Daniel's been engraving. He bought a Dremel 7760 a few months ago, expecting the main obstacle to be his asthma — respirator, vacuuming, the whole dust-management routine. Turns out that part's manageable. The real puzzle, he says, is the bits. He's got a 0.7 millimeter carbide for fine work, but for PCB part numbers even that feels like writing with a crayon. And when he switches to fine carbides for 1 or 2 millimeter numbers on HDPE, the result is chicken-scrawl — multiple thin lines that look hand-drawn and messy. So he's asking four things. One: what are the actual limits of human precision for readable micro-engraving on PCBs? Two: how do you select the right mixture of carbides for very fine work versus larger markings, and why do the fine ones fail so badly at 1 to 2 millimeters? Three: what tools exist for cleaning the fur out of engraved channels in HDPE and ABS, beyond a spudger and coarse sandpaper? And four: for high-volume work, is there a quick-change system for Dremel engraving bits — something like the interchangeable heads or magnetic chucks you get with drills — or do people just get fast at swapping collets?
The answer to most of those questions sits inside a single counterintuitive fact about how plastic wants to be cut. And it's not the fact most people reach for. So let's start by defining what we're actually talking about when we say micro-engraving here — because the term covers a range that spans about four orders of magnitude.
From "I can read it" to "nobody can read it without a microscope and a prescription."
Right. At one end you've got Graham Short — we'll get to him — engraving letters the size of a red blood cell. That's not what Daniel's doing. For this episode, micro-engraving means the practical hobbyist zone: 0.5 to 2 millimeter text on PCBs and plastics, with a handheld rotary tool. The Dremel 7760 he bought is a capable machine — 8,000 to 25,000 RPM, 4-volt lithium-ion, about fifty bucks. But its finest standard engraving bit, the 9924, has a 0.8 millimeter tip. That's already too coarse for PCB part numbers. So the moment you want to go finer, you're out of Dremel's catalog and into third-party suppliers — PreciseBits, Vortex Carbide — and suddenly you're making geometry choices with no guidance.
And the geometry is the whole game. That's where we're going. First, the limits of what a hand can do. Then the bit paradox. Then the fur problem and its surprising fix. Then the workflow tools, and finally the state of quick-change systems — which, spoiler, is not what anyone hopes it is.
The first question Daniel raised is the most fundamental: what can a human hand actually do at this scale? And there are two answers — the extreme outer limit, and the practical hobbyist ceiling. The extreme limit is a man named Graham Short. In 2011 he completed a piece called "Nothing is Impossible" on the edge of a razor blade. The letters were 0.1 millimeters. Invisible to the naked eye — you need 400x magnification to read them. And that's not even his smallest work.
His smallest is 6 microns. 0.006 millimeters. That's the size of a red blood cell. A human hair is about a hundred microns, so we're talking about lettering sixteen times thinner than a hair.
The process to get there is, honestly, closer to a medical procedure than a craft. He works only at night — midnight to 5 AM, three nights a week — to avoid traffic vibrations. He takes beta blockers to drop his heart rate to around 20 beats per minute. He uses Botox to suppress blinking and facial twitches. He loops a leather strap around his elbow and anchors it to a heavy press to lock his arm rigid. In a five-hour session he completes 7 or 8 cuts. Seven or eight.
So the answer to "what can a human do" is: 6 microns, if you're willing to chemically suppress your own pulse and work between heartbeats. That's not a workflow. That's performance art with a carbide tip.
And Graham Short himself says he's at the limit. He told Sky News, "I can't do any more now, I'm at the limit." He also said something revealing about mistakes — he expects them. "It's part of the work." At that scale, the slip is built into the process. He just factors it in.
Which brings us to the practical ceiling. James Stanley — he's the one who built the drag engraver we'll talk about later — did a series of tests engraving text on aluminum at different sizes. His results are the best benchmark I've found for what a skilled hobbyist can expect. At 1 millimeter, the text is legible. At 0.5, you're squinting. At 0.4 to 0.3, it's legible only if you already know what it says — your brain fills in the gaps. At 0.2 millimeters, it's illegible even if you know the text.
And that's on aluminum, which cuts cleaner than plastic. For PCBs, it's even harder. The surface is never truly flat, so cut depth varies across the board. And here's the thing most people don't realize about tiny V-bits: they're rarely concentric. The tip spins in a small circle — runout — so the slot you get is much wider than the nominal tip size. You think you're cutting a 0.1 millimeter line, but the tool is actually routing something closer to 0.3 or 0.4.
So the tool lies to you. The number on the package is the tip diameter under ideal conditions, not the width of the cut it actually makes when it's spinning at 20,000 RPM in a handheld tool.
Stanley's proposed solution for PCB work is interesting — he suggests using a drag engraver. Spindle off. Spring-loaded diamond tip that just... scratches. No rotation means no runout, and the spring means it self-levels on uneven surfaces. It's elegant. But it's also an open experiment — nobody has published results showing it works at production scale. He proposed it, but as far as I can tell, he hasn't tested it systematically.
So the practical limit for a hobbyist with a Dremel and steady hands is somewhere around 0.5 millimeters for "I can read this" and 1 millimeter for "this looks clean." Below that, you're fighting physics — runout, surface flatness, the tremor in your own fingers. And no amount of beta blockers is going to turn a Dremel into a medical microscope.
Now, the bit geometry paradox. This is the core technical insight, and it's what Daniel's running into with his 1 to 2 millimeter numbers. He's got a fine carbide — say a 0.1 millimeter tip V-bit from Vortex Carbide, 20-degree angle, 3.175 millimeter shank, fits the Dremel Lite perfectly. Beautiful tool. But when he tries to write a 2 millimeter number with it, the result looks terrible.
Chicken-scrawl.
And the reason is straightforward once you think about it. A very fine V-bit cuts a single thin line. To make a visible number at 2 millimeters, the stroke width of each digit needs to be maybe 0.3 or 0.4 millimeters wide. The fine bit's line is much narrower than that. So you have to trace the outline of each digit, or make multiple passes side by side. And because you're doing this by hand, those passes don't line up perfectly. You get an irregular, hand-drawn look — exactly the chicken-scrawl Daniel's describing.
So the bit that's perfect for 0.5 millimeter text is exactly wrong for 2 millimeter text. And the fix isn't "get better at tracing" — it's "use a different bit."
A wider bit — a ball-nose or a larger V-bit — cuts the full stroke width in a single pass. One pass, one clean line. The number looks machine-made. This means the ideal engraving kit needs bits at both extremes. Sub-0.2 millimeter for micro work — think PreciseBits, they do a 0.005 inch tip, that's 0.13 millimeters, and they go down to 0.003 inches, which is 0.076 millimeters, both with 1/8 inch shanks, fully compatible with the 7760. And then at the other end, 1 to 2 millimeter ball-nose or V-bits for readable numbers. Nothing in the middle does both jobs well.
The bit that tries to split the difference gives you mediocre micro-work and mediocre large work. It's the engraving equivalent of an all-season tire — fine in no conditions.
And these third-party bits are serious tools. PreciseBits are 100 percent optically inspected, submicrograin tungsten carbide, hardness HRa 92. They're not the little diamond-coated grinding bits that come in a twenty-pack for eight dollars. Those have their place, but not for this.
So if the bit geometry paradox means you need bits at both extremes, the next question is how to actually use them on plastic without making a mess. Daniel mentioned HDPE and ABS, and the fur that comes up unpredictably from the surface. He's discovered the spudger as a cleaning tool, which is creative — but the fur problem starts before you ever pick up a cleaning tool.
The fur is caused by heat. Specifically, heat from multi-flute bits. Most beginners assume more flutes means a smoother cut — more cutting edges, cleaner finish. For HDPE, the opposite is true. There's a post on the Easel forum from 2015 that nails this. A user named BillArnold was cutting HDPE with a 2-flute spiral bit, 1/32 inch, and getting — his words — "tons of fuzzies" that a brush couldn't remove. He tried different feed rates, different spindle speeds. Nothing worked. The fix was a single-flute bit.
Why does that work?
A multi-flute bit on plastic generates friction. The extra cutting edges don't clear chips fast enough, so the tool rubs instead of shearing. The plastic partially melts, and that melted plastic re-solidifies as fur along the edges of the cut. A single-flute bit — often called an O-flute or a downcut bit — has one big chip channel. It shears the plastic cleanly, the chip evacuates immediately, and the heat leaves with the chip.
So the rule is: make chips, don't rub. If you see dust or fuzz or smears, you're generating friction, not cutting.
That's rule number one from the Interstate Plastics HDPE fabrication guide, almost verbatim. Rule two: get the chip out of the cut immediately — air blasts, vacuums, whatever you've got. Rule three: use sharp carbide bits with proper chip load. And this is why the Dremel's low-speed setting — 8,000 RPM — and shallow passes help. High RPM with a fine bit on plastic is a recipe for melting.
Jim Boslice put out a guide in June of this year — 2026 — that adds a few practical tricks. Painter's tape on the surface before engraving reduces burrs. Lowest speed setting. Shallow passes. Clear debris frequently with a soft brush or compressed air. For ABS specifically, the insight is that melting isn't caused by high speeds per se — it's caused by rubbing. If your feed rate is too slow or the tool is dull, the cutter rubs. The fix is higher feed rates so heat gets carried away in the chip.
Which feels wrong when you're doing delicate work — your instinct is to go slow. But going too slow is exactly what melts the plastic.
So you've got your single-flute bit, you've got your low RPM, you've got your tape, and you still get some fur. Now what? Daniel mentioned the spudger, and he's right that it works. But there are tools actually designed for this.
Hand deburring tools are the professional answer. The KLEMA NB1100 is a good example — ergonomic handle, 360-degree rotating high-speed steel blade, designed specifically for plastics, copper, aluminum. It gets into engraved channels and slices off burrs cleanly. STARVAST makes a similar tool with 15 interchangeable blades. Slice UK has deburring tools built for the finishing stages of plastic manufacturing — they're designed to remove roll-over burrs, tear burrs, and cut-off burrs from injection-molded or machined parts.
For rotary work, XEBEC makes ceramic fiber deburring stones that go in a micro-motor system. They get into hard-to-reach edges and internal features. But for the hobbyist with a Dremel, the community-tested approach for HDPE fuzz includes a few things Daniel might already have: a stiff nylon brush, a razor blade for manual cleanup, and — this one surprised me — a light pass with a torch.
That comes from a user named MichaelHursh on the Easel forum, 2017. A quick pass with a flame burns off the fuzzies. It's essentially flame polishing — you're applying an oxygen-rich flame briefly to melt the outer layer, which smooths micro-ridges and removes frosted edges. You have to work fast to avoid oxidation and discoloration, but it works.
A torch. For cleaning up plastic. The same tool you'd use for creme brulée.
It connects to something important: flame polishing is a known technique in HDPE fabrication. It's not a hack — it's a standard process, just usually done at larger scale. The hobbyist version is a small butane torch and quick hands.
Daniel also asked about workflow — specifically, whether there's a quick-change system for Dremel engraving bits. The kind of thing where you pop one head off and snap another on, like the magnetic chucks on drills. And the answer is... no. Flat no. There is no magnetic chuck, no interchangeable-head system, no keyless chuck equivalent for Dremel engraving accessories.
The Dremel 7760 has the EZ Twist nose cap. You press the shaft lock button, twist the nose cap to loosen the collet, swap the bit, tighten. That's it. It's tool-free, which is better than the old wrench system, but it's still a collet change. The EZ SpeedClic system — the one that's "six times faster" — is for cutting wheels and sanding discs only. Not engraving bits. The shanks don't fit.
The answer to "do people just get fast at changing bits" is yes. That's exactly what they do. The EZ Twist is a single motion once you've got the muscle memory. Press, twist, swap, twist. Experienced users do it in seconds. It's not elegant, but it's the system we've got.
The one innovation worth mentioning is the PreciseBits depth ring system for CNC users. You can order their micro-engraving bits with depth rings positioned to plus or minus 0.004 inches, so you don't have to reset Z-axis zero between tool changes. That's clever for CNC work, but it doesn't help with a hand-held Dremel. For Daniel's use case, the workflow is: get fast at the EZ Twist, and design your engravings to minimize bit changes.
Which is actually a design problem, not a tool problem. If you plan your engravings so all the fine work happens with one bit and all the large work happens with another, you're changing bits once per piece instead of three times.
That's where the single-bit philosophy comes in.
Hilbert: You're all talking about bits and collets and deburring tools, but the real secret is the old guy never changed his bit.
Wait —
Hilbert: I worked a summer at a shop that made custom nameplates for industrial equipment. "DO NOT OPERATE WITHOUT GUARDS" in 2 millimeter letters on HDPE panels. Hundreds of them a week. The owner was this old German machinist, and he had one bit. A 1.2 millimeter single-flute carbide. He did everything with it.
Everything?
Hilbert: Micro-engraving, he'd tell you to buy a laser. He said the Dremel was for marking, not art. But for nameplates? One bit. The small text got one pass, the large text got two passes side by side, and because he'd been doing it for twenty years the two passes looked like one stroke. Machine-made. No chicken-scratch, because he wasn't tracing — he was filling. Two parallel lines from a single-flute bit, and they met perfectly.
The chicken-scrawl problem might be a workflow problem, not a bit selection problem. If you commit to one bit size and design your markings around it, you eliminate the quick-change bottleneck entirely.
Hilbert: He also used a torch for cleanup. I thought he was insane the first time I saw him do it — one quick pass over a fresh nameplate, all the fuzzies gone, surface polished clear, no ink needed because the flame left it glossy. But he had a respirator that looked like it belonged in a cleanroom, and he'd vacuum the whole bench before starting. Every time. So the asthma thing? He'd have told your friend to get a better vacuum. And a downdraft table if he could afford one.
A downdraft table for a Dremel engraving station. That's... actually not a bad idea.
Hilbert: The shop had one he built himself. Plywood box, furnace filter, old fan from a scrapped air handler. Pulled the dust straight down through the work surface. Cost him maybe forty dollars in 1979.
1979?
Hilbert: I was seventeen. The bit he used cost twelve dollars and he sharpened it himself on a diamond stone. He said the secret to HDPE wasn't the tool, it was never letting the plastic get warm. If the panel felt warm to the touch, you were already doing it wrong.
The single-bit philosophy — one 1.2 millimeter single-flute carbide, everything designed around that dimension, torch for finish, and obsessive dust control. That's a complete system.
Hilbert: He retired in '03 and sold the business to a guy who bought a laser engraver and went bankrupt in eighteen months. The laser made pretty pictures but the nameplates looked cheap. No depth. You could feel the difference with your fingers.
That tactile quality matters for industrial nameplates. A laser etch sits on the surface — it's a color change, essentially. An engraved mark has topography. It survives abrasion, solvents, decades of people touching it.
Hilbert: The old guy's nameplates are still on machines all over the Midwest. I saw one two years ago in a factory in Ohio. "EMERGENCY STOP — DO NOT BYPASS." Still legible. Forty-some years.
Forty years on a single-flute cut.
Hilbert: The plastic yellows but the letters don't go anywhere.
Where does that leave us? The drag engraver for PCB isolation routing is still an open question — nobody's tested it at production scale, and it might be the missing tool for micro-engraving on uneven surfaces. And the single-bit philosophy raises its own question: does it scale to high-volume work, or does it just shift the bottleneck from bit changes to pass planning? If you're doing two passes for every large character, that's double the machine time.
But zero tool-change time. The math depends on the job. For a hobbyist doing small batches, the single-bit approach probably wins. For production, you'd want to run the numbers.
The bigger point is that the gap between "possible" and "practical" is enormous — Graham Short at 6 microns on one end, a clean 2 millimeter number on HDPE on the other — but the hobbyist with a Dremel and a respirator is closer to the practical limit than they think. The answers are mostly about geometry and heat management, not steady hands. Single-flute bits for plastic. Bits at both extremes of the size range. A torch and a vacuum. And the willingness to treat bit selection as a design decision, not a shopping problem.
If you've got a micro-engraving tip you swear by, or a technique for taming HDPE fur that we missed, send it in. We might do a follow-up.
Thanks to our producer Hilbert Flumingtop for keeping this show running — and apparently for knowing more about single-flute carbides than I do.
This has been My Weird Prompts. Email us at show at my weird prompts dot com.
We'll be back soon.
