How to Weld Brass: TIG and MIG Techniques, Filler Metals, and Zinc Fume Safety

A restorer once brought us a set of brass architectural fittings — the kind with a century of polish on them and owners who care about sixteenths of an inch. One had cracked along a seam. The obvious move was to fuse it: same metal, same colour, invisible repair. The second obvious move was to put an argon torch on a bench whose extraction had been working fine for steel for years.

That second move is where people get hurt, and it is why this article starts with the safety section before the technique. Brass is copper-zinc, and zinc boils at about 907 °C — below the melting range of the brass itself. Every time you melt brass, you boil zinc out of it: white zinc-oxide fume into the air you are standing in, and gas into the weld pool that comes back as porosity. Welding brass is not really a technique problem. It is a fume-and-porosity problem you are choosing to take on, with a joinery decision that should have been made before the torch came out.

Most brass should not be fusion welded at all. It should be brazed. The exceptions are real and worth knowing, and when you are inside one of them, the technique matters down to the arc length. Here is the full picture — alloy by alloy, process by process, with the health data that comes with the territory.

Short answer: Brass is copper-zinc; zinc boils (~907 °C) below brass’s melting range (915–955 °C for cartridge brass), so fusion welding boils zinc out — causing porosity, lost ductility, colour change, and zinc-oxide fume. Most brass should be brazed (silver BAg alloys, RBCuZn-A, or BCuP with flux), not welded. Fusion welding is for low-zinc alloys, appearance-matching repairs, or joints exceeding braze strength — using TIG (DCEN argon, thin sections ≤10 mm; MIG above), short arc, ERCuSi-A or ERCuSn-A filler, clean joint, and local exhaust ventilation. Zinc-oxide fume causes metal fume fever 4–8 hours after exposure; OSHA PEL is 5 mg/m³ zinc oxide fume, copper fume 0.1 mg/m³.

What follows is organised the way the job itself unfolds. We start with identification, move to the reason brass behaves the way it does, then the decision between welding and brazing, then the health data that has to be in place before anything is lit, and only then the processes: filler metals, TIG, MIG, oxy-fuel, soldering, joint preparation, distortion, a full worked repair, starting parameters, troubleshooting, the myths that cost parts, storage, and the questions readers ask most.

Know Your Alloy

Brass is a family, not a formula, and composition ranges matter more here than on almost any other material the shop sees — because zinc content drives everything that follows. Zinc decides how much fume you make, how porous the bead gets, whether fusion welding is even a sensible question, and how far you should have been from the torch in the first place.

Start with the UNS number, or with the best identification you can make. The five groups you will meet most often look like this:

Read the third column as a risk ranking. Cartridge brass at roughly 30 percent zinc is the everyday case: weldable, but only with the fume and porosity controls this article describes. Muntz and naval brass sit near 40 percent zinc, which is where the thermodynamics stops being negotiable and the material’s own data sheets start steering you back towards brazing. The low-zinc group — red brass and similar alloys at about 15 percent zinc or less — is the category that welding engineers actually mean when they say a brass is weldable.

Two clarifications are worth stating because they are constantly blurred. Bronze is copper-tin, not copper-zinc — bronze and brass are different materials with different welding behaviour, and half the advice given for one does not transfer to the other. And leaded or free-cutting brasses should not be fusion welded: lead does not dissolve into the weld pool in a useful way, it promotes cracking, and the fume adds its own problem to the zinc one. If the chip curls off yellow and greasy, or the part is stamped “L” or “Pb,” the arc question is closed.

Check the data sheet before you light the torch

A published C46400 naval brass data sheet rates brazing as Excellent, soldering as Excellent, and gas-shielded arc welding as Fair. That one line is the industry consensus without anyone needing to explain it: when a material’s own data sheet rates its own weldability as Fair, the answer is usually to find another joint — not a hotter setting.

This habit scales. Before any brass job, find the alloy designation, pull the supplier data sheet, and read the joinability column first. A sheet that says “Excellent” for brazing and “Fair” for arc welding is not giving you a tuning range; it is telling you where the material wants to go. Our companion piece on whether you can TIG weld brass works through the same question from the process side.

Why Zinc Is the Whole Problem

Zinc boils at approximately 907 °C. Cartridge brass melts between 915 and 955 °C. The zinc reaches its boiling point before — or right as — the brass melts, so there is no practical way to melt brass without vaporizing a meaningful quantity of zinc. This is not a minor side reaction that better technique eliminates; it is thermodynamics.

The vapour-pressure numbers show the scale. At around 1,080 °C, zinc vapour pressure is approximately 39,500 pascals while copper’s is about 0.02 pascals — a partial-pressure ratio greater than five hundred thousand to one in favour of zinc leaving. At welding temperature, zinc is the most volatile thing in the pot by an enormous margin, and copper essentially stays home.

That is why brass fumes white. The zinc leaves the melt, reacts with oxygen in the air around the arc, and condenses as fine zinc-oxide particles — the same white smoke that gives brass foundry work its historic nicknames. It is also why the weld pool gasifies: the same vapour that escapes upward also gets trapped as the pool solidifies, and comes back as porosity.

Three consequences

• Porosity. TWI describes the core problem directly: weld-metal porosity caused by zinc boiling off during melting, with volatilization producing white fumes of zinc oxide. Gas evolves during solidification and gets trapped. Research on thin brass TIG work confirms that the weld metal becomes porous and that the brass loses the physical and chemical properties it normally possesses.
• Loss of ductility and properties. The same zinc depletion that causes porosity means the weld metal is not brass in the way the parent metal is — it is closer to copper with porosity in it. Empirically this shows up as a more brittle, less ductile joint, which is exactly what a repair of a cracked fitting cannot afford to be.
• Visible colour change. The parent metal is yellow; the weld is not. Tin-bronze fillers ERCuSn-A and ERCuSn-C are recommended partly for this reason — they blend with the yellow of many brasses far better than silicon bronze does.

Notice that all three are the same phenomenon measured in different ways: zinc leaving. Less zinc in the pool means more gas during solidification, a weld metal that is compositionally no longer the parent alloy, and a colour shift because the copper left behind is redder than the brass it came from.

Fume scales with voltage

One detail links the welding parameters directly to the safety problem: fume generation increases with arc voltage. Gray and Hewitt’s work in the 1960s found that changing voltage across a 10-to-50-volt range changed the fume rate by roughly a factor of five. Same wire, same settings, slightly longer arc — dramatically more zinc fume.

This is the technical reason the short arc is not just a style preference on brass. It is an exposure control. The technique section and the ventilation section are the same section, seen from two sides. It also explains why “I only ran a few passes” is a weaker defence than it sounds: those passes were run at whatever arc length the operator found comfortable, and a long, wandering arc on brass multiplies the fume that a short one would have produced from the identical joint.

Shop note: Arc length is a dose setting. On brass there is no arc length that is merely a style choice — the moment you stretch the arc you raise voltage, voltage raises fume rate, and the fume lands in the breathing zone of whoever is standing there.

The Honest Answer: When to Weld, When to Braze

TWI’s position is that brazing with an oxy-acetylene torch and a copper-silver filler is the viable alternative for most brass work, and that only low-zinc brasses are weldable using fusion welding. Its naval brass guidance goes further: MMA is not recommended for naval brass or any other zinc-bearing copper alloy. Those two sentences, from the welding research body that wrote the joinery textbooks, are the spine of every decision in this article.

In practice the choice is a checklist. Fusion welding earns its place when several of the conditions on the left column are true at once; brazing wins almost every time the right column is what the job actually looks like.

Read the pairs vertically rather than across. “The part is thin” sits on the braze side because heat control on thin brass is brutally unforgiving and braze input is lower and more diffuse. “The specification calls for it” sits on the weld side because a drawing outranks your preference — but a drawing that calls for fusion welding on C46400 is a drawing someone should query before the shop absorbs the risk.

What brazing actually involves

AWS defines brazing as a joining process where the filler metal liquidus is above 840 °F (450 °C) and below the solidus of the base metal — that lower boundary is what separates brazing from welding, and it is also why brazing solves the brass problem: the brass never melts, so the zinc never boils.

Silver brazing alloys under AWS A5.8 cover the usual range. The common classifications and their melting windows look like this:

Flux is borate or borax based — AWS FB3 class — and it has to be rated for brass rather than assumed from copper practice. One caution worth repeating because it surprises people: Johnson Matthey’s Sil-fos products require a separate flux on brass, because the self-fluxing behaviour works only on copper.

Brazing is genuinely a good process for brass, not a consolation prize. The joint is ductile, the temperature is low enough to protect the rest of the part — critical on anything with polish, plating, or temper — and you are not fighting vapor evolution at all. The restorer’s cracked fitting from the opening got brazed. It is still invisible.

Safety: Zinc Oxide Fume and Metal Fume Fever

This section is the most important thing on the page. Everything else is technique; this is health. Read it before you read the amperage tables, and if you weld brass only occasionally, read it again before the next job — because the exposure pattern is what makes this hazard quietly persistent.

Metal fume fever

Metal fume fever is the acute flu-like illness caused by breathing metal oxide fume. It has more names than most occupational illnesses — copper fever, brazier’s disease, welder’s ague, foundry fever, brass chills, spelter shakes, and the Monday morning syndrome — which tells you how well known it is among the people who get it. NIOSH’s own account of welding fume estimates that 30 to 40 percent of all welders have experienced it at some point.

The Monday name is not a joke. Symptoms typically appear four to eight hours after exposure (reported range: three to twelve), so you finish a shift feeling fine and wake up the next morning with a fever. Symptoms include fever that can reach 40 °C, chills and rigors, muscle aches, headache, nausea, a dry throat, a dry non-productive cough, chest tightness, shortness of breath, and a metallic taste in the mouth. The NIOSH Pocket Guide lists exactly this symptom cluster against zinc oxide, which is one reason the entry is worth bookmarking.

It is self-limiting: a typical episode lasts one to three hours, severe cases 24 to 48, and it is not usually dangerous in a healthy person. What makes it worth taking seriously is the repeat-exposure pattern. Tachyphylaxis means the body builds tolerance — the fever after the fifth exposure is blander than after the first. Welders get the mild version, feel like they got away with it, and go back to the same exposure. That is how an afternoon’s illness becomes a chronic respiratory problem.

The exposure limits, correctly stated

Getting these right matters, because one common figure circulating is wrong. Here are the numbers with their sources attached:

The confusion to defuse is the ceiling. People quote “15 mg/m³” as the zinc oxide limit because they have seen it in a table — but that figure belongs to zinc oxide dust, not to the fume welding produces (OSHA carries it as an eight-hour total-dust average, NIOSH as a dust ceiling). For the fume, OSHA’s PEL is 5 mg/m³ as an eight-hour TWA, and the NIOSH Pocket Guide’s zinc oxide entry carries the same 5 mg/m³ TWA for fume with a 10 mg/m³ short-term limit.

Across all welding: OSHA has no enforceable general PEL for welding fume as a single substance — the 5 mg/m³ figure from a 1989 rule was remanded and never took effect, a history NIOSH documents on its welding fume PEL page, which also notes that fumes from copper and other metals “are considerably more toxic and shall be kept at or below the levels required by their respective PELs.” Compliance runs through constituent limits. NIOSH goes further: its Pocket Guide lists welding fumes with no OSHA PEL and a NIOSH REL of “Ca” — a potential occupational carcinogen — and NIOSH recommends reducing exposure to the lowest feasible concentration, noting that welders can be harmed even when individual constituents sit below their PELs. The NIOSH Pocket Guide entry for welding fumes is the one-page version of that position.

One practical consequence: on brass, the copper-fume limit of 0.1 mg/m³ is fifty times tighter than the zinc-oxide limit of 5 mg/m³. It is usually the number you approach first, which surprises people who have spent the whole job thinking about zinc.

Controls

• Local exhaust ventilation at the source — the primary control, not an optional extra. Hoods or extractor guns positioned to capture at the arc, with capture velocity around 0.5 m/s (100 feet per minute). OSHA’s welding, cutting and brazing hazards guidance frames metal fume exposure as a hazard to be controlled with work practices and engineering controls rather than tolerated.
• Respirators where local exhaust is not feasible, and only as part of a written respiratory protection program with fit testing and medical clearance under OSHA’s respiratory protection standard, 29 CFR 1910.134. A respirator is the second line, never the first.
• Short arcs — fume rate scales hard with voltage; the arc length is a dose-control decision.
• Clean the work area. Metal fume deposits on surfaces and gets re-suspended by ordinary activity — the fume you did not inhale at the arc is still on the bench for the next person.
• Constituent limits, not a fume average. A single “total welding fume” number is not a compliance strategy. Sample for what is in this fume — on brass that means zinc and copper — and manage to those limits.

Worth repeating because it is counter-intuitive: ventilation is not a courtesy, it is part of the procedure. If the extraction arm is not running before the arc strikes, the weld should not start. Our guide to MIG welding gas settings covers shielding-gas discipline, which is the adjacent habit: gas that does its job at the arc keeps the fume where the extraction can catch it.

If someone develops symptoms

Because the onset is delayed, the affected person rarely connects the illness to the weld — that is what makes the pattern repeat. The response is practical:

1. Remove the person from exposure immediately and get them into fresh air.
2. Rest and fluids; the episode is typically self-limiting within hours.
3. Seek medical attention if breathing difficulty, chest pain, or a fever that does not settle develops.
4. Report the episode and treat it as data: recurrent episodes are a ventilation failure, not bad luck, and the fix belongs to the extraction system rather than to the welder’s constitution.

Anyone welding brass regularly should know that the fever they feel on Saturday morning is a diagnosis of Friday’s ventilation, and say so — because the tolerance that develops will hide it from them personally while the damage accumulates quietly. NIOSH’s broader welding fumes and manganese material makes the same point about cumulative exposure: effects show up in workers whose individual readings looked acceptable, which is another argument for treating ventilation as a permanent fixture rather than a per-job concession.

Filler Metals for Brass

Filler choice on brass is where the “brass is brass” assumption does the most damage. The wrong rod gives you a joint that cracks, wets badly, or looks wrong — and it does so for reasons that have nothing to do with your hand. Start from the specification, not from the rack.

For fusion welding

• ERCuSi-A (AWS A5.7, ~3% silicon copper, Cu6560 / EN ISO 24373 SCu 6560) — the general-purpose choice for brass, used for both welding and brazing, and what TWI and most suppliers reach for first. The 3% silicon alloy flows well and tolerates the heat cycle better than a composition-matching brass filler would.
• ERCuSn-A and ERCuSn-C — tin bronzes. Recommended when colour match matters, since they blend with the yellow of many brasses far better than silicon bronze does. Tin-bronze weld metal is more sluggish than silicon bronze, which also means a wider joint angle and a slower, more deliberate pool.
• ECuAl-A2 (Cu6180) — the correct choice for welding manganese bronze castings. Worth noting because manganese bronze is frequently handed out as a general brass filler, and it is not one.

The underlying rule, from TWI’s copper-alloy guidance, is that MIG and TIG fillers for brass should not contain zinc at all: zinc in the wire vaporises in the arc, adding to the zinc already leaving the parent metal. That is why the practical filler list for brass is copper-silicon and copper-tin — fillers that bring the elements you want to keep and none of the one you are trying to manage.

The AWS A5.7 specification is the source document for these classifications; a supplier labelling a rod only as “brass rod” is telling you nothing useful.

For brazing

• RBCuZn-A — sometimes called naval bronze, roughly 60% copper, the correct brazing rod for brass fittings and brass itself. Washington Alloy describes it as primarily used for acetylene welding of brass.
• BAg silver alloys — the general brazing choice; BAg-7 is the most widely used, with BAg-1, BAg-5 and BAg-20 covering faster-flowing and wider-melting-range needs.
• BCuP copper-phosphorus — used on brass with two real caveats: they require a separate flux on brass rather than being self-fluxing, and phosphorus reacts with iron or nickel to form brittle compounds, so they must never be used on steel or nickel alloys. High-phosphorus grades are also brittle and notch sensitive.

Put the two lists side by side and the selection table most shops end up making for themselves looks like this:

The manganese bronze mistake

Manganese bronze — RBCuZn-B, also called nickel bronze — is a ferrous braze-welding filler. Washington Alloy lists nickel bronze as primarily used to braze or oxyacetylene weld steel, cast iron, brass, and bronze. Reaching for it because the part is brass gets you a filler that was not designed as the best choice for brass, when RBCuZn-A or BAg was sitting right there on the same shelf.

The distinction is worth slowing down for. “It is on the brass rack” is not a selection criterion; “it is specified for this joint” is. Manganese bronze does appear on brass work — legitimately, on steel-to-brass transitions and on castings where a ferrous-compatible filler is wanted — but as a default rod for a brass fitting it is a compromise nobody was asked to make. The same logic applies everywhere in the shop: a filler that is right for one material family does not become right for another because the two look similar on the shelf. Our look at whether aluminium brazing rods work on steel is the same mistake wearing different clothes.

Cadmium in silver brazing alloys, and when it matters

Several common silver brazing alloys contain cadmium: BAg-1, BAg-1a, BAg-2, BAg-2a, and BAg-3 are cadmium-bearing, typically at 16 to 25 percent by weight. AWS A5.8 devotes a section to this — “Precautionary Information for Brazing Filler Metals Containing Cadmium.” Cadmium has been banned as a brazing and soldering filler in the European Union under REACH Regulation 494/2011 since December 2011.

For US work, the driving requirement is usually the application: cadmium-free alloys are required for food-handling equipment, surgical instruments, and potable water. Lucas-Milhaupt makes it explicit — Silvaloy 560 (a BAg-7 alloy) as a cadmium-free alloy for food handling, Silvaloy 630 for surgical and food handling. If the joint touches food or drinking water, specify cadmium-free from the start; substituting later is expensive. Note also that cadmium fume is its own hazard with its own limits, which is why the cadmium-bearing grades carry ventilation warnings on the tin. And a naming caution: the cadmium question and the thermite welding process often abbreviated “CAD” or Cadweld are entirely different problems with similar names — TWI’s explainer on CAD welding covers the other one.

TIG on Brass

TIG is the better fusion process for brass because it gives independent control over heat input — the variable that controls zinc loss. You choose the amperage, the arc length, the travel speed and the addition rate separately, and on brass every one of them is a lever on how much zinc leaves the pool. For a process-by-process primer before you start, our guide to TIG welding brass is the companion to this section.

Gas and polarity

TWI recommends 100 percent argon for sections under about 3 mm, and 75 percent helium with 25 percent argon for thicker sections where deeper penetration is needed — helium raises arc energy so the puddle gets enough heat without you stretching the arc to get it. Lincoln’s technical manual states that, as with steel, DC electrode negative polarity is employed with argon.

There is a genuine disagreement worth knowing about. Wieland — a major copper alloy supplier — recommends AC for gas-shielded welding of copper-zinc alloys, on the reasoning that heating of the melt is lower compared with direct current. Other shops use AC because the polarity reversal halves the effective heat and makes thin brass far more controllable. If you are welding thin brass and finding DC too hot to manage, AC is worth testing: it is a normal shop solution, not a fringe technique. The tungsten discipline matters either way — on non-ferrous metal, tungsten can dissolve into the puddle, and tungsten inclusions in brass cause cracking. Our TIG welding settings chart covers the grinding discipline that applies here too.

Thickness limit

TWI puts TIG on brass at a practical limit of about 10 mm, recommending MIG above that. Respect the line: TIG’s advantage is control, and control evaporates on thick sections that eat heat faster than the torch can manage it. Past 10 mm you are pouring heat into a high-conductivity sponge and boiling zinc the whole time, with none of the precision that justified choosing TIG in the first place.

Technique

• Keep the arc short. The single most important instruction. A long arc raises voltage, and voltage drives fume rate — you are managing exposure control with your hand position.
• Strike on the filler. Some shops establish the arc on the filler rod rather than the base metal, limiting how much parent brass heats before filler arrives. Trade practice rather than standard, but common among people who do this work regularly.
• Use a copper backing strip where practical: absorbs heat, supports the pool, controls distortion — three problems, one piece of copper.
• Pulse TIG for thin material. Research on thin brass TIG specifically examined pulsed welding for heat input control; it is the most effective way to keep the pool small enough to control on light sections. The pulse caps peak temperature cycle by cycle, which is exactly the control zinc vaporisation demands.

MIG on Brass

MIG on brass is viable but more demanding, with a narrower comfort zone than TIG. It is the process you move to when section thickness or production rate forces the issue — not the one you start with.

• Filler wires: ERCuSi-A or ECuAl-A2 with argon shielding. TWI notes that preheat reduces the current needed, which in turn reduces zinc loss — one of the few places preheat earns itself on high-zinc material.
• Pulse MIG is strongly preferred — a small, repeatable pool that prevents the burn-through steady droplet transfer produces on a material that is already fluid and conductive.
• Wire diameter matters more than on steel. 0.030 inch wire on thin brass deposits too much heat per unit length; go as small as your feeder handles reliably.
• Expect more spatter than on steel — silicon bronze wire is not as fluid as steel wire in argon. Check gas flow and stickout before you start chasing voltage.

Gas discipline decides whether this works. The usual choice is DCEN with 100 percent argon; some shops add small amounts of oxygen or helium, which change arc character and zinc behaviour, but whichever gas you use, keep it consistent across the job and set the regulator from a real specification rather than a guess — our page on argon pressure for MIG welding covers the flow-rate side of that.

If you are new to brass and have a TIG machine, use the TIG machine. MIG on brass is a production optimization, not a first choice. For the general contrast between the two processes — and where each one genuinely wins — see our MIG versus TIG breakdown.

Oxy-Fuel on Brass

It is done, and it has a history. A C46400 data sheet rates oxy-acetylene welding of naval brass as Good; RBCuZn-A is described as primarily used for acetylene welding of brass; TWI lists torch brazing as the viable alternative generally. Against that, Rapid Welding’s copper and brass guide states that gas welding of brass is not recommended because the zinc tends to vaporize.

Both positions are defensible and not really in conflict: a skilled operator with a fine flame and years of experience can gas-weld brass; a beginner cannot. The vaporization risk is identical either way — the difference is control, and control is exactly what a beginner lacks. A torch has no amperage dial and no meter, so heat input is entirely a judgement call made in real time over a metal that conducts heat away fast enough to make local readings misleading.

Use the torch primarily for brazing, where its lower temperature is an advantage rather than a liability. The flame can be brought up patiently, the whole area warmed evenly, and the filler introduced only when the base metal is ready — the technique that the worked repair further down this page depends on. Reserved for fusion work, the same torch becomes the hardest process to control on the material least forgiving of uncontrolled heat.

Soldering: The Third Option

Between welding and brazing sits soldering, and on brass it is used far more than either — for exactly the joints where the fusion question was never real. Soldering works below 450 °C, well under the point where zinc starts to care, so the parent metal never softens, never distorts, and never fumes. The trade is strength: a solder joint is a low-strength, low-temperature joint, and it behaves like one under load and under heat.

Where it belongs: instrument and decorative assemblies that will never see real load, thin-gauge work where even brazing input would warp the part, electrical and plumbing joints where solder is the specified connection, and repair of antique items where the goal is to preserve what is there rather than to add strength. Tin-lead and tin-silver solders with appropriate flux are the standard consumables; the flux choice follows the alloy and the contaminant load, and — as with brazing — brass wants a flux rated for brass rather than one assumed from copper practice.

Where it does not belong: anything structural, anything hot in service, and anything that will be machined after joining — solder does not machine like metal, it tears.

The decision sequence in practice is therefore three questions in order:

1. Does this joint need strength or temperature beyond solder?
2. Does it need to be a full-fusion weld for metallurgical reasons, or will a braze serve?
3. Only if both answers point upward do you light an arc on brass — and then the entire safety and technique apparatus of this article comes into play.

Choosing the Process: Side by Side

The five ways to join brass in one view, with the trade-offs that decide between them:

Read down the “zinc loss / fume” column and the decision makes itself for most joints: the two processes at the top never melt the brass, so the problem this article spends thousands of words managing simply does not exist there. Fusion welding earns its place in the bottom three rows only when the joint’s requirements — strength, temperature, appearance, procedure — reach past what brazing can deliver. If you are still weighing which fusion process to use when fusion is genuinely required, our MIG vs TIG vs stick comparison walks the trade-offs from the equipment side.

Four jobs, four answers

A cracked brass door fitting, visible surface. Brazed with a silver alloy, dressed back, polished. The repair is invisible because the filler polishes like the parent, the heat never reached the rest of the casting, and no fume was generated. Fusion welding here would add porosity risk to a cosmetic part for zero benefit.

A low-zinc red brass flange on a steam line, service temperature above braze limits. Fusion welded — TIG, ERCuSi-A, clean joint, extraction running. The alloy is low-zinc so zinc loss is manageable, the service condition rules out braze, and the joint will be inspected. This is the case the process exists for.

Naval brass seawater hardware. Brazed or soldered per its own data sheet, which rates both Excellent and arc welding Fair. Dezincification-resistant alloy, marine environment, and a weldability rating that argues against you — the joint gets the process the material recommends.

Thin-gauge brass instrument shell, 0.020 inch, needs a seam. Soldered if the load allows; TIG-brazed with silicon bronze if it needs more strength than solder gives; fusion TIG at 20–60 amps only if the specification demands a fusion weld. At that thickness, heat control decides the part’s flatness, and every process above fusion in the table controls heat better than fusion does.

Joint Preparation

Preparation is more important on brass than on most materials, and it is more often skipped. Andersen’s guidance is that edge preparation and cleaning will in most cases be the most important step; Lampert advises removing oxide by grinding or acid cleaner and getting the surface free of grease. This is not boilerplate: every millimetre of surface oxide that goes into the pool is copper and zinc oxides adding gas.

• Degrease and mechanically clean immediately before welding — not an hour before; brass re-oxidizes and re-contaminates. The moment between cleaning and welding is part of the procedure, not dead time.
• Joint angles differ by filler: TWI specifies 60° included angle for copper-silicon filler and 70° or more for copper-tin filler. The tin-bronze pool is more sluggish, so it needs more room to wet the sidewalls.
• Backing: a copper backing strip does two jobs at once — absorbing heat and controlling distortion.
• Fit-up: brass brazing lives or dies on capillary gap. A tight, even gap pulls filler; a wandering gap leaves dry spots that look wetted from the front and are not.

Preheat, honestly

Preheat guidance for brass genuinely varies, and any article giving you one number is oversimplifying: TWI suggests 100–300 °C (preheat allows lower current and therefore less zinc loss); TWI’s naval brass guidance says preheating is not usually necessary and allows up to 250 °C; Andersen suggests 100–350 °C; Rapid Welding advises limiting preheat below 80 °C to avoid structural change.

The pattern in that list is the useful part. Preheat exists to fix fusion problems, and fusion problems are mainly a low-zinc-brass issue — high-zinc brass does not need it. Welding naval brass? You probably do not need preheat. Welding low-zinc red brass and fighting lack of fusion? Preheat is the fix, at the low end of the ranges, with the current reduced to match. On every job: preheat at the lowest temperature that works, never as a substitute for cleaning, and never as a habit carried over from steel.

Distortion

Brass distorts badly, and predictably. Cartridge brass C26000 has a coefficient of thermal expansion of about 19.2–20.0 µm/m·K — roughly double steel’s — and thermal conductivity of 120–121 W/m·K. High expansion plus high conductivity is a difficult combination: the material wants to move a lot, and the conductivity spreads heat fast and unevenly, so the part warps rather than simply growing.

Control it the same way you control distortion anywhere: restrain it, use a backing bar, weld in short sequences, back-step. And consider brazing instead if distortion tolerance is tight — TIG brazing with silicon bronze is a lower-temperature method that reduces distortion, which is another argument for the braze route when the joint does not need full fusion. On brass every degree saved is a millimetre of flatness kept, and a fitting that came in polished and flat is the last thing you want to hand back with a twist in it.

A Practical Workflow

If you are welding brass in fusion mode, this order works — and the order matters, because steps one through four decide whether the weld succeeds before the arc ever strikes:

1. Identify the alloy. Get the UNS number and check its zinc content — high-zinc or low-zinc.
2. Decide whether fusion is even right. Above about 30 percent zinc, ask whether brazing serves the joint better.
3. Degrease and mechanically clean — last thing, right before you light the torch.
4. Set up local exhaust ventilation and confirm it is actually capturing at the arc, not near it.
5. Fit a copper backing strip.
6. Use TIG with 100 percent argon, DCEN, on thin sections; test AC if heat control is the problem.
7. Hold a short arc. Short means short — it is a fume control as much as a weld control.
8. Use ERCuSi-A for welding, or ERCuSn-A if colour match matters.
9. Keep the pulse short and the pool small.
10. Expect porosity — and cut it out if it is significant rather than welding over it.

Steps four and seven are the ones people skip, and they are the two that decide whether this job is a repair or a health event. Extraction positioned by feel at the arc, arc length held short on purpose — that pair converts an unavoidable zinc problem into a managed one.

A repair, start to finish

To make the workflow concrete, here is how a cracked fitting actually gets handled — every step in order, with the reason attached, because the reasons are what transfer to your next job.

First, the alloy question. The fitting was unmarked, which is normal on architectural brass — so it got filed, and the file cut with the characteristic bite of a ~65-percent copper alloy rather than the skitter of a high-zinc naval brass. Cartridge-brass territory: around 30 percent zinc, weldable in principle, brazable without argument. Had the file skittered, the decision would already have been made — high-zinc goes to brazing, full stop.

Second, the honest question: does this need fusion at all? The crack was in a non-structural seam on a visible fitting. Braze strength exceeds what the part ever sees in service, the polish on the surface argues for a filler that polishes, and brazing’s low heat protects the rest of the casting’s temper. Fusion welding would have been chosen against the material’s interests to satisfy an aesthetic preference for “real” welds. It got brazed.

Third, preparation — the step that decides the outcome. The crack was opened slightly at the root to give the filler a path — a small V along the fracture line — then degreased with solvent, then mechanically cleaned to bright metal within minutes of joining. Not an hour before; brass re-oxidizes, and oxide in the joint is either gas in the joint (if you move to fusion) or a wetting failure (if you do not). The rod came out of its sealed bag and got wiped before use.

Fourth, the setup. Flux suited to brass — not the self-fluxing copper rod’s assumption — painted into the joint. A heat sink clamped behind the seam: a thick copper backing that would eat heat away from the areas not being joined and hold the part flat. Extraction on, arm positioned at the joint, confirmed capturing by feel at eight inches before committing.

Fifth, heat discipline. A neutral oxy-acetylene flame, brought up patiently rather than aimed at the crack — brass conducts heat away fast, so the whole area gets warm before the filler moves, and localizing the flame on the crack itself is how people overheat one spot while the filler refuses to wet elsewhere. Filler applied when the base reached temperature, not before: touched to the joint, watched it drawn in by capillary action along the V, then off the heat. The brazing rod is the thermometer — when the base metal melts the rod on contact, you are already at the limit.

Sixth, after. Flux residue removed while warm — brass flux residue is corrosive if left, which is the classic way a perfect braze fails six months later from the outside. Joint dressed, checked for wetting along the full length, polished. No porosity to chase, no fume to manage beyond flux smoke, no distortion — the heat sink did its work and the base metal never melted.

That is the entire decision tree executing in about forty minutes. If the part had genuinely required fusion — wrong alloy family, temperature requirement, or written procedure — the sequence would have forked at step two, and everything from ventilation position to arc length would have been governed by the rest of this article instead. The technique is the last thing you choose, not the first.

Positions and technique compared

• TIG: DC electrode negative. AC is not normally used for brass TIG in most shops because the risk is tungsten contamination — on non-ferrous base metal tungsten can dissolve into the puddle, and tungsten inclusions in brass cause cracking. Keep the tungsten well off the work and grind it to the correct taper before starting. (The AC-versus-DC disagreement above applies to specific heat-control situations, not to casual practice.)
• MIG: DCEN with 100 percent argon as the usual choice; some shops use small oxygen or helium additions, which change arc character and zinc behavior — whichever gas you use, keep it consistent across the job.
• Vertical and overhead: brass is fluid and thin wall sags readily. Short stringer beads, frequent repositioning, and a chill bar behind the joint are the minimum for vertical work. Overhead brass welding on anything thin is genuinely difficult and is usually a brazing job instead.
• Thin brass: on 0.020–0.030 inch material, work at 20–60 amps with a tungsten sized for that range, and expect to spend most of your time on control rather than speed.

Positional work is where the case for brazing gets strongest. Every argument about sagging, overhead access and repositioning disappears when the base metal stays solid; you are only ever asking a filler to flow into a joint you can reach with a torch.

Starting Parameters

Brass settings do not transfer from steel, and they barely transfer between brasses — but honest starting points exist, and every one of them carries the same two modifiers: go smaller than you think on heat input, and adjust from a test bead on offcuts of the actual part rather than from a chart alone.

Two notes on the table. Pulse — TIG or MIG — is the single most useful feature available on brass work: it caps the pool’s peak temperature cycle by cycle, which is exactly the control zinc vaporization demands, and it is worth more than any arc-starting or synergy feature on the machine. And on every row, the arc-or-flame length is a dose setting as much as a weld setting: fume generation scales with voltage, so the transition from “technique preference” to “exposure control” is only true in one direction — there is no arc length on brass that is merely a style choice.

Troubleshooting Brass Welds

The pattern worth noticing: most brass problems are a heat problem. Brass has a wide melting range and high thermal conductivity, so the puddle stays fluid and zinc boils out before the base metal has fully wetted. Every entry in the fix column is some version of “less heat, more control.” When a fault does not respond to that, move up one level — is the gas right, is the filler clean, is the alloy what you think it is — rather than turning the machine up.

Four Myths That Cost You Parts

“It’s brass, so it brazes like copper.” Brazes similarly, maybe — with a completely different flux requirement. Self-fluxing Sil-fos works on copper and not on brass; use it on brass without separate flux and the joint refuses to wet. Same family, different chemistry.

“Manganese bronze is the brass rod.” It is a ferrous braze filler aimed at steel and cast iron work that happens to also be used on brass. The correct brass rod — RBCuZn-A or a BAg silver alloy — is usually one slot away on the same rack. This myth alone has produced thousands of wrong-colour, wrong-property joints.

“More preheat is safer on thick brass.” Preheat is a fusion aid for low-zinc alloys, not a general blessing. On high-zinc material it adds zinc loss and distortion without solving anything — Rapid Welding’s advice to keep preheat below 80 °C exists for exactly this reason. Preheat when fusion demands it, at the lowest temperature that works, with the current turned down to match.

“The fume is just smoke — open a door.” Zinc oxide fume is the specific cause of metal fume fever, its production scales with arc voltage, and the illness arrives hours after the exposure when nobody is looking. A door across the shop is not capture at the arc; the extraction arm twelve inches from the pool is the control, and it should be running before the arc strikes.

Storage and Handling of Brass

• Filler rod tarnishes fast. Brass rod picks up zinc oxide from the air, and contaminated filler is a direct route to porosity and cracking. Keep it in a sealed bag or lidded box, and wipe the rod before loading it into the torch or feeding it into the gun.
• Brass work-hardens. Hammered or heavily formed brass becomes stiff and more crack-prone. Do not straighten a welded brass assembly by pounding it — the repair you just made is the part that fails.
• Grinding brass is a dust hazard. Brass dust is toxic, and it will also contaminate a steel grinding wheel permanently. Use a dedicated wheel, wet grinding where practical, and local extraction — the same extraction discipline you use at the arc, aimed at a different hazard.
• Store the part, not just the rod. Greasy fingerprints, polishing compound and flux residue all survive to the weld if you leave them. Clean parts go into the job; clean parts come out of it.

Frequently Asked Questions

Is it difficult to weld brass?

Considerably more difficult than steel, and the reason is chemical rather than technical. Zinc boils below the melting point of brass, so every weld volatilizes zinc — causing porosity and fume simultaneously. TWI rates gas-shielded arc weldability of C46400 naval brass as Fair and recommends against MMA on zinc-bearing copper alloys altogether. The technique is learnable; the thermodynamics is not negotiable.

How do you join two pieces of brass together?

Soldering for thin, low-strength joints; brazing with a silver or brass filler for most structural and repair work; fusion welding for low-zinc alloys, thin sections, or joints that must match appearance. Most brass jobs should be brazed — the material’s own data sheets say so.

Can I weld brass with a MIG welder?

Yes, using ERCuSi-A or ECuAl-A2 wire with argon shielding, preferably pulsed. TWI recommends MIG rather than TIG above about 10 mm. Expect to fight spatter and control the pool more aggressively than on steel, and go as small on wire diameter as your feeder will feed reliably.

Do you weld or solder brass?

It depends on load and temperature. Bending or thin instrument work gets soldered; most repairs and hardware get brazed; fusion welding is reserved for low-zinc alloys, appearance-matching work, or joints that must exceed braze strength.

What filler should I use for brass?

For TIG welding: ERCuSi-A is the general choice, ERCuSn-A or ERCuSn-C if colour must match, ECuAl-A2 only for manganese bronze castings. For brazing: RBCuZn-A or a BAg silver alloy, with separate flux on brass. Do not reach for manganese bronze as a general brass filler, and never use phosphorus-bearing fillers on steel or nickel.

Is welding brass dangerous?

Yes, specifically because of zinc oxide fume. Zinc boils at about 907 °C and fumes hard at welding temperature, producing metal fume fever four to eight hours after exposure — which is why the connection is missed. OSHA’s PEL for zinc oxide fume is 5 mg/m³ eight-hour average and copper fume 0.1 mg/m³; local exhaust at the source is the primary control, with the short arc as a dose-reduction technique.

Can I braze brass with BCuP?

Yes — BCuP is used on brass, bronze, and gunmetal — but it needs a separate flux on brass rather than being self-fluxing. Never use it on steel or nickel: phosphorus forms brittle compounds with iron and nickel, and the joint will fail in a way no amount of brazing technique can rescue.

Do I need cadmium-free brazing filler?

If the joint touches food, drinking water, or a patient, yes. Cadmium-bearing silver brazing alloys (BAg-1, BAg-1a, BAg-2, BAg-2a, BAg-3 — 16–25% cadmium) are common in general work, cadmium is banned as a brazing filler in the EU under REACH, and US applications like food handling, surgical, and potable water require cadmium-free from the specification stage.

Should I preheat brass?

Sometimes, and it depends on the alloy. Preheat helps low-zinc brasses where fusion is fighting you, because lower current then suffices — and less current means less zinc loss. High-zinc brasses like naval brass generally do not need it. Avoid overheating: excess preheat on brass is a cause of distortion, not a cure for anything.

Why is my brass weld porous no matter what I do?

Work the list in order: is the arc short (long arc = five times the fume and more gas evolution)? Is the joint freshly cleaned? Is the filler rod out of its bag or has it been sitting on the bench picking up oxide? Is the alloy high-zinc — in which case the porosity is the zinc, and the fix is not parameters but a different joint design or brazing? Is the extraction actually capturing? If all five check out and the porosity persists on a 30-percent-zinc alloy, you have reached the material’s limit, and the honest answer is that this joint should have been brazed.

Can I weld brass to steel?

Not by fusion, and this is a joint where people waste a lot of rod learning why. Brass and steel do not alloy with each other — you get a brittle intermetallic at the interface rather than a mixed pool, and diluting brass with iron compounds the zinc-gas problem. The standard route is mechanical (brazed or threaded transition, or a brazed sleeve joint), and where a metallurgical transition is genuinely required, brazing with a filler chosen for both parents is the normal answer. The same “two metals that will not mix” logic applies to welding cast iron to steel, where the joint design does the work the metallurgy will not. Galvanized steel makes it worse: the galvanizing layer is zinc, and heating it puts you right back in fume-land.

Does brass change colour when you weld it?

Yes, and the colour change is direct evidence of zinc loss — the copper that stays behind is redder than the original alloy, so the weld zone and heat-affected zone often come out pink-to-red against the yellow parent. It is not a surface stain that polishes off entirely; some of it is the alloy actually changing composition near the fusion line. On appearance-matching work this matters more than strength: low-zinc alloys with closer colour match, shorter arc, and lower heat input all help, and on decorative work the usual plan is to accept a brazed joint that polishes back to match.

How do I know what brass alloy I have?

Look for a stamp, engraving, or a supplier mark first — cast and extruded architectural pieces often carry one. Failing that: a file test gives a rough read (cutting easily and leaving a coppery smear suggests a lower-zinc cartridge brass; skittering suggests higher-zinc naval or Muntz metal), a spark test is of limited value on non-ferrous metal, and a handheld alloy sorter or XRF gun gives certainty if you have access to one. When the alloy genuinely cannot be identified, treat it as high-zinc and braze — that decision is safe across the whole family, while guessing low-zinc and guessing wrong puts you welding material the process cannot handle.

Wrapping Up

Brass resists fusion welding for one reason: its zinc leaves when you heat it. Everything else follows from that one fact — the porosity, the colour change, the lost ductility, the fume, the health risk, the distortion. They are the same problem showing up in different places, and no setting on any machine changes the boiling point of zinc.

So decide the joining method before you decide the settings. If the part can be brazed, brazing gives you a better joint with none of the zinc exposure — and the brass never melts, which resolves every problem on this page simultaneously. If it genuinely needs a fusion weld, keep the zinc content low, keep the arc short, keep the pool small, clean the joint immediately before welding, and have the extraction running before you strike.

And when you do fume zinc, remember the fever shows up four to eight hours later. Working through it because you felt fine at the end of the shift is how people end up with a recurring problem — the tolerance builds, the warning fades, and the ventilation never gets fixed. The restorer’s fitting got brazed, the extraction stays on, and the only thing that fumed was the flux.