Ninety-nine point nine seven percent. That's what a HEPA H13 filter catches at zero point three microns. And the dust coming off a PCB milling cutter is mostly smaller than that. So if you're running a shop vac with a standard filter, you've built a fibreglass aerosol dispenser.
And the person who told you to get a small vacuum didn't mention any of that.
Here's what Daniel wrote in. He's setting up an electronics bench, fine work, PCB milling and engraving, drilling FR-4, desoldering, rework, general precision assembly. And he keeps getting the same advice. Get a small vacuum. That's it. That's the whole instruction. And his point is that the word vacuum is quietly doing three different jobs, and nobody says which one they mean. One, particulate capture at the source. Milling and engraving throws off fine fibreglass and resin dust, drilling makes swarf. That's a capture-at-the-cutter problem, not a clean-up-afterwards problem. Two, fume extraction. Solder flux fume is not dust. It's not vacuumed at all. It's drawn through activated carbon or ducted outside. Different machine. Three, electrostatic discharge. An ordinary vacuum is mechanically a static generator, air and particles moving fast through an insulating plastic tube. And if the thing you're cleaning is an assembled board, that's a problem.
Three machines wearing one name.
And then he wants the practical survey. What actually works at each budget tier, from a bench-top mini-vac with fine nozzles, through hobbyist fume extractors, up to what a small production bench runs. Which nozzles matter. Where the compressed-air instinct is actively wrong. So today we're separating those jobs, and then we're going to look at what actually works at each tier.
The thing that keeps getting conflated here is physics. Dust is a solid. Fume is a gas. Static is a charge. Those are three different states of matter and one of them isn't matter at all. But the marketing doesn't care. A shop vac is sold as a universal tool. A bench vacuum is sold as a smaller universal tool. And the failure modes are invisible until you've already contaminated your shop or zapped a board.
The zap is the only one that announces itself immediately. The dust problem takes years. The fume problem takes years and you can't see it happening.
Right. And that's why the instruction is so bad. Get a small vacuum presumes the problem is one thing. But if you're milling boards, your problem is a cloud of sub-ten-micron fibreglass particles being generated at the cutter. If you're soldering, your problem is vaporised flux, mostly colophony, which is rosin, and it's a respiratory sensitiser. If you're cleaning an assembled board, your problem is that the act of cleaning might destroy the board.
So let's start with the first job hiding in that word. The dust that comes off a milling cutter.
FR-4 is fibreglass cloth impregnated with epoxy resin. When a milling bit cuts through it, it doesn't make shavings like wood. It makes a fine dust. Some of it is visible, the swarf from drilling is little curls of copper and fibreglass, but a lot of it is below ten microns. That's the stuff that stays airborne for minutes or hours. And it's not just nuisance dust. Fibreglass is a mechanical irritant, the resin component has been linked to respiratory problems with chronic exposure, and there's enough in the literature to treat it as a potential carcinogen over years of daily exposure.
So this is not about keeping the bench tidy.
It's not about keeping the bench tidy. It's about what you're breathing while you run the mill. And here's the part most people miss. The dust is being generated at the cutter. The bit is spinning at ten, twenty thousand RPM. It's throwing particles in every direction. If you wait until the job is done and then vacuum the bench, you've already inhaled the worst of it. Capture has to happen at the source, while the dust is being made.
Which is why the ordinary shop vac fails in a specific way. It's not that it lacks suction. It has plenty. It's that the filter is a screen door.
Standard shop vac filters are typically rated to catch particles down to maybe five or ten microns, and only if the filter is in good condition and seated properly. The dust from FR-4 milling is often smaller than that. So the vacuum pulls in the air, catches the big stuff, and then exhausts the fine dust right back into the room through the motor cooling path. The machine becomes a distribution system. You can watch this happen. Someone runs a shop vac next to a mill for ten minutes and then a haze forms around the machine. That's the sub-micron fraction that got through.
And the motor itself is part of the problem. Shop vacs cool the motor with the exhaust air. So the fine dust passes through the motor, gets heated, and comes out as a warm aerosol. It's not just recirculated, it's conditioned.
Which is why the filter class matters so much. HEPA H13 is the standard you want. It captures ninety-nine point nine seven percent of particles at zero point three microns, which is the most penetrating particle size. Anything smaller or larger is actually captured more efficiently. H13 or higher. Not HEPA-type, not HEPA-like, not washable HEPA from the hardware store. Actual H13 with a sealed gasket.
And the sealed construction is the part people skip. A HEPA filter in a leaky housing is a HEPA filter that's being bypassed.
If the housing leaks, air takes the path of least resistance around the filter. So you need a sealed canister. And then you want cyclonic pre-separation. That's the cone-shaped thing that spins the air and drops the heavy particles out before they reach the filter. It keeps the HEPA from clogging in the first ten minutes. Without it, you're changing a forty dollar filter every week.
And then the geometry of capture. You said capture at the cutter. How close does the nozzle need to be?
Closer than people think. The capture velocity of a vacuum nozzle drops off with the square of the distance. So a nozzle two centimetres from the cutter captures most of the dust. A nozzle ten centimetres away captures a fraction. At twenty centimetres it's basically a decoration. The raw suction power matters less than where the nozzle is. This is why the pro setups use a small enclosure around the mill with a downdraft table, or a dedicated mini-vac with a fine nozzle positioned right at the spindle.
So the bench-scale answer is either contain the whole thing or put the nozzle right on the cutter.
And for a bench, the enclosure is often the better answer. A small acrylic or polycarbonate enclosure around the mill, with a downdraft table pulling air through the base. The enclosure does two jobs. It keeps the dust from spreading, and it lets the extraction work in a smaller volume. You don't need a huge airflow if the volume is small. A HEPA-filtered downdraft table with a modest fan can do more than a big shop vac pointed vaguely at the cutter.
And the health angle here is the part that gets skipped. Hobbyists will spend four hundred dollars on a spindle and nothing on the air they breathe while it runs.
The johal.in article makes this point about fume extraction, that it's the most skipped thing on a hobbyist bench. But the same neglect applies to particulate. Someone milling FR-4 for a few hours a week, no mask, no extraction, is doing something to their lungs. It won't show up next month. It shows up in twenty years. I spent enough time in medicine to know that the respiratory system keeps a ledger.
A ledger with compound interest.
And fibreglass dust is a nasty entry in it. So for particulate, the summary is, HEPA H13 or better, sealed housing, cyclonic pre-separation, and capture at the cutter. Either a nozzle on the spindle or an enclosure with a downdraft. That's the first job.
But dust is only half the story. Because the other thing you're breathing at the bench is not dust at all.
Solder fume. And this is where the word vacuum really breaks down. You cannot vacuum a gas. A vacuum cleaner moves air and captures particulates. Solder fume is mostly vaporised flux. The solid part is smoke, and some of that is particulate, but the thing that irritates your throat and lungs is the gas phase. Colophony, rosin, the stuff that makes flux work. When it's heated to soldering temperature, it vaporises and then condenses into a fine mist of resin acids suspended in the air.
So a particulate filter does nothing to it.
A HEPA filter will catch some of the condensed droplets. But the vapour phase goes straight through. HEPA is mechanical filtration. It catches particles. It does not catch molecules. To deal with organic vapours, you need adsorption. Activated carbon. The carbon has an enormous internal surface area, and the vapour molecules stick to it. That's a completely different mechanism.
Which is why fume extraction is a different machine. And it's the one most likely to be skipped.
Because it's invisible. Dust you can see on the bench. Fume you can maybe see as a wisp of smoke, but mostly you just smell it. And the health effects are delayed. Colophony is a respiratory sensitiser. That means repeated exposure can make you more sensitive to it over time. Some people develop occupational asthma from it. The hobbyist soldering for an hour on the weekend is probably fine. The person at the bench for four hours a day is not.
And the capture distance thing is even more dramatic for fume than for dust.
Much more. The numbers here are stark. A fume extractor with a nozzle five centimetres from the solder joint captures something like ninety percent of the fumes. Move it to twenty centimetres and you're under thirty percent. At thirty centimetres it's nearly decorative. The fume rises in a thermal plume from the iron, and it spreads fast. You have to catch it before it spreads.
So the flexible arm is not a convenience feature. It's the whole machine.
The arm is the machine. A bench-top fume extractor is basically a fan, a carbon filter, and a positioning arm. The arm is what makes it work. You put the nozzle right next to the joint, you solder, the fume goes into the nozzle. If the arm is stiff or too short or you don't bother repositioning it, you've bought a fan that makes noise.
And then the carbon saturates.
The carbon saturates. That's the consumable people forget. Activated carbon has a finite capacity. Once the surface is covered in adsorbed molecules, it stops working. It doesn't clog visibly like a dust filter. It just stops adsorbing. And then the extractor is blowing the fume back at you while looking like it's working. The carbon needs replacement on a schedule. Depending on how much you solder, that's every few months to a year.
And some units have a pre-filter for the particulate smoke, so the carbon lasts longer.
Right. A cheap foam or paper pre-filter catches the condensed droplets and the larger smoke particles, and the carbon handles the vapour. But the pre-filter also saturates and needs cleaning or replacement. The whole thing is a consumable system, not a one-time purchase.
So that's job two. Now the third one. The one that actually kills boards.
ESD. Electrostatic discharge. And the thing to understand is that an ordinary vacuum is a static generator by design. Not by accident. By physics. Air moving fast through an insulating plastic hose creates charge separation. The particles rubbing against the hose wall, the air itself, it all generates static. The hose is plastic. Plastic is an insulator. So the charge has nowhere to go. It builds up on the hose, on the nozzle, on the tool.
And then you touch it to a board.
And then you touch it to a board. The charge finds a path through the nearest sensitive component. A MOSFET gate oxide is a few nanometres thick. It doesn't take much to punch through it. The discharge might be too small to feel. You won't get a shock. The component just dies silently.
What makes an ESD-safe vacuum different?
The entire airflow path is conductive or dissipative. The hose is made of a dissipative material, usually carbon-loaded plastic. The nozzle is conductive. The whole assembly is bonded to earth. So any charge that builds up is drained to ground instead of accumulating. It's not just a grounded cord. The cord being grounded doesn't help if the hose is still an insulator. The charge builds up on the hose.
The grounding has to be continuous from the nozzle all the way back.
Continuous. Nozzle to hose to body to ground. And that's why these things are expensive. It's not a marketing gimmick. It's a different set of materials and a different assembly process. A proper ESD-safe vacuum runs three, four hundred dollars and up. You can get dissipative hoses and nozzles as aftermarket parts, but retrofitting a standard vacuum is fiddly and you have to verify continuity.
When does a hobbyist actually need one?
Here's where the risk gets overstated. If you're cleaning bare boards before assembly, or you're cleaning around the bench, you don't need ESD-safe. Bare fibreglass doesn't care about static. The risk is with assembled boards, populated with sensitive parts. CMOS logic, MOSFETs, anything with a gate. If you're doing rework on assembled boards and you want to vacuum off the flux residue or the dust, that's when the ESD-safe vacuum matters.
And the cheaper mitigations?
A grounded wrist strap costs five dollars. An anti-static brush costs ten. If you're cleaning an assembled board, brush it gently with an anti-static brush while you're grounded. That avoids the static problem entirely. The vacuum is for when you need suction, and if you need suction on an assembled board, you need the ESD-safe version.
Then there's the compressed air instinct. The thing everyone reaches for.
The compressed air can is the worst of both worlds. First, blowing air across a surface generates static through the triboelectric effect. The air itself moving across the board creates charge. That's the same physics as the vacuum hose, except now you're pointing it directly at the components. Second, compressed air doesn't remove dust. It redistributes it. You blow the dust off the board, it goes into the air, and then it settles back onto the board. Or onto the next board. Or into your lungs.
The instinct is actively wrong. It's not just ineffective, it's counterproductive.
The canned air thing is even worse because the propellant can spit liquid if you tilt the can. Now you've got a board covered in a cold solvent residue. The whole practice is bad. Brush or vacuum. Don't blow.
Let's do the budget survey. Someone's setting up a bench and they want to know what to actually buy.
Start with the cheap end. A bench-top mini-vac with fine nozzles and brushes. Fifty dollars or so. These are small, low-power, and they're fine for loose dust and debris. Swarf from drilling, component leads, general cleanup. They're not HEPA, they're not ESD-safe, and they won't capture milling dust at the source. But for picking up the mess after the fact, they're better than a dustpan and brush.
The nozzles matter more than the motor at this tier.
The nozzles are the whole thing. A fine pointed nozzle for getting into tight spots. A brush attachment for sweeping dust off a board without scratching it. A narrow slot nozzle for running along the edge of a bench. The wide floor nozzle that comes with a shop vac is useless at this scale. It's designed for carpet.
The fifty dollar mini-vac is the cleanup tool, not the capture tool.
It's the broom. It's not the dust collector.
Then the fume extractor tier.
Hobbyist fume extractors run from about a hundred fifty to three hundred dollars. These are the bench-top units with a carbon filter and a flexible arm. Some have a pre-filter. The good ones have a decent fan and an arm that holds its position. The bad ones have a weak fan and an arm that flops over. You're paying for the arm.
The replacement carbon filters are an ongoing cost.
Budget for it. A replacement carbon filter is twenty to forty dollars depending on the unit. If you solder regularly, you're replacing it a few times a year. That's part of the cost of the machine, not an optional extra.
Then the professional tier.
Above five hundred dollars you're into proper rework station territory. These are the units with a real extraction arm, a HEPA or ULPA filter, a carbon stage, and often a built-in holder for the iron. They're designed for continuous use. The airflow is higher, the filtration is better, and the arm is actually engineered rather than a bendy straw. For a small production bench, this is the baseline.
The particulate side at the professional tier?
That's where you see the downdraft tables and the enclosed mills with integrated extraction. A proper downdraft table for electronics work runs several hundred dollars. It's a perforated work surface with a fan and a HEPA filter underneath. You work on the table, the dust is pulled down and trapped. For milling, you want the enclosure. For hand work, the downdraft table.
The ESD-safe vacuum is its own purchase, separate from all of this.
Separate machine. If you're cleaning assembled boards, a proper ESD-safe vacuum is four hundred dollars and up. Atrix and a few other companies make dedicated units. They come with dissipative hoses, grounded nozzles, and they're certified for electronics work. If you're not cleaning assembled boards, skip it. If you are, it's not optional.
The answer to get a small vacuum is, which problem do you actually have?
The answer is usually, more than one. If you're milling and soldering and doing rework, you have all three problems. That's three machines. Or two machines and a brush. The person who buys one vacuum and expects it to do everything is going to be disappointed in a way they won't notice until it's too late.
The dust problem announces itself slowly. The fume problem announces itself in twenty years. The ESD problem announces itself as a dead board with no visible cause.
That's the cruel part. The ESD failure looks like a bad component or a bad design. You don't blame the vacuum. You blame the supplier.
The framework is, capture at the source for dust, carbon for fume, ground path for static. Three jobs, three machines.
One more thing worth saying. The compressed air instinct is the one that feels the most professional and is the most wrong. People think of it as the proper way to clean electronics. It's not. It's the fastest way to generate static and redistribute dust at the same time. If you take nothing else from this, stop blowing on boards.
If you're setting up a bench from scratch, the order of purchases matters. If you're milling, the particulate capture comes first. That's the one with the most immediate health impact. Then the fume extractor when you start soldering regularly. Then the ESD-safe vacuum only when you're cleaning assembled boards.
The fifty dollar mini-vac is fine as a day-one purchase for general cleanup. But it's not the solution to any of the three real problems. It's a dustpan.
A dustpan with a brush attachment.
Hilbert: It's not three machines. It's four.
Four?
Hilbert: You're forgetting the board holder. Flex PCBs. You lay a flex board flat and it curls. So you put a vacuum nozzle under it, hold it down with suction, and solder it. Works fine. Terrible for ESD, obviously. But it works.
The vacuum is a fixturing tool.
Hilbert: It was in the shop where I worked. Small prototyping outfit. We did boards for GPS modules. This was the late two thousands. And we had one vacuum for everything. A modified domestic Hoover. The owner's wife's old one. We taped a fine nozzle to the hose and used it for milling dust, for general cleanup, and for holding flex boards. One machine, four jobs.
How did that work out?
Hilbert: The Hoover caught fire. Fibreglass dust got through the bag and into the motor. Built up on the windings. One day it just seized and then the insulation went. Small fire, mostly smoke. The owner's wife was not pleased.
The distribution system became a combustion system.
Hilbert: We also used an air compressor to blow off boards. The boss swore by it. We had a batch of prototype GPS modules, maybe forty units. After the air gun, half of them wouldn't lock on to satellites. The boss blamed the supplier. I knew it was the air gun. Static through the RF front end. But you don't tell the boss his favourite method is killing his boards.
Did you say anything?
Hilbert: I said the air gun might be a factor. He said the air gun was how he'd always done it. So we kept doing it. And the next batch lost a third.
The flex board holder trick.
Hilbert: That actually worked. For a while. Then the suction would slip and the board would move mid-joint. So you'd get a crooked solder joint and a static-charged flex board at the same time. Two problems, one nozzle.
Do you still use compressed air on electronics?
Hilbert: I use it on my keyboard at home. I know better. But it's a keyboard. If it dies, I'll buy another one.
The man who watched half a batch of GPS modules die from static still blows on his keyboard.
Hilbert: The keyboard doesn't have a GPS module in it. If it did, I'd use a brush.
Fair enough.
Hilbert: The point is, the fourth job is real. If you're doing flex boards, you'll discover it on your own. And then you'll have four problems and still only one vacuum.
The next time someone tells you to get a small vacuum, ask them which of the three jobs they mean. Or four. And if they say all of them, that's the person whose Hoover caught fire.
The one thing I'd leave people with is this. The word vacuum is doing too much work. It's one word for three separate engineering problems, and the machine that solves one of them is actively dangerous for the other two. The shop vac that recirculates fibreglass dust. The fume extractor that doesn't catch static. The ESD-safe vacuum that can't filter fumes. Separate the problems first, then buy the machine for the problem you actually have.
The problem you actually have is usually the one you can't see. The dust is visible, so people buy for dust. The fume is invisible, so they skip it. The static is invisible, so they don't believe in it. The invisible problems are the ones that cost you.
That's a good place to land. Thanks to our producer Hilbert Flumingtop for keeping us on schedule.
This has been My Weird Prompts, the human-AI collaboration podcast.
If you've got a bench setup question or a weird prompt of your own, email us at show at my weird prompts dot com.
We'll be back soon.
