Most duct shops add a plasma table to cut blanks and stiffener notches, and almost every duct blank is galvanized. That combination is what makes dust collection a design decision rather than an accessory purchase. When the arc cuts through zinc, part of the coating vaporizes and comes off the plate as zinc oxide fume. It is a fine, white, breathable particle, and if you breathe enough of it you get metal fume fever, which feels like a sudden flu and puts your best operator in bed for a day. In several markets, welding and cutting fume exposure limits have tightened over the last decade, so the extractor you buy with the table is not a nice-to-have. It is part of the cutting cell.
I have visited shops where the plasma table ran with nothing but a roof fan. The white dust on the rafters told the whole story. The table itself still cut fine, but the shop failed an insurance inspection, and the cleanup bill was bigger than a proper dust collector would have cost. So before we talk about filter types and airflow numbers, accept the premise: the collector is sized for zinc oxide fume and fine particulate, not just coarse slag dust.
There are two mainstream ways to capture plasma fume on a duct fabrication line. A downdraft table pulls air down through slotted panels across the cutting surface, so fume is captured at the source as the torch moves. A full enclosure or side-draft booth surrounds the cutting zone and moves a larger air volume across the whole plume. Most duct shops with a plasma cutting cell integrated into the line end up with a downdraft table, because the table is already part of the material flow: sheets come in, blanks go out, no extra handling.
The sizing rule of thumb I use is capture velocity at the table surface. You want roughly 100 to 150 feet per minute of downward air velocity across the open table area, a bit more if you cut near the table edges often. On a 2 by 6 meter table that works out to roughly 12,000 to 18,000 cubic meters per hour of airflow depending on how much of the surface is open and how the plenum is baffled. If a supplier quotes a collector at half that number for a full-size table, ask them to show the capture velocity calc, because fume will escape around the torch and you will end up running the shop doors open to dilute it, which defeats the point.
Two tables with the same fan can capture very differently. The plenum under the slats needs baffles that balance flow along the table length; otherwise the zone near the fan hogging the airflow leaves the far end dead. On one job in Lagos the far third of the table collected visible fume until we added two adjustable dampers in the plenum and rebalanced. Thirty minutes of work, and the capture improved more than a fan upgrade would have. When you do your site acceptance, smoke-test the far corners of the table, not just the middle where the supplier demonstrates.
The dust from plasma cutting galvanized is fine and mostly non-sticky, which is exactly what cartridge collectors handle well. Bag houses are the traditional choice for heavier dust loads; wet collectors solve the spark and combustible dust problem at the cost of sludge handling. Here is how they compare for a typical duct shop running one plasma table:
| Type | Typical airflow | Filtration | Strengths | Watch-outs |
|---|---|---|---|---|
| Cartridge collector | 10,000 to 20,000 m³/h | 99.9% at 0.5 µm with pleated media | Compact, pulse-cleaned, low ΔP drop-in media | Needs spark arrestor; media blinds if oil mist present |
| Bag house | 15,000 to 40,000 m³/h | 99%+ at coarse range | Handles heavy dust load, cheap media | Large footprint, weaker on sub-micron zinc fume |
| Wet collector | 8,000 to 15,000 m³/h | 99.7% with mist eliminator | Spark-safe by design, good for aluminum too | Sludge disposal, water treatment, higher running cost |
For galvanized duct blanks on one table, my default recommendation is a pulse-jet cartridge collector with a spark arrestor box in front of it. Reserve wet collectors for shops cutting a lot of aluminum, where the dust is both abrasive and flammable, or where the local fire code forces the issue.
A plasma arc throws molten droplets. In a downdraft system those droplets get pulled straight into the ductwork toward your filter media, and a cartridge fire is a spectacular way to lose a week of production. The defenses are layered and none of them is expensive. A spark arrestor or drop-out box slows the air and lets heavy sparks fall out before the collector. A fire-retardant pre-filter layer catches the strays. Inside the collector, a sprinkler or a suppression cartridge is worth pricing if your insurer asks, and more insurers are asking.
Housekeeping sits in the same bucket. Plasma dust from galvanized sheet is not classified as combustible the way aluminum or steel-grit dust can be, but mixed shops generate mixed dust. If the same collector serves a grinding bench, check what the local code says about combustible dust thresholds before you combine the duct runs.
Pulse-jet collectors clean their own cartridges with compressed air blasts, controlled either by a timer or by the pressure drop across the media. Set the cleaning to trigger on ΔP, not on a blind timer. A healthy cartridge set on galvanized plasma dust runs at maybe 800 to 1,200 pascals of pressure drop. When ΔP climbs past roughly 1,800 to 2,000 pascals and stays there after a pulse cycle, the media is loading up. Running past that point costs you fan energy and capture velocity at the same time, because the choked filter pulls less air, and less air means fume escaping the table.
The schedule that works in real shops: check ΔP daily for a week to learn your baseline, blow down the pulse system and drain the moisture trap weekly, and change cartridges when ΔP will not recover, typically every 9 to 14 months on single-shift plasma duty. If your ΔP jumps suddenly rather than drifting, look for a torn cartridge or a moisture problem in the compressed air before you buy new media. Wet compressed air pulsed into a dry dust cake makes paste, and paste is the number one cause of premature cartridge death I see in tropical shops.
Once the system is running, verify it. A cheap photometer or even smoke pencils around the table edge during a cut will show escaping fume. For formal peace of mind, an occupational hygienist sampling zinc oxide at the operator station once a year costs a fraction of one claim, and in several countries it is becoming a requirement rather than a suggestion. Keep the log with your maintenance records; it also makes resale and audits easier.
Two running costs catch buyers off guard. First, the pulse cleaning needs clean, dry compressed air, typically 0.5 to 0.8 cubic meters per minute at 6 bar for a mid-size collector. If your shop compressor is already at its limit, the collector and the plasma will fight over it, and weak pulses mean dirty cartridges. Size the compressor with the collector included, not as an afterthought. Second, the collector fan is a noise source, usually 80 to 90 decibels at the inlet. Mount it outside or in an acoustic enclosure if the collector sits near your office or your folding benches, because your crew already has enough noise exposure from the shears and the plasma itself.
Can I duct the plasma collector together with my welding stations?
Technically yes, but I advise against it on a duct line. Welding fume is finer and often oilier, and combining duct runs means your whole collection system has to meet the strictest classification of everything connected to it. Separate collectors for the table and the welding bays cost a bit more up front and save years of blame-shifting.
How often do cartridges really need replacing on galvanized?
Single shift, properly pulse-cleaned, 9 to 14 months is the honest range. Two shifts roughly halves it. The best predictor is not the calendar, it is the ΔP trend; replace when the pressure drop will not come back below your baseline after a full pulse cycle.
Is zinc fume from galvanized really a health issue, or just unpleasant?
It is a real issue. Metal fume fever from zinc oxide exposure is well documented, and repeated exposure is not something to write off. Beyond health, fume escaping the table settles on your finished duct, on your machine guideways, and in your painting area if you have one. Capture at the table is cheaper than cleaning zinc dust out of everything downstream.
Do I need the collector running for every cut, even short jobs?
Yes. Fume generation starts with the arc, not with the job length. The short notches and corner cuts that happen all day on a duct line are exactly the cuts operators skip the collector for, and they are enough to keep the shop dusty. Interlock the collector with the table so it cannot be forgotten.