, not to repair an aluminum part.

That independence, plus a joint that cannot work itself loose, is why CAD welding still matters. The next section breaks down what actually happens inside the mold during those few seconds of reaction.

The Chemistry Behind the Exothermic Reaction

Ask most people how a CAD weld works and they will describe it as “welding with fire.” That is close, but not precise. A CAD weld is not fusion from an arc or a flame. It is a chemical reaction that manufactures its own heat and its own weld metal at the same time. Understanding that chemistry is what separates an operator who pushes the right button from one who actually knows what is happening inside the mold.

The Thermite Reaction in Plain Language

The reaction at the heart of the CADWELD exothermic connection is a variation of the classic thermite reaction. The charge inside the crucible is essentially two powders mixed together: copper oxide (the oxidizer) and finely divided aluminum powder (the fuel). Once ignited, aluminum does what aluminum does best: it goes after oxygen more aggressively than copper can hold it. The aluminum strips the oxygen away from the copper oxide and produces two new substances.

In chemical shorthand: 3CuO + 2Al → 3Cu + Al₂O₃

Read that in English and you have the entire process. The products are molten copper and aluminum oxide. The molten copper becomes the weld metal that fuses the conductors together. The aluminum oxide becomes slag, the grayish crust you see after the mold comes apart. No electrode, no generator, and no compressed gas are involved. The reaction carries everything it needs inside the powder, which is why a full welding kit can fit in a small case.

The heat released is extreme. According to the exothermic welding manual, the process produces molten material in excess of 2500°F (1400°C), and the entire reaction runs to completion in only a matter of seconds. That combination of temperature and speed is what allows the copper to flow into every gap around the cleaned conductors and fuse with them before it solidifies.

Anatomy of the Layered Charge

The reaction does not happen all at once in one loose pile. The charge is carefully layered inside a graphite crucible, and each layer has a job. From top to bottom, the standard arrangement looks like this:

Layer	What it is	Its job
Starting powder	Highly ignition-sensitive granules spread on top	Catches the spark or flame and starts the reaction reliably
Welding material	Copper oxide and aluminum powder premixed	Reacts to create molten copper and aluminum oxide slag
Steel disk	Thin steel shim covering the tap hole	Temporary plug that holds the charge until full reaction temperature is reached
Crucible	Upper graphite chamber with a tap hole at the base	Contains the reaction and meters the pour
Weld cavity	Lower graphite chamber holding the conductors	Molds the molten copper around the bare conductors into the final joint shape

Two details in that table do most of the work. First, the steel disk. It plugs the tap hole so the powder cannot fall into the weld cavity before it reacts. As the reaction builds, the disk absorbs heat until it reaches its melting point and melts through at exactly the right moment, releasing a fully molten, fully reacted charge of copper downward into the cavity. If the disk let go too early, unreacted powder would pour out and give you an incomplete, crumbly joint. The disk is essentially a mechanical timer built into the chemistry.

Second, the slag. Aluminum oxide is far less dense than liquid copper, roughly a third as heavy, so it floats to the top of the pour and rides up into the crucible instead of sinking into the joint. That buoyancy is why the weld metal underneath solidifies clean and why the slag can be chipped away after the mold is removed. The separation happens on its own, with no skimming and no flux added by hand.

The same aluminum chemistry shows up across the wider welding world, though usually with less dramatic intent. Its reactivity is exactly why welders spend so much effort removing the oxide layer before laying a bead, as covered in Aluminum Cleaner for Welding: How to Prep Aluminum Right, and why flux is used to shield and clean the puddle in processes described in Aluminum Flux Welding: How It Works and When to Use It. In a CAD weld, that same appetite for oxygen is harnessed deliberately as the heat source itself.

The modern methods of exothermic welding literature makes the point clearly: the graphite mold, the layered charge, and the timed release are what turn a violent chemical reaction into a controlled, repeatable, and inspectable electrical connection. Once the pour is complete, the copper solidifies within about 30 seconds, and the conductor-to-conductor bond that remains is metallurgical rather than mechanical. That distinction matters for everything that follows in this guide, from current-carrying capacity to long-term resistance stability.

Step-by-Step: How a CAD Weld Is Made in the Field

Exothermic welding looks dramatic from a distance, a flare and a puff of white smoke, but the real work happens before anyone strikes an igniter. The reaction itself lasts only a couple of seconds. Everything that decides whether that joint will carry fault current for the next forty years happens in the minutes beforehand: the part you chose, how clean the copper was, how dry the mold was, and which way the mold was pointed.

A typical field kit is refreshingly small. A two-piece graphite mold, a clamp, a wire brush, a propane torch, tubs of welding material and starting powder, a flint gun or electronic igniter, and your PPE. No compressor, no power source, no crimping tool, no torque wrench. nVent’s CADWELD line organizes all of it into numbered part combinations, and reading that chart correctly is step one.

The Field Workflow, Start to Finish

1. Select the mold and part number. Match the part number to the exact conductor combination: cable-to-cable, cable-to-rod, cable-to-rail, cable-to-terminal. Conductor size, strand count, and material all matter. A part sized for 4/0 cable will under-fill a 250 kcmil run, and a cavity cut for copper will not seat an aluminum conductor. Check the selection chart before you open the case.
2. Clean the conductors to bright metal. Strip back insulation, then remove oxide, dirt, oil, and flux residue with a wire brush or abrasive cloth until the surface shines. Harger’s installer guidance is explicit that conductors must be clean and completely dry, because contamination vaporizes inside the cavity and leaves porosity or an unfused pocket in the joint. The prep discipline covered in Aluminum Cleaner for Welding: How to Prep Aluminum Right transfers directly to exothermic joint quality.
3. Assemble the mold and clamp. Seat both conductors fully into their cavities so the ends meet in the weld chamber with no visible gap. Close the mold halves and tighten the clamp firmly. A loose clamp is the single most common cause of molten copper leaking out around the parting line, and a leak both ruins the weld and damages the graphite.
4. Preheat with a torch. Play a propane flame through the mold cavity and across the bare conductor ends for a minute or two. This drives off absorbed moisture that would otherwise flash to steam mid-reaction, degrading the weld or cracking the mold. Preheat is non-negotiable on cold mornings, in humid climates, or after molds have ridden around in an unheated truck. Cylinder selection and handling basics are covered in AllGas Welding Supply: Gases, Services & Safety Guide.
5. Load the charge. On traditional systems, place the steel disk over the tap hole, pour the welding material into the crucible, and lay a strip of starting powder across the lip so it runs to the ignition point. On modern self-contained products, you instead drop in a sealed cartridge or ignition insert that already holds the measured charge.
6. Step back, then ignite. Move clear of the mold opening, do not lean over the crucible, and fire the flint gun or squeeze the electronic trigger. Expect a bright flare and white smoke within a second.
7. Wait roughly 30 seconds. Let the reaction finish and allow about 30 seconds for the mold to cool before removing it from the wires. Pulling it early means disturbing semi-solid metal and pulling the conductor out of alignment.
8. Remove, brush, and inspect. Release the clamp, open the mold, tap the slag from the crucible, and brush the new joint bright. Look for a continuous fillet of copper around every conductor, no cracks, no holes, no unfused area, and no evidence that molten metal escaped at the parting line.

Why the Mold Must Stay Vertical

This is the one rule installers break most often, and it is a physics rule rather than a manufacturer preference. The system is gravity-fed: the reaction melts the steel disk, and molten copper then falls through the tap hole into the weld cavity below. Tip the mold sideways or upside down and the metal cannot reach the cavity cleanly, the pour stalls, or copper runs out of the parting line instead of into the joint. Mold part numbers are designed and printed for a specific orientation, so read the cast marking on the body and set the assembly up the way the drawing shows it, every time.

Two habits finish the job properly. First, log the part number, mold identification, and location for every connection, so a failed weld can be traced back to a specific mold nearing the end of its rated life. Second, treat visual inspection as a pass/fail gate rather than a formality; a joint with a thin fillet, discoloration, or trapped slag belongs on the list for rebuild or mechanical backup, not buried in a trench. The Harger installer guide walks through these same acceptance cues alongside the part-selection charts, and it is worth reading before your first live job.

Traditional vs. Modern Exothermic Welding Systems

The chemistry never changed. Copper oxide is still reduced by aluminum, and the joint is still made by molten copper rather than by torque. What has changed since the first rail bonds in 1938 is how crews start the reaction and how much gear they haul to the job site. Crews now choose between two families of systems, and the choice affects reliability, weather tolerance, training time, and per-joint cost.

The Traditional Powder-and-Flint System

The classic setup is what most veteran grounding contractors learned on. You clamp the graphite mold around the conductors, pour in a tin of welding material, sprinkle a pinch of starting powder across the top of the charge, and then strike a flint ignitor to throw a shower of sparks onto it. The starting powder catches, drops through the tap hole, and ignites the main weld material. The steel retaining disk melts at the right moment, and the molten copper drops into the weld cavity.

Done well, this method produces excellent joints, and it has decades of field history behind it. But it is genuinely technique-sensitive. Wind can blow starting powder off the crucible before you strike the flint. A damp mold or humid conditions can spoil ignition or crack the graphite, which is why preheating with a propane torch is mandatory in most circumstances. The quantity of starting powder is judged by eye, so two operators can produce noticeably different ignition behavior. And leaning over an open mold with a flint gun puts you closer to the flare-up than you need to be.

Self-Contained, Electronically Ignited Systems

The modern answer packages the welding material and the ignition source into one sealed unit. nVent ERICO CADWELD Plus combines an electronic ignition system with a self-contained weld material package: the starting charge is factory-metered and sealed inside the assembly, and a battery-powered firing unit sets it off at the push of a button. Harger’s UltraShot and NUWTube products and TE Connectivity’s AMPWELD cartridge-style systems follow the same basic philosophy. There is no flint gun, no loose bottle of starting powder, and no exposed charge for wind or moisture to interfere with.

That single change fixes several traditional pain points at once:

• Ignition consistency. The starting charge is measured at the factory rather than sprinkled by hand, so every weld receives identical ignition energy regardless of operator habit.
• Wind and weather tolerance. There is no loose powder to blow away. Preheat is still required to drive moisture out of the mold, but ignition itself becomes far less weather-dependent.
• Portability. A cartridge system cuts the number of loose consumables a crew carries into a substation, a vault, or a hillside right-of-way.
• Safety margin. The operator triggers ignition electronically and can stand clear, reducing exposure to molten metal and fume.

Side-by-Side Comparison

Factor	Traditional (flint + starting powder)	Modern (self-contained/electronic)
Ignition method	Flint ignitor and loose starting powder	Sealed electronic ignition cartridge
Wind sensitivity	High; loose powder can scatter	Low; charge is enclosed
Operator skill needed	Higher; technique-sensitive	Lower; more repeatable
Parts to carry	Molds, welding material, starting powder, flint gun	Molds, self-contained weld units, firing tool
Consumable cost per weld	Lower	Higher
Failed-weld/rework risk	Higher on exposed sites	Generally lower

Cost, Training, and Where Each System Wins

Cost trade-offs are subtler than they look. Traditional welding material is cheaper per joint on paper, but a single failed weld in a live substation can cost far more than the difference in consumables once you count return trips, rework, and outage time. Crews that do occasional grounding work often prefer cartridges because they shorten the learning curve and cut down on forgotten accessories. High-volume crews with experienced operators may still find traditional kits economical on large jobs with favorable weather.

Surface preparation matters in both systems, especially with aluminum conductors. A quick refresher on aluminum cleaner for welding is worth the five minutes before you commit to a weld on an aluminum run.

The Mixing Problem: Never Cross Manufacturers

This is the failure mode that surprises people. A mold and its welding material are engineered and tested as a matched pair. Fill volume, reaction rate, gas generation, and cavity geometry are all tuned together. Filling one manufacturer’s mold with another maker’s welding material can change the reaction dynamics enough to produce porous, slag-contaminated, or under-filled welds that meet neither manufacturer’s standards. ERICO’s technical literature states it plainly: mixing molds and welding materials from different sources can yield welds that conform to neither. If you inherit a mixed inventory from a previous contractor, quarantine it and standardize before the next job starts.

Finally, qualification follows the assembly, not the brand slogan. Systems tested to IEEE 837 passed that sequential test as a complete configuration. Substitute a component and you are outside the tested combination. Keep the mold, the weld material, and the ignition method exactly as the manufacturer specifies, and the joint you make today will outlive the conductor it connects.

Standards, Codes, and Qualification Requirements

A CAD weld can look flawless and still be unproven. In substation and rail work, visual acceptance is only the first gate: engineers need documented evidence that the connection type survives corrosion, mechanical stress, and fault current before anyone is allowed to make one in the field. That evidence comes from qualification standards, and for exothermic welding three documents do most of the heavy lifting.

IEEE 837: The Benchmark for Substation Grounding Connections

IEEE 837, with its 2014 revision, is the standard for qualifying permanent connections used in substation grounding. The nVent ERICO CADWELD guidance on the updated IEEE Std 837-2014 requirements explains why the revision mattered: it tightened the test sequence and raised the bar for how connections demonstrate they can handle real fault conditions.

The key word is sequential. A single specimen is put through a series of tests in a fixed order, so the connection must absorb earlier punishment and still perform at the end. Environmental exposure, accelerated aging and corrosion, and mechanical loading all come before the final exam: the electromagnetic force (EMF) test. Placing the EMF test last is deliberate. It answers the question that matters most during a fault event: will the joint stay intact when enormous magnetic forces try to rip it apart?

For that test, the standard specifies test currents of 65 kA and 75 kA, which correspond to 175 kA and 202 kA first-cycle positive peak currents. Those are the magnitudes a substation ground lead can realistically see during a short circuit, not laboratory abstractions. A connection that stays bonded, keeps its resistance low, and shows no mechanical displacement under those conditions has demonstrated something torque values on a lug never could.

IEEE 80 and the Bigger Grounding Picture

IEEE 837 does not exist in isolation. IEEE Standard 80 sets the safety framework for substation grounding, including step and touch potential limits that keep people on a substation pad safe during a fault. The grid calculations behind those limits quietly assume every connection in the grid behaves as designed. That is why IEEE 837 explicitly notes that connections meeting its requirements should satisfy the connection criteria of IEEE Standard 80. Passing 837 is how a manufacturer closes that loop: the joint assumed in the safety math is a joint proven to still be a joint after abuse.

AWS D15.2 for Railroad Rail and Rail Bonds

Railroad work has its own qualification culture. AWS D15.2/D15.2M specifies minimum standards for welding rails and rail components, including rail bonds and thermite welding. Where IEEE 837 is written around substation grounding, D15.2 is written around the demands of track: cyclic loading, vibration, thermal cycling across seasons, and the need for bond integrity that survives heavy axle traffic. Rail bonding applications such as signal circuit bonds and cable-to-rail connections follow this standard, and exothermic rail connections are qualified against it rather than against a general structural code. Railway connection solutions documentation typically references D15.2 test results for exactly this reason.

Standards at a Glance

Standard	Scope	What qualification demonstrates
IEEE 837 (2014)	Permanent connections in substation grounding	Sequential survival of aging, corrosion, mechanical loading, then EMF testing at 65 kA and 75 kA
IEEE 80	Substation grounding safety design	Connection criteria supporting step and touch potential safety limits
AWS D15.2/D15.2M	Rails, rail components, rail bonds, thermite welding	Minimum acceptance for track-service joints under cyclic and thermal stress

What Passing These Tests Actually Proves (and What They Do Not)

Qualification proves performance stability, not just conductivity. A passing connection has shown that its resistance does not climb after corrosion exposure, that it does not loosen, and that it stays physically intact when fault-level magnetic forces hit. That is fundamentally different from a torque reading, which is a snapshot of clamping force on the day it was taken.

What standards cannot do is install the weld correctly. Field acceptance still rests on trained operators, mold wear limits, preheat discipline, and visual reject criteria such as cracks, porosity, incomplete fill, or evidence of blowout. The broader lesson applies across disciplines: whether someone is reading aerospace welding codes or rail standards, qualification tells you the process can work, while inspection confirms that this particular joint did work. Document both, and the connection you sign off on today remains defensible years later when a fault event puts it to the test.

Benefits: Why Engineers Specify Exothermic Connections

Ask a substation engineer why a grounding design calls for exothermic welds instead of bolted lugs, and you rarely hear a romantic answer about fire and graphite molds. You hear about resistance readings, fault current, and a maintenance budget that keeps shrinking. Exothermic connections earn their place on drawings because they solve the exact problems that mechanical connectors never fully escape: loosening, corrosion, and resistance that creeps upward year after year.

A Molecular Bond That Cannot Come Loose

The defining advantage of a CAD weld is that it is not a clamp. A bolted or crimped joint holds conductors together with external pressure. An exothermic weld fuses them into a single body of copper. As Harger’s installation guidance states plainly, a molecular bond formed by exothermic welding will not loosen and shows no increase in resistance over time. There is no torque spec to lose, no bolt to back off under thermal cycling, no creep in a spring washer. Once made, the joint simply stays the same, decade after decade.

That permanence matters most where nobody ever visits. Buried grounding grids, cathodic protection leads, and tower footings are not convenient inspection points. A connection you cannot re-torque is only acceptable if it never needs re-torquing, and that is precisely what a fused copper joint delivers.

Current-Carrying Capacity Equal to the Conductor

Because the weld metal is the same copper as the conductor, the connection’s current-carrying capacity equals or exceeds that of the conductor being welded. The joint is not a weak link or a bottleneck; in practice, the conductor itself will fail before the weld does. When a fault dumps thousands of amperes through a grounding system, current takes the path of least resistance, and an exothermic joint offers no more resistance than the cable it joins.

This is a meaningful contrast with mechanical connectors, where interface resistance depends on contact pressure, surface condition, and how well the joint was maintained. If you are weighing connection methods across a grounding project, our advantages of stick welding guide explores a similar principle in structural work: fusion beats pressure whenever the joint has to be trusted without supervision.

Performance Under Fault Current and Independent Qualification

The industry’s proving ground for substation grounding is IEEE 837, which qualifies permanent connections through a punishing sequence: high-current fault simulation, thermal cycling, and the electromagnetic force (EMF) test that tries to shake the connection apart. Independent laboratory certification has shown that Cadweld connections consistently exceeded IEEE 837-2014 requirements, including the EMF test that mechanical joints often struggle to survive without loosening.

Corrosion Resistance and Longevity

Galvanic and crevice corrosion attack mechanical joints relentlessly, especially where copper meets aluminum or where moisture sits inside a clamp interface. A fused joint has no interface at all. The weld fills the conductor strands into one solid mass, eliminating the gaps where water and soil chemistry do their worst work. Field experience, and the nVent railway connection literature, reflect a common result: exothermic connections frequently outlive the conductors they join. When a cable finally reaches end of life, the bond itself is often still sound.

Lower Tooling Burden Than Mechanical Connections

There is a practical, logistical benefit that shows up in the van and on the daily report. Exothermic kits weigh less and take up less space than the tools and materials needed for mechanical connections, according to nVent. No hydraulic crimpers, no calibrated torque wrenches, no assortment of lugs, bolts, and anti-oxidant compounds. A crew climbing a transmission tower or hiking to a remote cathodic protection station carries a mold, a clamp, a small torch, and the charges. That portability also simplifies training and logistics, since the same mold kit replaces a truckload of mechanical fitting options.

It is worth noting that surface preparation discipline still applies. Exothermic welding is a copper-to-copper fusion process, so it does not replace aluminum-specific techniques such as those covered in our guide to aluminum cleaner for welding, where oxide removal is a separate, deliberate step with its own chemistry.

The Engineering Bottom Line

Engineers specify exothermic connections because they convert a maintenance liability into a one-time decision. The joint is as conductive as the conductor, immune to loosening, resistant to corrosion, qualified against fault-current standards, and simpler to carry into remote terrain than the mechanical alternative. Against those strengths, the trade-offs around permanence and training are known, documented, and manageable, which is exactly what a good specification is built on.

Limitations, Hazards, and Safety Requirements

Everything that makes an exothermic connection attractive is also what makes it unforgiving. The same fire that creates a molecular bond releases a violent burst of heat, light, and fume in open air, and once that reaction runs, nothing about the joint can be undone. A fair assessment of the CAD welding process has to account for both sides of that coin.

Heat, Flash, and Fume

The reaction inside the crucible exceeds 2500°F (1400°C) and produces a localized burst of smoke the moment the charge ignites. Operators are trained never to look directly at the ignition flash, and to stand upwind and back away as the reaction runs. Molten copper can spatter if the mold is wet, cracked, or clamped loosely, so face shields, safety glasses with side shields, heat-resistant gloves, long sleeves, and leather footwear are standard kit, not optional extras.

OSHA groups the primary dangers of welding, cutting, and brazing operations as burns, eye damage, electrical shock, and cuts, and separately calls out health hazards from metal fume exposure and ultraviolet radiation, noting that both are controllable through work practices and personal protective equipment. It is worth understanding a nuance here: welding smoke is a mixture of fine particles and gases, and regulators set exposure limits on individual components rather than on total welding smoke. That means the local exhaust and respiratory decisions on your job are driven by what is actually in the charge and the conductors being joined.

Moisture, Mold Wear, and Blowout

Water is the single most common cause of a bad or dangerous weld. Residual moisture in a mold turns to steam instantly at reaction temperature and can crack the graphite or blow molten metal out through the clamp seam. This is why preheating with a torch, sometimes multiple times in humid weather, is non-negotiable; a solid grounding in torch handling and heat safety pays off directly here. Graphite molds also wear. Conducter holes elongate, the tap hole erodes, and repeated thermal cycling degrades the block. Manufacturer instructions specify a service life in welds, and exceeding it is how leaks and misshapen joints happen. Incompatible pairings are equally risky: molds and welding material from different manufacturers are not designed to work together and should never be mixed.

The Joint Never Comes Out

Unlike a bolted or crimped lug, an exothermic weld cannot be unmade. If you guessed the wrong part number, positioned the conductor incorrectly, or the connection later needs to be relocated, the options are cutting the conductor out and losing material, or sleeving a new splice on. That permanence is a genuine design limitation in retrofit work, occupied facilities, and anywhere future disassembly is plausible.

Trained Operators and Documented Controls

Exothermic welding is not a task a general laborer can be handed on day one. Most owners require operators to complete manufacturer certification training covering part selection, mold assembly, preheat discipline, ignition, and post-weld inspection, which is why structured programs such as an aluminium welding course are useful foundations before someone attempts field bonding. Site rules normally add a hot-work permit, combustible clearance, ventilation or fume extraction, and documented PPE issuance; consumable and cylinder storage guidance often comes from the same supplier that provides your welding gas and safety equipment.

Hazard	Typical Control
Flash and UV exposure	Face shield and shaded eye protection; turn away at ignition
Molten metal spatter	Dry molds, correct clamp torque, heat-resistant PPE
Fume inhalation	Work upwind or outdoors, ventilation, respiratory protection
Fire to nearby materials	Clear drop zone, hot-work permit, extinguisher on hand
Blowout from wet molds	Thorough preheat, storage in dry sealed containers

None of these drawbacks outweigh the performance case for exothermic connections in grounding and bonding, but each one is a real cost: training time, inspection discipline, and permanent geometry. Respect them during planning, and the field work becomes routine rather than risky.

Common CAD Welding Mistakes and How to Troubleshoot Them

Exothermic welding is simple, but it is not forgiving. The reaction almost never fails; the preparation around it does. Nearly every rejected CAD weld traces back to a wet mold, sloppy fit-up, an unlocked clamp, or a part number that does not match the conductors. All are visible before you strike the arc, so a thirty-second inspection saves a cut-out and a re-weld.

Wet or Contaminated Molds

Graphite is porous and quietly absorbs humidity from the air and from storage. Trapped moisture turns to steam the instant the reaction starts, cracking the mold, blowing molten metal out of the parting line, or leaving a porous, oxidized weld. Preheat every mold with a clean torch flame until all visible moisture is gone, and never skip it because the schedule is tight. Contamination works the same way: graphite dust, oil, sawdust, or slag in the cavity contaminates the pour, so blow out the cavity and wipe the conductors clean before assembly.

Worn Conductor Holes and Weld Leakage

This is the single most common cause of rejected connections. Check every conductor hole for wear, erosion, or cracks before each use, and confirm each conductor fits snugly against the hole walls. A gap of even a millimeter gives molten copper a path out of the mold, and once it runs it follows the insulation to the ground below. Harger’s installer guidance calls worn or cracked molds a leading cause of weld leakage and rejected connections, which is why molds carry rated uses rather than an open-ended service life.

Loose Clamps and Unlocked Handles

The mold halves must meet with real force. If the handle does not close tightly or lock into its detent, the parting line opens under the pressure of the pour and weld metal escapes. Check the clamp hardware, the handle spring, and the alignment pins, and replace the handle assembly when it has lost its grip rather than compensating by squeezing harder.

Non-Vertical Molds and Under-Filled Conductors

The system is gravity-fed: powder sits on top, the reaction drops through the tap hole, and metal falls into the weld cavity. Tilt the mold and the charge segregates, the pour short-circuits, and you get an incomplete joint, so assemble in true vertical orientation, clamped so it cannot shift. Under-filled conductor holes starve the joint as well: fill the cavity fully with conductor and leave no large voids, because a joint that is mostly poured copper rather than fused conductor is mechanically weak.

Wrong Part Number

Every mold and every welding material charge is built for a specific combination of conductor sizes and types. A mold specced for two 4/0 bare copper cables will not seal a 1/0 to threaded rod connection. Confirm the part number against the actual conductors on the truck, not the drawing, and never mix molds and charges from different manufacturers, since cavity volumes and charge weights are matched as a set.

Skipped Preheat and Cold Joints

Skipping preheat is the classic shortcut. Beyond moisture, a cold mold chills the molten metal before it fully fuses, producing a dull, gray, porous bead with visible shrinkage cracks. That is a cold joint, and it fails both visually and electrically. There is no rework: cut it out, discard the damaged conductor section, and start over at a fresh mold position. Our air acetylene welding guide covers torch handling, flame control, and safety that transfer directly to preheating.

Reusing Molds Beyond Their Service Life

Graphite erodes a little with every pour. As cavity walls thin and hole edges round over, leakage risk climbs sharply. Track use counts per mold, inspect conductor holes under light before each use, and retire molds at the manufacturer’s limit even if they still look serviceable. A replacement mold costs far less than a failed connection on an energized site and keeps you aligned with inspection expectations, including the qualification testing described in nVent ERICO’s CADWELD documentation.

Failure Mode	What You See	Corrective Action
Wet mold	Cracks, blowout, smoke at parting line	Preheat fully; store molds dry
Worn conductor hole	Metal run-out on conductor	Replace mold; verify snug fit
Loose clamp	Flash along parting line	Replace handle; lock before ignition
Non-vertical mold	Uneven, incomplete fill	Re-clamp truly vertical
Skipped preheat	Dull, porous, cracked bead	Cut out and re-weld
Wrong part number	Poor fit, gaps, leakage	Match part number to conductors

Because every weld releases molten metal above 2,500°F and a burst of fume, troubleshooting has a safety side: clear combustibles, wear proper eye and body protection, and follow OSHA’s welding, cutting, and brazing guidance. Fix the process before you ignite it, and most rework disappears.

CAD Welding vs. Crimping, Bolting, and Pressure Connectors

Choosing between an exothermic weld and a mechanical connector is a choice between two joining philosophies. One creates a permanent molecular bond that will outlive the conductor. The other creates a serviceable joint you can open, inspect, and take apart again. Both belong in a well-run grounding or bonding program, and knowing where each fits is the difference between a connection that lasts forty years and one that gets re-torqued every maintenance cycle.

Side-by-Side Comparison

The table below compares CAD welding against the three common mechanical alternatives across the criteria that matter most on a job site.

Criterion	Exothermic (CAD) Weld	Crimped, Bolted & Pressure Connectors
Resistance stability	Equal to or lower than the conductor itself; stable for life	Can rise over time as the joint relaxes, frets, or corrodes
Maintenance	None once accepted; nothing to re-torque	Periodic re-torque or hardware replacement
Installation speed	About 30 seconds of reaction time plus preheat and prep per weld	Fast per joint; crimping is quickest, bolting slower on large conductors
Inspectability	Visual check of fill and surface; no hidden interface	Torque marks and crimp indentations checked, but the contact interface stays hidden
Reworkability	Not removable; cutting the conductor is the only way back	Fully removable, adjustable, and reusable
Tooling vs. per-joint cost	Cheap tooling, moderate per-joint consumable cost	Expensive crimp or torque tooling, low per-joint cost
Standards qualification	Qualified to IEEE 837-2014 sequential testing	Qualified through manufacturer listings and application-specific testing
Best suited to	Permanent grounding, lightning, cathodic protection, rail bonding	Removable joints, retrofits, tight spaces, future disassembly

Where Exothermic Welding Has the Clear Edge

The core argument for a CAD weld is that there is no mechanical interface to fail. Mechanical pressure connections can loosen and corrode over time, which is why maintenance schedules exist for them in the first place. As nVent’s comparison of connection types points out, a bolted or clamped joint depends on continuous contact pressure that vibration, thermal cycling, and corrosion slowly erode. An exothermic joint has none of those failure modes because the two conductors have literally become one piece of metal. Resistance does not creep upward, and a fault current sees a full cross-section of copper or copper alloy rather than a constrained contact patch.

That also means no ongoing labor. There is no torque log to keep, no annual inspection for stretched hardware, and no galvanic washer to replace. On a substation grid or a buried cathodic protection circuit that nobody wants to open again for decades, that is a decisive advantage.

Surface prep follows the same discipline as any other welding trade. Conductors and molds must be cleaned to bright, dry metal before the reaction runs, which is why the moisture-removal preheat step is non-negotiable. The same clean-metal mindset covered in Aluminum Cleaner for Welding: How to Prep Aluminum Right applies here, only the contaminant being chased is water rather than oxide.

When Mechanical Connectors Are the Better Choice

Exothermic welding is permanent, and permanence is not always a feature. Crimped and bolted joints remain preferable whenever future disassembly or relocation is expected. Specifically, reach for a mechanical connector when:

– The joint must come apart later — equipment upgrades, temporary bonds, or circuits planned for redesign.
– Space is tight — inside a crowded pull box, switchgear cabinet, or meter pan where a graphite mold and clamp physically will not fit.
– The area is energized or occupied — an open flame near live busbars, gas lines, or fuel storage is simply not acceptable, so mechanical or arc-based joining methods applied under controlled conditions take priority.
– The conductor is aluminum and the specification calls for a listed compression fitting with approved anti-oxidant compound rather than a copper reaction weld.
– The application demands a listed, field-adjustable fitting that an inspector can verify by torque value alone.

Standards and Qualification: What Actually Proves Performance

On paper, both families can be qualified, but the tests differ in severity. Substation grounding connections are commonly specified against IEEE 837, a sequential qualification test that runs thermal cycling, fault current, corrosion exposure, and mechanical shock on the same sample set. Products such as TE’s AMP-Weld exothermic connectors are tested and qualified to IEEE 837-2014, giving engineers a documented basis for specification rather than a manufacturer’s claim. Railroad applications lean on AWS D15.2 for rail and rail bond joints, while mechanical connectors are typically evaluated through product listings and torque or pull-out data.

The practical takeaway is simple. Specify exothermic welding where the joint is permanent, buried, or hard to reach, and where rising resistance would be dangerous. Specify crimped, bolted, or pressure connectors where service access, relocation, or future disassembly matters more than lifetime stability. On most real projects, you will use both.

Practical Tips, Training, and Final Takeaways

You have followed the chemistry, the field steps, the standards, and the failure modes. Now it comes down to execution. Exothermic welding rewards crews who plan the job before they strike the first spark, so use this closing checklist on your next grounding project.

The Pre-Job Checklist

• Select the part number first. Match the mold to the exact conductor combination and size, not to the nearest convenient number. A 4/0 copper to half-inch ground rod takes a different part than two 4/0 conductors, and guessing here is the single most common source of leaks and under-filled joints.
• Inspect everything before you leave the shop. Check mold cavities for wear, verify clamp condition, confirm conductor ends are cut square, and confirm you have the correct welding material for the mold manufacturer. Never mix components from different brands.
• Store welding materials dry and in date. Starting powder that has absorbed moisture is a frequent cause of ignition failure. Keep containers sealed, off damp concrete, and within the shelf life printed on the package.
• Clean conductors to bright metal. Oxidized or grease-coated ends weld poorly, no matter how good your mold is.

Weather, Orientation, and Ignition Discipline

Keep the mold vertical. The entire system is gravity-fed: the steel disk holds the charge until it melts, and the molten copper then drops into the weld cavity. Tilt the mold and you interrupt that sequence. Wind matters too. Loose starting powder can blow out of the crucible before ignition, which is one reason electronically ignited, self-contained systems have gained ground on exposed sites. Plan a weather contingency: if you cannot preheat and hold a dry, vertical mold safely, reschedule rather than force the weld.

Post-Weld Acceptance and Documentation

After the mold is removed and the joint is brushed, inspect against standard visual criteria: full fill around every conductor, no visible pores or unfused edges, no carbon inclusions, and a smooth flow line where the weld metal entered the cavity. Reject and cut out anything that fails. Photograph each connection, log the part number, conductor sizes, date, operator, and mold serial number, and file it with the project record. That paperwork matters when an inspector or utility owner asks how a specific bond was made months later.

Then measure the system, not just the joint. Fall-of-potential ground resistance testing confirms the installation does what it was designed to do, and some jurisdictions require grounding resistance not exceeding 5 ohms as common acceptance practice.

Training and Certification

Exothermic welding looks simple from a distance and punishes untrained hands up close. Manufacturers such as Harger offer hands-on certification training covering tools, inspection, safety, and troubleshooting for exothermic welding. A proper program teaches mold selection, preheat, ignition, visual acceptance, and the corrective actions covered earlier in this guide. If your crew is new to the process, invest in that training before the first production weld, and pair new operators with experienced ones until their welds pass consistently.

While exothermic work is its own discipline, broader Aluminium Welding Course: What You Learn, Costs & Careers material also helps crews understand how metallurgy, joint prep, and inspection habits carry across processes.

When CAD Welding Is the Right Choice

Specify exothermic connections when you need a permanent, low-resistance bond that will not loosen, corrode over time, or need periodic torque checks, especially in substation grounding, lightning protection, cathodic protection, rail bonding, and telecom ground systems. It works on copper, steel rail, copper-clad steel, brass, stainless steel, cast iron, and galvanized conductors. Choose mechanical connectors when the joint must be removable, when you are working in a cramped or energized space, or when future disassembly is part of the design.

Frequently Asked Questions

What does CAD welding stand for?

CAD welding refers to CADWELD, a brand name for exothermic welding invented for making permanent electrical connections. The term became generic, so tradespeople now use ‘CAD weld’ to describe any thermite-based copper-to-copper or copper-to-steel molecular bond made in a graphite mold.

How hot does a CAD weld get?

The exothermic reaction exceeds 2500°F (1400°C), producing molten copper that flows into the weld cavity. Because the reaction is self-sustaining, no external power or heat source is required beyond the initial ignition.

Is an exothermic weld stronger than the conductor itself?

Yes. The molecular bond has a current-carrying capacity equal to or greater than the conductor being joined, and it will last longer than the conductor. It cannot loosen because there is no mechanical interface to relax.

How long does a CAD weld take to complete?

The reaction itself completes in seconds. After ignition you should wait about 30 seconds for the mold to cool before removing it, so a complete joint typically takes a few minutes including preparation, cleaning, and preheating.

Can you make a CAD weld in the rain or on wet equipment?

No. Moisture is the most common cause of failed welds and can cause steam blowouts of molten metal. Conductors and molds must be thoroughly dried with a propane torch, and work should be stopped in wet conditions.

Which standard covers exothermic grounding connections?

IEEE 837 (current revision 2014) is the standard for qualifying permanent connections used in substation grounding, including sequential and electromagnetic force testing. Railroad rail bonds fall under AWS D15.2/D15.2M.

Is CAD welding a permanent connection?

Yes. Once made, an exothermic weld cannot be separated without cutting the conductor. If future disconnection is required, use a bolted or pressure connector instead.

Do CAD welds need maintenance or re-torquing?

No. Unlike mechanical lugs, exothermic connections require no periodic re-torque and show no increase in resistance over time, which is why they are specified for grounding grids and rail bonds that must survive decades of fault current.