Fumevac Robotic Welding Ventilation Hoods Explained

A manufacturing plant switches a manual MIG station to a robotic cell, cuts two people from the job, and celebrates the productivity win. Six months later, an occupational nurse flags a welder’s cough and the plant discovers the cell has no fume capture at all — because the fume is still being made, just faster and in a place nobody was watching. A robot removes the arc from a human hand; it does not remove the fume. That gap is exactly where welding ventilation hoods like the Fumevac range come in, and it is where most robot cell ventilation plans quietly fail.

Why a Robot Cell Still Needs a Ventilation Plan

Automation is excellent at removing hands from a job. It is not very good at removing chemistry from a room.

A robot cell gets sold on two numbers: cycle time and headcount. Two welders replaced, more parts per shift. That is real money. But an arc does not care who holds the torch. It still vaporises base metal, filler metal and flux, condenses that vapour into metal fume, and produces gases such as ozone and nitrogen oxides. A robot changes the geometry and timing of the plume, not the fact that fume is being made. Often it makes more per shift than the station it replaced.

For a plant manager, safety lead or welding engineer pricing a new robotic cell, ventilation is not an afterthought. It belongs to the cell design, not the building.

What actually changes when you automate

Three things change when a manual bay becomes a robot cell, and only one of them is good news.

• Who is exposed. In a manual bay the welder’s breathing zone sits 6 to 12 inches from the arc. In a robot cell that welder may be gone, and exposure shifts to the loader, the grinder between cycles, the touch-up welder inside the enclosure, and the operator walking the aisle during a fault.
• How much fume per hour. Arc-on time and duty cycle rise sharply. A robot welds without rest breaks and without the fatigue that slows a manual operator, so one cell can pulse fume for hours.
• Where the fume goes. With no human leaning over the arc to disturb the plume, the fume column rises cleanly, spreads along the ceiling and drifts. It reshapes itself every time the robot moves, making any fixed capture point a compromise rather than a solution.

Exposure factor	Manual MIG bay	Robotic cell
Arc location	Held in hand at chest height	Programmed path, often high or in a fixture
Dominant exposure	Welder breathing zone, continuous	Operators and loaders, intermittent
Fume per shift	Limited by human endurance	Limited by duty cycle, often higher

The last row is the awkward one. Fume per hour matters, but fume per shift hour decides whether the cell needs dedicated extraction. Because the plume changes shape constantly, a hood that works on a static bench can miss on a moving robot.

Material choice compounds it. Flux-cored aluminum work carries its own fume profile (see our guide to aluminum flux core welding), while alloy selections bring their own considerations in the guide to alloy steel welding electrodes. Each shifts what the hood must capture.

Breathing zone versus ambient zone

Two different measurements, and you must satisfy both. The breathing zone is the air around a person’s head and chest, where exposure limits are evaluated. The ambient zone is the general air inside the enclosure. Extraction must clean the breathing zone for whoever is inside and stop the cell filling with haze that drifts into the next bay. Handling only the air near the torch still lets the cell atmosphere build up.

The legal baseline, before you buy anything

This is not a grey area. OSHA 1910.252(c)(1)(iii) requires local exhaust or general ventilating systems to be provided and arranged to keep toxic fumes, gases and dusts below the maximum allowable concentration in 1910.1000. It also names what drives contamination: the dimensions of the space, especially ceiling height, the number of welders, and the possible evolution of hazardous fumes, gases or dust according to the metals involved.

Over in the UK, HSE states plainly that all welding fume can cause lung cancer and that welding fume falls under the COSHH Regulations. HSE also flags what applies directly to robotics: a local exhaust ventilation system in a robotic cell may not capture all the fume, so how far operators can work becomes a main factor in deciding whether mechanical general ventilation is also needed.

What this guide covers

If you are researching Fumevac robotic welding ventilation hoods, you are probably asking one question: what has to be true for a hood to control fume in a cell where the arc moves? This guide answers it in five parts:

• What welding fume is, and why automating a weld changes nothing about the hazard.
• How a hood captures the plume, and the three capture arrangements used in robot cells.
• How to size the hood from the far point of the robot’s work envelope using OSHA’s capture table.
• Where to place it so the robot, part and positioner do not defeat it.
• What the standards require, and the mistakes that fail in practice.

Automation moved the arc, not the hazard. Everything that follows is about engineering the control around where the arc goes.

What Welding Fume Is (and Why Automating Welding Doesn’t Remove It)

Welding fume is not campfire smoke. It is a fine metal dust formed when the arc melts base metal, filler rod or wire, and sometimes a flux coating. The heat vaporises the metal and the vapour cools instantly into microscopic particles that hang in the air and ride the heat upward. Alongside come gases: ozone, nitrogen oxides and, depending on process and shielding gas, carbon monoxide. Flux-cored and stick consumables add coating decomposition products, so they carry a separate hazard profile.

None of that changes because a robot holds the torch. The chemistry at the arc is identical. What changes is the rate, the pattern and who is standing where.

What a Robot Cell Changes

Four shifts matter for ventilation design:

• Duty cycle rises sharply. A manual welder stops, repositions, checks fit-up and grabs a coffee. A robot runs the same seam with no idle time, so fume per shift can be several times higher even when the weld is identical.
• Arc-on time is higher and steadier. Concentration settles into a predictable background level instead of spiking and decaying.
• The plume moves differently. With no welder leaning over the arc, no body blocks the thermal draft. The plume rises more freely, travels further along the fixture and can drift across the cell before mixing with room air.
• The plume comes in a repeatable pulse. Each seam drops a fresh cloud at the same relative position, so a hood can be aimed at a known point. An advantage, and only if the hood is actually there.

Not All Fume Is the Same

• Mild steel gives the familiar iron-oxide dominated fume, plus manganese from base metal and wire. It is the volume hazard in most fab shops and the one driving the manganese limits below.
• Stainless steel adds chromium and nickel. Cutting, grinding or welding stainless with a stainless wire can generate hexavalent chromium, which is why stainless fume can cause occupational asthma and why OSHA treats stainless with its own specific ventilation rules in 1910.252. Processes like aluminum flux core welding and aluminum MIG welding bring a third set of constituents entirely, aluminium and its flux residues, needing their own capture approach.
• Galvanised (zinc-coated) steel releases zinc oxide fume when the arc breaks the coating, the classic trigger for metal fume fever, the flu-like illness welders report after a bad afternoon on galvanised steel. Galvanised material also brings lead into the mix, and lead is one of the materials where OSHA mandates local exhaust ventilation indoors rather than leaving it to judgement.
• Coated and painted steel adds organic binders and pigments that decompose into fume, and some coating chemistries are more hazardous than the metal underneath. The consumable matters too: a low-hydrogen electrode does not produce the same flux decomposition products as a cellulose electrode, and a stainless wire behaves differently from a mild steel wire on the same part.

The Health Picture

Welding fume is a serious long-term health hazard, and the HSE health risks guidance is blunt about it. The International Agency for Research on Cancer classifies all welding fume as a Group 1 carcinogen, meaning definitely carcinogenic to humans, with lung and kidney cancer among the outcomes of concern. HSE issued a safety alert in February 2019 on the link between mild steel welding fume and cancer.

• Lung cancer, the headline long-latency risk.
• Metal fume fever, the flu-like shivers and chest tightness linked to hot work on galvanised metal. HSE reports that 40 to 50 welders are hospitalised each year in Britain through breathing metal fume at work.
• Occupational asthma, which HSE links to hexavalent chromium, nickel and cobalt in stainless fume.
• Welder’s lung, a pneumoconiosis from long-term iron oxide exposure, associated with chronic bronchitis and emphysema.
• Neurological effects from manganese, including the movement and behaviour changes of manganism.
• Pneumonia risk, especially in enclosed or poorly ventilated spaces where fume builds up.

The Numbers: Manganese and the ACGIH Limits

Authority and basis	Limit
NIOSH recommended exposure limit (REL), TWA	1 mg/m³
NIOSH short-term exposure limit (STEL)	3 mg/m³
OSHA permissible limit, ceiling	5 mg/m³
NIOSH IDLH (immediately dangerous to life or health)	500 mg/m³
ACGIH TLV-TWA, respirable fraction	0.02 mg/m³
ACGIH TLV-TWA, inhalable fraction	0.1 mg/m³

Per the NIOSH welding health topic page, some studies indicate welders exposed to low levels of manganese, below 0.2 mg/m³, perform more poorly on tests of brain function and motor skills, and prolonged exposure above 1 mg/m³ may lead to Parkinsonian symptoms known as manganism. That is why the ACGIH numbers sit so far below the OSHA ceiling of 5 mg/m³: a plant that complies perfectly with the OSHA PEL may still be well above the level at which neuro effects have been observed.

That gap is also why ACGIH TLVs and BEIs are guidelines, not standards. ACGIH states plainly that these are health-based guidelines, set without consideration of economic or technical feasibility. They are not law, but they are the current scientific view of a safe exposure, and courts, insurers and regulators increasingly treat them as the benchmark. Where OSHA and ACGIH disagree, the practical question is which one your client, your insurer and your safety lead will hold you to.

Why the Control Logic Doesn’t Change

The hazard profile in a robotic cell is the same hazard profile as a manual bay, in the same order of severity, for the same reasons. So the control logic is identical too: capture at the source, keep it out of the breathing zone, fall back on respiratory protection for what escapes. The HSE controls guidance makes the same point in its own terms, and OSHA’s respiratory protection standard places accepted engineering controls, including local exhaust ventilation, ahead of respirators in the required hierarchy.

What automation changes is the engineering problem. The source moves, it moves faster, and nobody is watching it. That is why a fixed fume extraction hood sized for a stationary manual workstation under-performs in a robot cell, and why the sizing, positioning and interlock work that follows matters so much.

How Fumevac Robotic Welding Ventilation Hoods Actually Capture Fume

Strip away the sheet metal and the spec sheet and a fume extraction hood does one job: move air in a controlled direction so the plume is captured at the arc and carried away before it reaches anybody. Fume leaving a hot arc is buoyant, so it shoots upward, expands as it cools, and spreads into a drifting cloud that follows whatever air currents exist in the cell. A robot running a clean aluminum MIG setup or a cell running aerospace grade TIG produces that plume whether or not a human is holding the torch.

A hood creates a controlled flow of air at the capture point. Air is pulled in through the hood opening, the rising plume is entrained and drawn into the inlet, and the contaminated air travels down the duct to a filter, through a fan, and out as clean air. Capture happens at the hood face. Everything downstream only moves what the hood already collected.

The Three Components of Any Extraction System

Component	Job	What goes wrong without it
Capture hood	Creates the airflow field that entrains the plume at the arc	Fume is generated faster than it is captured, so it drifts into the cell and the operator’s breathing zone
Duct and filter housing	Transports captured fume and removes particulate before air returns to the shop	Contaminated air recirculates into the workspace instead of being cleaned
Fan and clean air discharge	Drives the flow and expels filtered air outside or to a safe discharge point	Static pressure falls short and the system under-performs its design figure

The hood is where the work happens. A bigger fan cannot rescue a badly placed hood; it just moves more clean air around a plume that was never captured. Nor can a better filter, because it only ever sees the fume that reached the duct. Put the best filter money can buy on a hood aimed at the wrong place and you still have a fume problem, just a more expensive one.

Three Ways to Arrange Capture in a Robotic Cell

• Articulating or fixed arm hood following the torch. A hood on a swing arm or linear track sits near the arc and follows the weld path. It is the classic arrangement, and it works well when somebody keeps it in the right place. In a robot cell, that somebody is a programmer or an interlock.
• Hood mounted near the workpiece or fixture. The hood is fixed over the part or positioner, so capture geometry stays repeatable cycle after cycle. It suits high-volume, repeatable weld paths where the arc stays inside the capture zone, but a positioner that swings the part out from under the hood undoes the design.
• Enclosure or booth with downdraft. The operation is enclosed and air is drawn down through the roof, capturing fume from welding, grinding and spatter handling in one layer. It best protects a whole cell and costs the most in floor space, heat and make-up air.

Capture Velocity and Control Velocity

Capture velocity is the speed of air at the hood opening needed to drag the plume in. Control velocity is the speed maintained throughout the zone of welding to stop the plume escaping. OSHA 1910.252(c)(3)(i) requires freely movable hoods to be placed as near as practicable to the work, with airflow sufficient to maintain a velocity toward the hood of 100 linear feet (30 m) per minute in the zone of welding when the hood is at its most remote distance from the point of welding. That clause bites in robot cells, because the robot’s most remote point is not its typical point. 1910.252(c)(3)(ii) sets the same 100 fpm figure for a fixed enclosure with a top and not less than two sides, directed away from the welder.

British practice is expressed as a range. HSE guidance on welding fume control describes capture velocities of 0.5 m/s to 1 m/s as appropriate for welding fume, soldering and liquid transfer, and warns that higher capture velocities may be needed where draughts interfere with the flow into the hood.

Three Ways Capture Fails in a Robot Cell

1. Thermal lift. The arc’s own buoyancy drives the plume upward faster than a weak hood can pull it down, and it is the most common reason a hood that worked on a bench top job fails over a production cell.
2. Cross drafts. An open cell door, a draughty aisle, extraction from a neighbouring machine or the shop’s general ventilation can push air across the plume, blowing it out of the hood or diluting the capture field. A gas-shielded process in a drafty bay is a hard case.
3. Robot and positioner interference. The torch, dress pack, part geometry or a rotating positioner moves through the capture field, blocks the inlet, or hits the hood. Operators respond by shoving the hood aside or switching it off, which turns a ventilation failure into a control failure.

Sizing the Hood: Turning OSHA’s Capture Table Into Real Airflow

Sizing a fume extraction hood for a robotic cell is less guesswork than most people assume. OSHA’s capture table gives a minimum airflow and matching duct diameter for any distance between hood and arc.

Start with the only distance that matters: the furthest the hood will be from the arc when the arc is lit. Not the average. The worst case.

What OSHA’s Capture Table Actually Says

The table in 29 CFR 1910.252(c)(2)(i) sets minimum ventilation requirements for a freely movable hood. Here is the working version most cell designers use:

Distance from arc or torch (inches)	Minimum airflow (cfm)	Minimum duct diameter (inches)
4 to 6	150	3
6 to 8	275	3½
8 to 10	425	4½
10 to 12	600	5½

Two things to notice. First, airflow roughly doubles with every two inches of standoff: six extra inches takes you from 150 cfm to 425 cfm. Second, the duct diameters assume a transport velocity near 4,000 feet per minute, high enough to carry spatter and slag without blocking the duct, yet low enough to keep friction loss manageable.

The Distance That Governs Is the Far Limit of the Envelope

Robot cells break the assumption behind the table. In a manual bay the welder moves the hood with the work; in a robotic cell nobody adjusts anything, so hood-to-arc distance varies across the work envelope by design.

Map the envelope first: the furthest point of the weld path from the hood’s reach, the furthest point from any wall or fixture the hood can mount to, and the positions where the positioner rotates the part so the seam ends up on the far side. If a two-position positioner flips a part 180 degrees, the hood must serve both orientations. The design distance is the maximum of those.

A robot swinging a 40-inch part across a 10-foot table can put the arc in places the hood has never been. A hood on a linear track following the torch keeps the design distance short; a hood fixed over one station, with the robot reaching into a table corner, means designing for 10 to 12 inches and 600 cfm. Those strategies cost very differently, so sizing belongs at the concept stage.

Adding Real Margin: Loaded Filters, Duct Losses and Duty Cycle

The OSHA figures are minimums measured at the hood with the system running. Real systems need more, for three reasons.

• Filter loading. As a filter fills, its resistance climbs, the fan pushes less air and volume at the hood falls, so a system sitting on the table minimum drifts below it within weeks. Specify the fan so it still meets the required cfm with the filter at its rated dirty capacity and full duct length installed.
• Duct and hose losses. Every elbow, every metre of flex hose and every transition costs static pressure. Flexible duct is worst because it flexes constantly, and ducting spanning a robot base has to survive being driven over.
• Duty cycle. Ask for the arc-on percentage and fume generation rate in the specification. Two cells welding the same part can differ enormously in fume output if one runs at 70 percent arc-on and the other at 25 percent, and filter loading rate follows from it.

Why Oversized Ducts Matter More Than Most People Think

Static pressure loss in a duct scales steeply with velocity: double the flow through the same pipe and friction loss roughly quadruples. A 5½ inch duct moving 600 cfm loses far less pressure than a 4½ inch duct moving the same volume, leaving the fan more headroom for hood and filter. In a cell that headroom keeps airflow up as the filter loads and allows a long duct run without a second fan. The trade is space and cost.

Two Worked Examples

Example one: single-station MIG robot welding 6-inch tube. The tube is clamped in a positioner at bench height, hood on a short articulated arm off the robot’s base column, following the torch around the circumference. Design distance from hood mouth to arc is about 6 inches, though the operator sometimes leaves the hood low for the root pass, pushing standoff to 7. Design on the 6 to 8 inch band: 275 cfm through a 3½ inch duct, fan curve specified for at least 300 cfm at the hood with the filter loaded. A 6-inch tube job is often under 25 percent arc-on, so a check every three to six months is realistic.

Example two: dual-station fixture handling 10-foot sheet metal. With two stations the hood cannot sit over one fixed spot, so the design distance is governed by the far end of the sheet: 10 to 12 inches, 600 cfm minimum on a 5½ inch duct, and specify nearer 700 cfm for margin. At station two the sheet is a wide flat plate that disperses fume along the seam, so the plume is longer than the arc itself, and the higher duty cycle loads the filter faster. Specify more filter media capacity plus a prefilter screen or spark arrestor, because a 10-foot sheet throws more spatter into the intake. In a tight cell an on-torch hose becomes attractive for a different reason: it follows the arc rather than covering a large moving plume.

Verification: Airflow Is Not the Same as Capture

A system can deliver every cfm you specified and still fail to capture. That is why HSE says that for on-torch extraction, capture velocity measurement is not an appropriate test, and that total volumetric flow should be checked against the manufacturer’s recommendations in the user manual. LEV performance should be judged on effective capture of welding fume rather than airflow measurements alone. Confirm the design flow, then confirm capture by watching where the plume goes with the cell running.

If the material is aluminium rather than steel, fume volume and filter loading behave differently enough to warrant their own review. Our guides to aluminum MIG welding and aluminum flux core welding cover the process side, and the ventilation logic is the same: different consumables, different fume, different loading rate.

Where to Put the Hood in a Robotic Cell (Reach, Interference, Positioners)

The best hood on the market does nothing if it sits three feet from the arc. Placement is a geometry problem before it is a purchasing problem, and most failed installations fail because nobody mapped the cell first.

Survey the cell before you choose anything

Before routing a single duct, walk the cell and record what actually moves: the robot’s working envelope marked on the floor, the full swept area of every positioner and turntable, fixture and part height above the table surface, the location of the wire feeder plus the gas and coolant lines feeding the torch, the route of the dress pack and cooling water hoses, the teach pendant and operator walkway, and the nearest obstruction in every direction.

The last two items matter more than people expect. The torch, wire feed and shielding gas setup set how far the manipulator can realistically stretch, because every extra metre of hose is drag, torsion and a trip hazard. A robot with a long dress pack reaches further than the catalogue suggests, but it reaches unstably at full extension, which changes where the weld can actually be made.

Know the legal baseline too, because it shapes the design. OSHA 1910.252(c)(2)(i) calls for mechanical ventilation when welding takes place in a space of less than 10,000 cubic feet per welder, in a room with a ceiling below 16 feet, in confined spaces, or where partitions, balconies or other structural barriers significantly obstruct cross ventilation. That describes most robotic cells. The UK position is comparable: HSE notes that mechanical general ventilation will be required in most welding workshops because natural ventilation from open doors and windows cannot disperse the fume generated by work tasks.

Placement strategies, most effective first

1. On-torch extraction at the gun. The nozzle sits a few centimetres from the arc, so capture happens at the source and travels with the weld. The strongest option for a moving arc.
2. Hood on the manipulator, or on a linear track that follows the arc. The hood tracks the weld path, so capture distance stays constant wherever the robot reaches.
3. Hood over the workpiece or fixture. Effective on repetitive parts where the weld path stays inside a small zone, provided the hood is wide enough to cover the whole plume spread.
4. Whole-cell downdraft booth or roof extraction. Captures everything in the enclosure, including grinding dust and spatter, but costs the most and needs a make-up air path.

The interference problem

A hood on a swing arm is the classic failure. The arm gets clipped by the robot on a fast reposition, catches the part as the positioner indexes, or swings into the operator’s load path. The predictable response is that someone parks the arm out of the way or switches extraction off so the robot can finish a cycle. That is a control failure dressed up as housekeeping, and it will happen whether or not it appeared in the risk assessment.

Mitigate it by mounting the hood so its swept volume sits outside the robot and positioner envelopes, using compliant or retraction mounts, and making hood position a taught position the robot references rather than something an operator sets by hand.

The last remote point rule

Size the system for the furthest point of the weld, not the nearest. With a fixed hood and a travelling arc, the governing distance is the far edge of the work envelope. With a hood tracking the arc on the manipulator, it is the tracking error, typically the clearance between hood lip and work at full reach. Test capture where the arc spends its time and where it is hardest to reach, because the fume does not care which is convenient.

The operator inside the enclosure

A robot cell changes who gets exposed. The loader, the grinder between cycles and the welder doing touch-up inside the fence all breathe the same air as the robot’s weld. Design the cell so those tasks happen with extraction running, and keep touch-up welding to a procedure that does not generate heavy fume, ideally outside the enclosure.

If any part of the cell qualifies as a confined space, 1910.252(c)(4)(i) requires all welding and cutting to be adequately ventilated to prevent accumulation of toxic materials or possible oxygen deficiency, applying not only to the welder but to helpers and other personnel in the immediate vicinity, with replacement air that is clean and respirable. And 1910.252(c)(4)(v) is unambiguous: oxygen shall never be used for ventilation.

Filtration, Spark Arrestors and Filter Change Strategy

A fume hood can be perfectly sized and positioned and still protect nobody, because the filter inside has loaded up and the fan pushes less than half the air it pushed on day one. In a robotic cell that failure is invisible: nobody stands at the machine with an anemometer every shift. The robot keeps welding and the fume finds its way back to whoever loads the parts. The filter, not the hood, decides whether capture still works.

The Filtration Chain, Stage by Stage

Most units stack stages in series, each catching what the last missed.

• Prefilter or spark screen: Mesh or perforated inlet screen, often behind a self-closing flap. Mechanical, not filtration: it stops spatter, slag and hot sparks reaching the pleated media, because a spark on a filter face can burn through it or melt the gasket in seconds.
• Primary stage (cartridge or bag): Coarse particulate, the heavy mass. Cartridges use pleated synthetic media; bag or “sock” filters hold a lot of dust at low pressure drop. This stage takes the abuse in cells that also grind and deburr.
• Secondary stage: Finer media at higher efficiency, where the respirable fraction reaching the alveoli gets stopped.
• Gas stage (activated carbon): Impregnated carbon that adsorbs gases and vapours. Essential when the cell runs stainless, galvanised or coated material, or a shielding gas leak is possible.
• HEPA stage: For very fine or hazardous particulate, or where discharge air is recirculated into a shared workshop.

Efficiency Is Not Capacity

Efficiency is the fraction of a given particle size the filter removes. European welding fume filters are classed under EN 1289 and EN ISO 16891, with classes such as E10, E20 and E30 for particle filters. Capacity is how much particulate the filter holds before its resistance climbs to an unacceptable level, measured in grams. A high-efficiency filter with low capacity clogs in a shift; a high-capacity filter with mediocre efficiency keeps the fan happy while letting fine fume through. For an unattended cell running long cycles, capacity is usually the number that matters.

Specification	What it tells you	Why a robot cell cares
Filter class	Capture percentage at stated particle sizes	Residual particulate in the cell when fresh
Capacity (grams)	Dust held at a given pressure drop	Hours of arc-on time per filter
Pressure drop at rated flow	Static resistance of the media	Fan sizing and flow loss as it loads
Carbon mass	Adsorbent capacity, not particulate capacity	Whether ozone, NOx or CO are removed

Both numbers scale with duty cycle. A cell welding one 20-second seam every three minutes generates a fraction of the fume of a cell running a 4-minute seam every 30 seconds, and that difference belongs in the maintenance plan, not in a guess.

Why Activated Carbon Matters

A particulate filter does not remove gases. It will not take ozone, nitrogen oxides or carbon monoxide out of the stream. HSE states plainly that particulate filters will not protect workers from welding gases contained in the fume. That matters in a robotic cell, because gases are exactly what is left once the particulate has been captured. Galvanised and zinc-plated work gives off zinc oxide fume, both particulate and a serious lung irritant and carcinogen. Stainless welding produces ozone and nitrogen oxides, which can trigger occupational asthma. Shielding gas leaks add argon or CO₂ displacement problems.

Filter Loading: The Silent Failure

As a filter loads, the media packs tighter, resistance climbs and the same fan delivers less flow. Because capture rests on a minimum capture velocity at the hood mouth, a filter that has quietly cost 30 percent of airflow has quietly cost a third of your capture performance. The plume still forms, it just drifts somewhere it used to be swept.

• Measure airflow or differential pressure across the filter bank at commissioning with a clean filter. That is your baseline.
• Check on a fixed interval and after every change; a magnehelic gauge or logged flow reading beats a fan that merely sounds normal.
• Set the change interval from real arc-on time, not calendar days. A cell running 12 hours a day, six days a week loads a filter three or four times faster than one that idles.
• Keep spares on site and log every change: date, filter identity, differential pressure and airflow before and after. The log is what proves the system was maintained when an inspector asks.

Filters Are a Backstop, Not a Control

A filter does not capture fume; it cleans air that has already been captured. If the hood sits on the wrong side of the arc, if the robot parks where the plume drifts, if the discharge blows back into the cell, no media will fix it. Filters are the last line of defence in a hierarchy that puts enclosure and local exhaust first, exactly as the NIOSH hierarchy of controls describes. Get the hood right, size it from the far point of the robot envelope, and only then choose media: filters rated for a flow the fan cannot deliver at the hood are money spent on the wrong problem.

Respirators sit downstream of all of this. Under OSHA 1910.134, protection against particulates requires an air-purifying respirator with a HEPA filter certified under 30 CFR Part 11, or certified for particulates by NIOSH under 42 CFR Part 84, and filtering facepieces or elastomeric half masks do nothing about the gas fraction. HSE’s reality check: work not expected to exceed one hour may be adequately controlled against particulates by an FFP3 tight-fitting disposable mask or a reusable half mask with a P3 filter, if the wearer is clean shaven. Adequate, and for an hour, and for particulates only.

Filters also touch the paperwork. 1910.252(c)(1)(iv) requires hazardous materials used in fluxes, coatings, coverings and filler metals to sit in your Hazard Communication programme, with access to labels and safety data sheets and training under 1910.1200. And 1910.252(c)(1)(v)(B) requires filler metals and fusible granular materials to carry the notice that they must not be used in areas without adequate ventilation, which points to ANSI Z49.1.

Consumables matter too. Flux-cored and gas-shielded MIG differ hugely in fume volume and spatter load on a spark screen, so if the cell runs aluminium flux-cored work, read the aluminum flux core welding guide alongside the ventilation plan.

What the Regulations Actually Require: OSHA, HSE, ANSI/AWS and ACGIH

Ventilation in a robot cell is not a nice-to-have line item on a layout drawing. Where your plant sits decides whether an undersized or missing extraction system is a legal breach, and the documentation trail often matters as much as the hardware.

United States: OSHA

Start with welding, cutting and brazing, covered by 29 CFR 1910.252. Paragraph (c) is headed “Protection of employees,” and (c)(1) is blunt about the intent: hazardous fume should be removed as close to the source as practical, with general ventilation handling only what local exhaust misses. Paragraph (c)(3) governs capture systems, covering hood and booth design, exhaust system location and discharge, and requiring a minimum air velocity of 100 feet per minute through the hood, with roughly 1,000 feet per minute recommended in the branch ducts.

Material-specific rules are where robot cells get caught. 1910.252(c)(6)(ii) requires that indoors, welding or cutting involving zinc-bearing base or filler metals, or metals coated with zinc-bearing materials, follows the local exhaust hood and booth requirements of (c)(3). Galvanised parts and zinc-coated wire consumables therefore force a genuine LEV design in an indoor cell. Lead, cadmium and mercury each have their own paragraph with similar local exhaust expectations, and cutting stainless steel is called out separately because the fume profile changes sharply. Consumable choice is part of this too, and how alloy selection changes the electrode and fume behaviour is worth understanding before you finalise the cell spec.

Beryllium is the sharpest provision. 1910.252(c)(8) requires welding or cutting involving beryllium-containing base or filler metals to be done using local exhaust ventilation and airline respirators, unless atmospheric tests under the most adverse conditions establish that exposure stays within acceptable concentrations.

Respirators are the fallback, not the plan

1910.134(a)(1) states that the primary objective in controlling occupational diseases caused by breathing contaminated air shall be to prevent atmospheric contamination, accomplished as far as feasible by accepted engineering control measures such as enclosure, general and local ventilation, and substitution of less toxic materials. Respirators come in only when effective engineering controls are not feasible, which is the same logic as the NIOSH hierarchy of controls. A cell with no capture hood cannot jump straight to a half mask.

If respiratory protection is used, the assigned protection factor in Table 1 of 1910.134 decides what you can rely on:

Respirator type	Assigned protection factor
Filtering facepiece (half mask)	10
PAPR with a half mask	50
PAPR with a full facepiece	50
Helmet or hood PAPR, and continuous-flow supplied-air helmet or hood	1,000, where the employer holds manufacturer evidence of testing at that level
Other PAPR and SAR helmets and hoods without that testing evidence	25, treated as loose-fitting facepiece respirators

On top of the ventilation and respirator rules sits 1910.252 in combination with the hazard communication standard 1910.1200: safety data sheets for the wire, flux and gases, container labelling, and training so operators loading, grinding and touch-up welding inside the enclosure know what is in the fume.

United Kingdom: COSHH and HSE

In the UK the governing law is COSHH, the Control of Substances Hazardous to Health Regulations 2002. Welding fume is a substance hazardous to health, and under the control ladder you must eliminate or reduce exposure at the source, which means adequate local exhaust ventilation well ahead of personal protective equipment. HSE’s welding control measures guidance points to capture velocities of roughly 0.5 to 1 metres per second as the normal design band for welding fume extraction, the metric counterpart to the OSHA figure in 1910.252(c)(3).

The recurring obligation is examination: HSE requires local exhaust ventilation to be thoroughly examined and tested at least every 14 months, and controls maintained in an efficient state, in efficient working order, in good repair and in clean condition. Frequencies and detail sit in HSE’s guidance on maintaining and testing welding control measures. For a robot cell this is where record keeping bites: filter changes, airflow readings and hose inspections all feed that 14-month cycle.

HSE also flags the robotic case specifically. The person nearest the fume is usually not the person who would have welded by hand, so an assessment built on a manual workstation is wrong. The exposure picture belongs to the operator at the load station, the grinder working between cycles and anyone doing touch-up welding inside the enclosure.

ANSI and AWS

ANSI Z49.1, Safety in Welding and Cutting, is the consensus standard OSHA’s welding rule was built on and is referenced directly in the regulation, so it is where you look for the reasoning behind a requirement, not just the requirement.

AWS carries its own safety standard, listed by the organisation as AWS A3.0M/A3.0:2025 within a portfolio of more than 350 authored standards. It is the reference for welding safety practice in automated cells, and AWS also offers the Certified Robotic Arc Welding credential, worth considering if you are building internal competence rather than buying in support. Where certification pathways matter for your own team, how AWS standards and credentials fit into regulated production work is useful background.

ACGIH: guidelines, not law

The limits quoted in fume datasheets come from ACGIH. As ACGIH explains, threshold limit values and biological exposure indices are health-based guidelines developed by a committee of scientists, not consensus standards and not legal limits. They are excellent engineering targets, and many corporate limits are set at or below them. But the enforceable duty sits with OSHA permissible exposure limits in the US and COSHH limits in the UK. Design to the enforceable limit, use the ACGIH figure to add margin, and keep the two separate in your records.

Common Mistakes That Make Robot Cell Ventilation Fail

Most robot cell ventilation problems are not exotic. They are the same handful of decisions, repeated badly, in cells commissioned in a hurry: a hood sized for the closest reach instead of the furthest, a FumeVac arm parked where the robot needed it, a filter swapped out “when it looked bad” with no airflow check afterwards. Each one is a management failure with a cheap fix and an easy way to verify it.

Sizing and positioning errors

• Sizing for the nearest point, not the most remote one. The robot works over one spot, then stretches to the far corner of the envelope. Size for the first and capture collapses in the second. Fix: size from the last remote point of the weld path. Verify: watch where the plume drifts with the arm extended.
• Mounting the hood where the robot hits it. A hood that collides with the torch, dress pack or a heavy part gets bumped, the interlock trips, and somebody disables it. The cell then runs with a hood tilted, dented or switched off. Fix: map the arm envelope and positioner rotation before choosing a mount. Verify: dry-run every program.
• Curtains and screens left in place. A spark curtain or non-combustible screen can sit between the hood and the plume, or block the flow path out of the cell. Fix: design screens as part of the ventilation layout, following OSHA 1910.252(b)(2)(iii), which requires the welder to be enclosed in an individual booth with a low-reflectivity finish or behind noncombustible screens, while booths and screens must still permit circulation of air at floor level. Verify: test capture with the screen in its working position, not removed.

Wrong assumptions about the air in the cell

• Using general dilution as the primary control. Fans pushing air across a cell with a low roof or blocked cross-flow do not remove fume, they distribute it. Worse with fume-heavy processes such as stick welding, and always worse in an enclosed cell. Fix: treat general ventilation as a background layer and capture at the arc. Verify: release a tracer at the far point and confirm it does not drift toward the operator.
• Assuming filtered discharge inside the cell is harmless. Filters remove particulate, not gases, and a dirty prefilter can puff loaded air into the cell on start-up. Fix: duct the discharge away from the breathing zone and provide a deliberate make-up air path. Verify: measure at head height through the worst cycle and confirm the discharge is not returning.
• Running the fence closed with no make-up inlet. A sealed cell is a negative pressure chamber. The fan fights the gap, airflow drops, and the interlock trips or gets bypassed. Fix: size and install a proper make-up inlet. Verify: confirm the cell holds slightly negative pressure against a measured inlet, not an improvised gap in the fence.

Filters, people and paperwork

• Never re-checking airflow after the filter loads. A loaded filter raises resistance and drops flow, so the hood quietly delivers less than the design capture velocity while the fume keeps coming. Fix: set a fixed check interval tied to real duty cycle. Verify: compare with the commissioning baseline, change the filter, confirm flow recovers.
• Ignoring the loaders, grinders and gougers. The robot arc is captured, but the person unloading, grinding the part or gouging out a bad weld stands in residual particulate. Grinding dust and carbon arc gouging fume are hazards the welding hood does not catch, and aluminium work adds cleaning dust, as the aluminium MIG welding guide explains. Fix: extend control to the whole cell and place the loader inside the extract. Verify: watch a full load-and-grind cycle and check where dust settles. Give those workers training and safety data sheets under OSHA 1910.1200.
• Treating residual particulate as zero and skipping RPE. This one has legal consequences. HSE is explicit that even when the welder works inside the capture zone, residual risk remains as uncaptured particulate, so additional respiratory protective equipment would be required. Fix: keep a full 1910.134 programme with medical evaluation, fit testing and training as the backstop. Verify: audit the RPE records, not just the hood.
• No inspection regime at all. Nothing is checked until a welder complains. Fix: put local exhaust ventilation on a written inspection and examination schedule. Verify: keep a dated log of airflow, filter condition and visual capture checks.

Work through this list in order and most cells come right quickly, because the failures compound: a poorly sized hood loses capture, operators nudge it aside, the filter loads, the interlock gets bypassed. Fix the geometry first, then the airflow, then the paperwork.

Comparing Hood Types: Mobile Units, Fixed Booths, On-Torch Extraction and Downdraft

No single hood wins in every robotic cell. The answer depends on how the robot moves, part size, how often the job changes and how many people share the space.

1. Portable mobile extractor with an articulated arm

A filter-and-fan cart with a flexible arm: the cheapest capture for a cell, no structural work, and it moves as the robot parks in different positions. It protects load-up and grinding staff inside the enclosure, but a six-axis robot knocks the arm over and someone ties it out of the way, making it decorative. It cannot follow a moving torch, so on a 10-foot fixture it captures one end and nothing at the other.

2. On-torch extraction with a hose to a central filter

The hood mounts on the torch, with a hose back to a central filter and fan, so capture happens at the source.

Upsides: Capture follows the arc, so design distance stays constant wherever the robot moves. HSE notes that on-torch extraction can be very effective for MIG welding, and that carries over to a robotic MIG cell. No operator intervention is needed.

Downsides: The hose has weight: on a light torch it pulls the torch out of position and skews the weld, and it adds drag to dress pack cables that already limit reach. The central filter must serve the whole cell, so expect real capital cost and a filter heavy enough to need a change crane or a two-person lift. It handles particulate, not gases, so flux-cored aluminium needs a discharge route filter media alone cannot satisfy.

3. Fixed hood over the workpiece or fixture

A permanent hood above the fixture, sized so the plume is drawn in. The geometry never changes, so capture is repeatable with no operator interaction and nothing to knock over, which suits a cell running the same part for months. It works only while the weld path stays in the capture zone: a longer part or a new product puts the far end of the seam outside the hood and capture quietly fails. It also counts as part of the cell’s safety envelope, and it sits where operators need to reach for load and unload.

4. Full downdraft booth or enclosure with roof extraction

A booth draws air down through the work area and out through the roof or a side plenum, treating the cell as a contained volume. It captures everything, so grinding dust, spatter, wire grind-off and coating debris are handled by the same system, making it the most robust answer for mixed processes. As a fixed enclosure it satisfies the fixed-enclosure ventilation requirements, and a sealed cell is easier to interlock and keep clean.

It is the most expensive option by a wide margin and takes the most floor space. It also introduces heat, since extracted air has to be replaced, and you need a make-up air path sized so extraction is not starved, often meaning ducting from outside or a designed inlet. A booth is also a one-way door: once the process or part mix changes, it may be the wrong shape for the work.

5. Dedicated cell extraction with a general ventilation layer as backup

For most robotic cells this is the combination that holds up: purpose-built extraction sized from the most remote point of the weld envelope, plus a general ventilation layer that handles the residual and protects the rest of the workshop. Extraction is never perfect, so the general layer covers the gap between the hood working and the hood doing nothing, and it gives a graceful failure mode, since the cell still has air movement if extraction stops and the interlocks can stop the robot.

For reference, HSE points out that resistance spot welding produces much less fume than MIG, and that good general ventilation is usually enough for low-intensity spot welding on clean uncoated metal, so where the work is not MIG the general layer may carry more of the load.

The comparison at a glance

Type	Best at	Weakest point	Capital
Mobile unit with arm	Cheap, flexible coverage	Knocked over, cannot follow a moving torch	Low
On-torch extraction	Capture at the source	Hose weight and drag on robot cabling	Medium to high
Fixed hood	Repeatability, no operator involved	Fails when the weld path leaves the zone	Medium
Downdraft booth	Arc plus grinding and spatter	Cost, floor space, heat, make-up air	Highest
Cell extraction plus general layer	Resilience across changing work	Needs careful design and commissioning	High

A decision framework

Ask four questions in order.

1. How variable is the work? A fixed hood or booth pays back on repeat parts and is wrong within a quarter if the cell changes weekly.
2. How big are the parts? Small parts suit a fixed hood; long parts need capture that travels with the arc.
3. What processes run in the cell? Grinding, coated stock or aluminium mean handling dust and gases, not just the plume.
4. What labour do you have? A design that runs a year between filter changes beats a cheaper system needing weekly attention.

On discharge, HSE is explicit that LEV discharging inside the work area must be filtered to reduce particulates to a negligible concentration, while air discharged outside should leave the building at a high enough level to be dispersed, and environmental legislation can apply on top. For mixed-material cells, where aluminium or stainless particulates are hard to filter, a filtered discharge inside the cell is safer and should drive the layout.

On sizing, HSE describes extracted benches as good for medium and smaller jobs, and for larger articles points to booths big enough to house the workpiece on a turntable so fume is drawn to the extraction at the back. A robotic cell is closer to that second description.

Mobile units buy you time, not compliance. Fixed hoods buy repeatability. Booths buy total coverage at a price. Cell extraction with a general ventilation backup costs more up front and usually wins on total cost of ownership, because it never depends on someone remembering to swing an arm into position.

Our guides to aluminium MIG welding and aluminium flux core welding cover the fume composition differences that change your filtration requirements.

Putting It Together: A Specification and Rollout Checklist

You now have the theory: what the fume is, how a hood captures it, how OSHA’s capture table translates into duct size and airflow, where the hood has to sit, and what the law demands. What follows is the part you can actually hand to an engineer, a fume extraction vendor or a plant manager on a Friday afternoon and say “build me this.”

Work through it in order. Skipping step one is how cells end up with a beautiful booth that captures the wrong fume.

Step 1: Characterise the Job

Before choosing equipment, write down five things on one sheet of paper:

• Materials. Mild steel, stainless, galvanised, aluminium or coated sheet each produce a different fume. If the cell runs stainless, the fume composition and any hexavalent chromium content change the filtration and respiratory assessment. If it runs galvanised, the zinc oxide fume drives a whole separate exposure concern.
• Consumables and process. Solid wire, flux core wire, MIG, TIG or laser all change fume volume and particle characteristics. The filler metal matters as much as the base metal. If aluminium is in the mix, flag the gas and fume differences early, because the reactivity of aluminium fume is one of the reasons aluminium MIG welding demands its own controls.
• Arc-on duty cycle. Ask the robot controller for arc-on time per cycle and cycles per shift. A cell welding 60 percent of the time generates roughly three times the fume of a cell welding 20 percent of the time, over the same shift. This number drives filter life, fan sizing and cooling load.
• Weld path envelope. Map the full three-dimensional volume the torch sweeps, including approach and retract strokes, not just the seam itself.
• Part geometry. Large fabrications with long seams, deep boxes and clamped parts all change how the plume rises and where the hood can physically go.

Step 2: Choose the Capture Architecture

Pick the arrangement that matches the job’s variability. A cell welding one part family at one station usually earns a fixed hood over the workpiece. A cell handling varied sizes across a wide envelope earns on-torch extraction or a hood on a track that follows the arc. A cell with heavy grinding, spatter and unloading alongside welding earns a downdraft booth with general ventilation as the backup layer. Document the choice and the reason, because a reviewer will ask later.

Step 3: Size From the Far Point

Size the hood from the most remote point of the weld envelope, not the typical or nearest point. Take the required airflow and duct diameter from OSHA’s capture table for welding fume, then apply a system curve at the loaded filter condition rather than quoting free-air fan capacity. A fan that reads well on the shop floor with a clean filter will not read the same once the cartridge has loaded and the duct losses have grown. Ask the vendor for airflow at the hood at end-of-filter-life, not at the fan inlet. If the system cannot hold the design flow, the answer is a bigger fan or a larger duct, not a bigger filter.

Step 4: Plan the Discharge

Where does the filtered air go? If the discharge re-enters the cell or the operator’s breathing zone, you are recirculating the problem. Discharge to a safe location outside the enclosure, or to a general ventilation system that handles the load. Then draw the make-up air path. A closed cell fence with continuous extraction and no make-up air is a negative-pressure cell that pulls unfiltered air through every gap, including the fence. Confirm the make-up air enters away from the discharge and does not sweep across the arc.

Step 5: Integrate With the Cell Safety System

Ventilation that stops working quietly is worse than no ventilation, because operators learn to ignore it. Wire an airflow or differential pressure indication into the robot’s safety circuit so the cell refuses to start the arc, or stops mid-cycle, when extraction drops below the set point. Show the status on the operator screen and on the pendant. Give operators a clear, visible indicator of filter condition rather than a gauge they have to find.

Step 6: Write the Respiratory Protection Fallback

Extraction reduces the hazard; it does not remove it entirely. Even a welder working inside the capture zone leaves residual risk, so a respiratory protection programme is often the safety net behind the hood.

If you are in the United States, 1910.134 governs this, and it is more than a box of masks. Under 1910.134(c)(1) you need a written respiratory protection programme with worksite-specific procedures covering respirator selection, medical evaluation, fit testing, use in routine and emergency situations, maintenance and storage, training, and regular evaluation of the programme’s effectiveness. Under (c)(3) you must designate a programme administrator qualified by training or experience commensurate with the complexity of the programme. Under (c)(4) respirators, training and medical evaluations must be provided at no cost to the employee. Under (e)(1) the medical evaluation must determine an employee’s ability to use a respirator before that employee is fit tested or required to wear one in the workplace.

In the UK, HSE recommends for arc welders a powered air respirator or a supplied air respirator combining respiratory, eye and face protection, with an assigned protection factor of at least 20, as the best option. The reasoning is practical: tight-fitting RPE gets uncomfortable enough that workers loosen or remove it, and a loose mask protects nobody. HSE also notes that welding curtains and screens should conform to BS EN ISO 25980 or offer the same level of protection, and that dark-coloured wall coatings reduce reflections from UV light.

Step 7: Commission and Document

At commissioning, measure and record: hood position and the arc positions it was tested at, measured airflow at the hood, filter baseline pressure drop, interlock function, and the arc-on duty cycle assumed in the design. Keep the sheet. It becomes the reference every future fault is measured against, and in a US inspection or a UK thorough examination it is the first document anyone asks for. Note the SDS for every material and consumable in the cell so1910.1200 hazard communication duties are covered.

Step 8: Set the Maintenance and Examination Regime

• Filter changes on duty cycle, not calendar. Tie the interval to arc-on hours and to measured pressure drop, not to a habit of changing on the first of the month. Keep spares on site so a change never waits for a purchase order.
• Airflow checks on a set schedule. Check at commissioning, then at regular intervals, and record the readings so a falling trend is visible before capture is compromised.
• UK: thorough examination and testing at least every 14 months for local exhaust ventilation, per HSE’s guidance on maintaining and testing control measures.
• Keep the filters as a backstop only. They are the last line, not the first. If a cell relies on filter quality rather than capture, the design is wrong.

Step 9: Train and Involve Operators

The people who load parts, grind between cycles and do touch-up welding inside the enclosure need to know how the extraction works, where it fails, and why they must not wedge a hood out of the way. Include them in the commissioning walk-through and in filter change training. Operators who helped design the process will defend it once you are not in the room. This is also where NIOSH’s hierarchy of controls earns its place: ventilation sits in the engineering control band, above administrative controls and PPE, because it removes the hazard before the worker meets it. Every shortcut around the hood pushes exposure downward the hierarchy into the controls that protect the individual last and least well.

The Bottom Line

A robot moves the arc out of a human hand. It does not move the fume out of the building, and it does not reduce the hazard. In a robotic cell, ventilation has to be engineered around a moving, unattended arc: sized from the far point of the envelope, positioned where the robot and the part cannot knock it, verified at the hood rather than at the fan, and interlocked so the cell cannot weld without it. Get those three things right and a correctly sized, correctly placed fume extraction hood is the only control that captures welding fume at the source, keeps it out of the breathing zone, and lets the productivity gain from automation stand on something other than deferred exposure.

Frequently Asked Questions

Do robotic welding cells actually need ventilation if there is no welder standing at the arc?

Yes. The fume is still generated at the same rate per unit of arc-on time, and robotic cells often run longer arc-on cycles with less time for fume to settle. The exposure simply moves from the welder’s breathing zone to the cell interior and to operators who load parts, grind between cycles and touch up welds inside the fence. OSHA 1910.252(c) does not exempt automated processes from ventilation requirements.

How far should a fume extraction hood be from the welding arc in a robot cell?

As close as the geometry allows, and in every case close enough that OSHA’s control velocity of 100 feet per minute is still achieved at the most remote point of the work envelope, not just at the closest. The OSHA table gives minimum airflow of 150 cfm at 4–6 inches, 275 cfm at 6–8 inches, 425 cfm at 8–10 inches and 600 cfm at 10–12 inches. Design the reach from the robot’s far limit, and verify airflow with the filter loaded.

Is a mobile fume extractor enough on its own for a robotic welding cell?

Usually not. Mobile units with swing-arm hoods work well for manual welding and for maintenance touch-ups, but they depend on someone positioning them, and in a robot cell nobody is standing there to do it. Most successful installations pair source capture (on-torch extraction, a manipulator-mounted hood, or a fixed booth) with mechanical general ventilation for background clearance and residual fume.

Do I still need respirators if I have a fume extraction hood?

Possibly. OSHA 1910.134(a)(1) puts engineering controls first, but respirators are still required where engineering controls are not feasible or are not fully effective, and HSE notes that even inside the capture zone there is residual risk from uncaptured particulate. If RPE is used, you need a written program with medical evaluation, fit testing, training and maintenance under 1910.134.

How often should filters on a robotic cell extractor be changed?

Base the interval on real arc-on duty cycle rather than a generic calendar figure, and check airflow or differential pressure at commissioning and at a set interval, because a loaded filter raises resistance and quietly reduces capture. Keep spare cartridges or bags on site, log every change, and remember that particulate filters remove dust but not welding gases — those need separate treatment.

What ventilation and inspection duties apply if my plant is in the UK rather than the US?

Welding fume is a COSHH-regulated substance, so the UK route is risk assessment, a prioritised control ladder ending in RPE, and local exhaust ventilation where capture at source is needed. LEV must be thoroughly examined and tested at least every 14 months, with capture velocity judged against the 0.5–1 m/s range, and measured against effective fume capture rather than airflow numbers alone.

Does galvanised or stainless steel need different ventilation from plain mild steel?

Yes, and the rules say so explicitly. OSHA 1910.252(c)(6)(ii) requires indoor welding of zinc-bearing or zinc-coated materials to use the local exhaust hood and booth provisions of paragraph (c)(3), and 1910.252(c)(7) does the same for lead-bearing metals, with workers in the immediate vicinity protected. Stainless fume carries hexavalent chromium, nickel and cobalt, which HSE links to occupational asthma and lung cancer risk.

Can the robot cell’s welding curtains or screens be used as the ventilation control?

No — they serve a different purpose: protecting people from arc light and UV. OSHA 1910.252(b)(2)(iii) requires booths and screens to permit circulation of air at floor level, and HSE notes that curtains conforming to BS EN ISO 25980 are for arc-light protection. In practice a screen positioned across a hood’s path can destroy capture performance, so keep the two systems physically separate.