How to Clean Welds — Between Passes and After the Final Pass

Job: get slag, silica islands, spatter, and oxide off every bead — between passes so the next pass fuses to metal instead of glass, and after the final pass so inspection sees the weld and not its crust.

Time: one to three minutes per pass. The labor is trivial; the discipline — every pass, every crater, no exceptions — is the actual skill.

Difficulty: first-month technique, career-long discipline. The shops that get stainless callbacks are not the ones that can't brush; they're the ones that let one carbon brush wander.

All MAXXGRIP PRO®: .40mm wire, dense fill, molded finger protectors — the knuckle guard earns its keep the first time a brush skips off a bead.

Everything else: a chipping hammer — this catalog does not carry one; buy it at any welding supply, because nothing else does its job on a full slag sheath · safety glasses with side shields, worn under the hood · welding gloves · a paint pen to mark the stainless brushes · a bright light you can rake across the finished bead.

Before you start: two questions and a law

Glasses on, under the hood, before the first hammer strike. Chipped slag is hot glass shrapnel, and a working wire brush sheds broken .40mm filaments at speed. OSHA 29 CFR 1910.133 names flying particles as the hazard class and requires side protection for it; look for the ANSI Z87.1 mark on the frame. Welding fume, arc rays, and hot-work fire watch belong to the welding operation itself — run those per your welding procedure and permits; this guide covers the cleanup hazard only.

Name your process, and you've named your debris. Stick and flux-core leave a solidified flux sheath — slag — over every bead. MIG leaves no sheath but scatters spatter and glassy brown silica islands across the bead surface. TIG leaves only oxide and heat tint. The steps below split exactly along those lines.

Name your metal, and you've named your brush. That one is the law, and it gets its own section.

The law: one brush, one alloy — forever

Carbon-steel brushes touch mild steel only. Stainless and aluminum get dedicated stainless brushes that have never — not once — touched carbon steel.

The reason is metallurgy, not fussiness. Every stroke of a carbon-steel brush deposits microscopic free iron in the surface it cleans. On mild steel headed for primer, nobody cares — it's iron on iron. On stainless, that embedded iron defeats the passive chromium-oxide layer that makes stainless stainless, and it comes back in weeks as rust bloom and pitting on a metal the customer paid to never rust (The Fabricator — brush technique, finishing.com — iron contamination). On aluminum, trapped iron debris corrodes galvanically.

Mark them. Paint pen on the stainless handles the day they arrive. Hyde sells every size in both fills, so the pairs are identical tools told apart only by your marking — which is exactly why the marking is not optional.

Rack them apart. Stainless brushes live at the stainless bench or in their own pouch, never loose in the common drawer.

Demote on contact. A stainless brush that touches mild steel once is a carbon brush for the rest of its life. Relabel it on the spot — it will still brush mild steel beautifully. The demotion runs one way and it is permanent.

The same segregation logic covers everything that touches stainless — flap discs, grinding wheels, clamps — but those are other guides' problems. Here it is brushes: one brush, one alloy.

One honest note on direction: stainless-on-mild-steel harms nothing but the brush's stainless status. Carbon-on-stainless harms the work, invisibly, until it blooms.

Between passes

Chip first where slag is heavy. Stick and flux-core beads wear a full slag jacket, and the chipping hammer — not the brush — takes it off. Let the bead lose its color for a moment: slag releases easiest as it cools and contracts, and low-hydrogen slag will often curl and lift in whole sheets. Work from the crater back along the bead, striking at an angle so shards fly away from you, and land deliberate hits along the toes, where slag keys into any undercut. The 46834's scraper nose earns its place here — popping a stubborn glob or a spatter cluster without a walk back to the bench — but it assists a chipping hammer; it does not replace one.

Then brush along the bead, and let the tips cut. A wire brush cuts with the points of its wire and nothing else. Moderate pressure, long strokes along the bead axis, then extra passes angled into each toe — that is where residue survives. Leaning on the brush folds the .40mm wire sideways so it wipes and polishes instead of cutting, and enough leaning sets the fill flat permanently. If you catch yourself leaning, you don't need more force — you need the hammer or the scraper nose first. The dense MAXXGRIP fill is the speed feature: more cutting points per square inch cuts faster *and* leaves a better surface, the same way a fine file at speed beats a coarse one.

Know your process's tell:

Stick (SMAW): after chipping, brush until the toes and ripple valleys show metal. Cellulosic rods (6010-class) leave a thin, tight slag that takes more brush and less hammer; low-hydrogen (7018-class) is the sheet-lifter that still hides chips in the toes.

Flux-core (FCAW): the same routine with more spatter — and watch the wagon tracks. The toe grooves of a multipass fillet trap continuous slag lines that must come out before the next bead rolls over them.

MIG (GMAW): no sheath, but do not skip the pass. Brush the soot band off the toes, then hunt the silica islands — glassy patches left by the wire's deoxidizers, usually at starts, stops, and ripple valleys. Here is the trap: they do not brush off. Wire over them and they polish up shiny and stay put. Pop them with the hammer point or the scraper nose; they click off like glass, because that is what they are.

TIG (GTAW): nothing to chip. On multipass stainless, a pass with the dedicated stainless brush strips loose oxide so the next puddle wets clean. On aluminum, the dedicated stainless brush is also the *pre-weld* ritual — the oxide skin melts at roughly three times the temperature of the metal underneath it, and brushing is what breaks that skin — plus a light pass between beads. Light hand: aluminum is soft, and you are removing oxide, not resurfacing the part.

The between-pass standard: bare metal on the bead and both toes before the arc restrikes. Anything glassy you leave becomes a slag inclusion the next pass buries — visible to X-ray, invisible to you, permanent. Craters and restarts get double attention; that is where slag pools deepest and silica collects. (Interpass temperature and preheat are welding-procedure territory — hold them per your WPS; they are not this guide's business.)

After the final pass

Clean the weld and an inch of parent metal on both sides. Full slag and spatter removal, then brush the soot and oxide band off the toes until the fusion line is honestly visible. The final clean is not cosmetic — it is the precondition for inspection. And know when to stop: the inspector needs the *surface*, not a buffed version of it. Brushing a mild-steel weld to a uniform shine can smear soft metal across the very toe line someone needs to read.

Stainless final clean: dedicated brush, honest expectations. The stainless brush takes off loose oxide and leaves no iron behind — that is its whole job. It will lighten but not erase heat tint, and it does not restore the passive layer; where the spec calls for pickling, passivation, or electropolish, the brush is the first step of that process, not a substitute for it. Write what was actually done on the traveler rather than letting the finish claim more than it is.

Rack discipline is part of the job. Carbon brushes back to the carbon spot, stainless back to its marked rack, and any brush that crossed the line relabeled before you walk away. A brush rack you can audit at a glance is the cheapest quality system a weld shop owns.

The inspection hand-off — what a brush proves, and what it can't

A properly brushed weld lets visual inspection tell the truth: undercut, overlap, toe cracks, crater pipes, and surface porosity are all findable because nothing is hiding them. That is the whole promise, and it is worth the three minutes. What the brush does not do: it fixes none of those defects, it cannot reach inclusions already buried under a covering pass, and NDT beyond visual — penetrant on stainless especially — carries its own cleanliness requirements from the procedure, not from this page. Hand the inspector a clean weld and an honest account of what was done to it; a polished weld with a vague story helps nobody.

What goes wrong (and the fix)

"The stainless is rusting." Someone's carbon brush — or grinder dust, or a bench clamp — put free iron in the surface, and it bloomed. There is no brushing it away after the fact: the affected surface needs mechanical removal and re-passivation per spec, which is a refinishing job, not a cleanup. The fix you actually control is prevention — dedicated, marked, demote-on-contact.

"The brush is shedding wire everywhere." You leaned on it and folded the fill, or it is simply spent. A shedding brush throws filaments at eye height and leaves steel fragments on the work — which on stainless is contamination by another route.

"The bead looks clean but shiny patches won't move." That is silica or polished slag, not metal — you have been buffing glass. Tap with the hammer point or scraper nose until it clicks off, then brush what was underneath.

"The toes look clean but soft." Smeared oxide from over-brushing — and smeared toes are exactly where undercut hides. Rake a light across the weld at a low angle: a real fusion line reads crisp; a smeared one goes soft-focus. Re-brush lightly along the toe and let the tips cut.

"I'm brushing forever." The brush is a finishing cutter, not a bulk remover. Heavy spatter wants anti-spatter compound or gun settings fixed upstream, and a scraper or hammer first; a brush asked to do a chisel's job just wears out angry.

The history footnote (because these pages should be worth reading)

The slag you just chipped is the invention working. That shield solidifies into the slag sheath, which means the chip-and-brush ritual is exactly as old as sound arc welds: before the coated electrode, bare-wire arcs made brittle, air-poisoned joints and left nothing to chip.

Tools this guide links back to: Hyde MAXXGRIP PRO® carbon scratch brushes 46830 (6" x 1") · 46841 (5-1/2" x 1") · 46832 (3" x 3/4" detail) — mild steel only · dedicated stainless 46831 · 46842 · 46833 — stainless and aluminum, marked and segregated · 46834 carbon with scraper nose — the slag-glob and spatter assist · stainless minis 46650 / assorted minis 46843 — boxed corners and limited access (the 46843's stainless member follows the same dedication law). Every linked PDP lists this guide under its applications — the graph runs both directions.