Daniel wrote in this week with a whole thing about engraving an heirloom menorah, and the question that follows you home from that satisfying scratch — will this mark still be legible in forty years? His actual questions are more specific. Engraving is drilling in miniature, a harder material spinning against a softer one. The art lives in the geometry of the carbide, which he wants to save for a later episode. Today he wants the narrower range of high-end carbides for stainless steel and harder materials, the basic formulae for matching a carbide grade to a given hardness, whether there's any way for a committed engraver to actually measure the hardness of a surface rather than guess at it, and where Dremel fits in a landscape of vendors who publish full specifications.
The menorah detail is doing real work there, because that's exactly the object where you stop caring whether the mark looks good this week and start caring whether it survives a grandson reading it.
And stainless is where the casual tool stops being honest with you. Below it, generic carbide chews through brass and aluminum and mild steel without complaint. At stainless, the grade and the geometry become the difference between a clean mark and a ruined heirloom.
So today we get under the scratch. What hardness actually is, how you measure it without destroying the thing you're measuring, and then how that measurement becomes a tool selection.
And the uncomfortable truth about the forty dollar bit in the wobbly collet.
Let's start with the thing most people think they know and don't. Hardness is not one property. The Mohs scale — talc is one, diamond is ten, your fingernail scratches gypsum — that's a field geologist's trick. It's qualitative. It tells you what scratches what, but it doesn't give you a number you can take to a tool catalog.
And it's the scale every beginner reaches for because it's the one that gets taught.
Right. The engineering standards are indentation-based. You press a specific indenter geometry into the surface under a known load, then you measure what it left behind. The three that matter for engravers are Vickers, Rockwell, and Knoop. Knoop is a long skinny diamond, used for very thin coatings and brittle materials. For our purposes, Vickers and Rockwell are the workhorses.
Walk me through Vickers. I've seen the numbers — HV three hundred, HV eight hundred — but I've never actually known what physically happens.
A diamond pyramid indenter, square base, one hundred thirty-six degrees between opposite faces. You apply a load — could be a few grams for microhardness, could be thirty kilograms for macro. The diamond presses into the surface and leaves a square impression. You measure the diagonal of that square under a microscope, calculate the surface area of the impression, and hardness is load divided by that area. It's a proper physical quantity, force per unit area.
So the number is real, not an index.
It's real in the sense that it has units. The clever part is that the diamond indenter is so much harder than anything you're testing that it doesn't deform, so all the deformation is in the sample. And because the geometry is self-similar — the impression scales with depth — Vickers gives you a number that's roughly constant across different loads for a uniform material. That's why it's the standard for thin materials and small parts.
Which sounds like exactly what an engraver would want, except for the part where you've now put a square hole in the thing you wanted to keep.
That's the catch, and it's not a small one. Vickers and Rockwell both leave a permanent indentation. On a finished workpiece, that's destruction. You cannot Vickers-test the underside of a menorah and then hand it to your son.
So what's Rockwell doing differently?
Rockwell measures depth of penetration rather than impression area. You apply a minor load to seat the indenter, zero the gauge, then apply the major load, release back to the minor load, and read the depth difference. The scale letter tells you the indenter and load combination. HRC — the C scale — uses a diamond cone indenter at one hundred fifty kilograms major load. That's the scale you'll see for tool steels and carbides. HRB uses a steel ball, for softer materials like brass and aluminum.
And the number goes up as the material gets harder?
Yes, but it's a depth measurement, so it's an inverse scale — the harder the material, the shallower the penetration, the higher the number. A typical annealed stainless might be in the high seventies on HRB, which is roughly twenty HRC. A hardened tool steel might be sixty HRC. A carbide tool can be ninety HRC or above, which is where Rockwell C starts to lose resolution and you switch to Rockwell A with a lighter load.
So the committed engraver has a problem. The real hardness tests are destructive, and the thing you want to test is the thing you want to keep.
This is where the indirect methods come in, and they're more practical than most people realize. First, calibrated file sets. These are files made from known hardness references, typically in five HRC increments. You try to scratch the workpiece with each file in sequence. The file that just barely bites tells you the workpiece is softer than that file. It's simple, it's cheap, and it leaves a tiny scratch that you'd probably be engraving over anyway.
And it's the same principle as Mohs, just with calibrated references.
But with actual engineering numbers attached. Second, portable Leeb rebound testers. These fire an impact body with a carbide tip at the surface and measure the rebound velocity. Harder surface, higher rebound. The device converts that to a hardness reading. The catch is that it's designed for large, massive parts — the rebound depends on the workpiece not moving, so a thin plate or a small object gives nonsense readings. You'd use it on a machine base, not a menorah.
And the third option?
Ultrasonic contact impedance, UCI. A rod with a Vickers diamond on the end is vibrated at its resonant frequency, then pressed against the surface. The resonant frequency shifts depending on how much the diamond penetrates. That shift gets converted to a Vickers-equivalent number. It works on thin parts, small parts, finished surfaces with minimal damage. The devices are pricey — a few thousand dollars for a good one — but they're what a serious shop would use for spot-checking.
So the committed engraver can measure hardness non-destructively, if they're willing to spend more on the tester than on the engraver.
Or they can do what most of us actually do, which is read the alloy designation and stop guessing. Three oh four versus three sixteen stainless. Both are austenitic, both work-harden, but three sixteen has molybdenum added for corrosion resistance, and the mechanical properties shift. In the annealed state, three oh four runs about seventy to eighty HRB, three sixteen about seventy-nine to ninety-five HRB depending on the specific heat. Neither is hard in the tool steel sense — we're talking maybe twenty HRC equivalent — but the work-hardening behavior is what bites you during engraving.
That's the part people don't expect. The surface feels soft, you start cutting, and the material gets harder as you go because you're deforming it faster than the heat can anneal it.
Stainless work-hardens at the cut zone. The first pass cuts like butter, the second pass is cutting something measurably harder, and by the time you're doing fine detail you're fighting a surface that's locally hardened well above the bulk material. That's why heat management matters, and that's where the coating conversation comes in later.
So the hardness is in the metallurgy, not the surface feel. The alloy designation plus heat treatment state tells you more than any field test.
And for most engravers, that's the honest answer. You know what you bought. If it's a mystery metal, you reach for the files or the UCI device. If it's a known alloy, you look up the spec and move on to tool selection.
Which is the part Daniel actually asked about. Once you know what you're cutting, how do you pick the carbide that cuts it?
Carbide is not a single material. It's a composite. Tungsten carbide particles — the hard, wear-resistant phase — held together by a cobalt binder. The cobalt is the glue. It's softer and tougher than the tungsten carbide grains. The ratio and the grain size determine everything about how the tool behaves.
So more cobalt means tougher but softer.
Right. More cobalt means the tool can absorb impact without shattering, but the overall hardness drops because cobalt is the soft phase. Less cobalt means harder, better wear resistance, but the tool gets brittle. And grain size matters independently. Finer tungsten carbide grains mean more grain boundary area, which makes the material harder for the same cobalt content, but also more brittle.
So the formula isn't just hardness versus workpiece hardness. It's a three-way trade between hardness, toughness, and grain size.
And for engraving hardened steels — say above forty-five HRC — you need a micrograin or sub-micrograin grade. Sub-micrograin means the tungsten carbide particles are below one micron. That gives you the edge sharpness and hardness to cut a hard workpiece without the edge crumbling. The cobalt content gets tuned to balance edge retention against chipping resistance. A typical micrograin grade for steel engraving might be six to ten percent cobalt, hardness in the range of ninety-one to ninety-three Rockwell A.
And the practical formula Daniel asked for? Workpiece hardness maps to a required tool hardness with a margin?
The rule of thumb is you want the tool hardness comfortably above the workpiece hardness — not by a little, by a lot. Carbide at ninety HRA is up around seventy HRC equivalent, which is harder than almost any steel you'd engrave. So for stainless, even work-hardened stainless, the hardness gap isn't the problem. The problem is toughness and heat.
The formula is more about cobalt content and geometry than about raw hardness.
That's the misconception worth killing. A harder carbide bit is not automatically better. If you grab a grade with three percent cobalt and run it into a work-hardening stainless, the edge chips. You've got a bit that's technically harder than the workpiece and it's failing anyway. The bit that survives is the one with enough cobalt to absorb the impact and enough grain refinement to hold an edge.
Then geometry. You said the art lives in the geometry.
The flute count, helix angle, and point geometry matter as much as the grade. A two-flute bit clears chips differently than a four-flute. Two flutes give more chip clearance, which matters in a shallow engraving cut where you're not moving much material but you're moving it constantly. Four flutes give a finer finish but can pack chips in a shallow cut. The rake angle determines whether the tool cuts or pushes. A positive rake slices; a negative rake scrapes. For stainless, you want a sharp positive rake that shears the material rather than rubbing it.
If you rub it, you work-harden it faster and generate heat.
Which brings us to coatings. Titanium aluminum nitride, TiAlN, applied by physical vapor deposition. It's a ceramic layer a few microns thick. Its job is to act as a thermal barrier. The tool can run hotter without the cobalt binder softening and the tungsten carbide grains pulling out. For stainless, that's critical, because the cut zone retains heat and the material work-hardens, so you're generating more heat with every pass.
The coating isn't making the tool sharper. It's making it survive its own heat.
TiAlN specifically forms an aluminum oxide layer at high temperature, which is a self-reinforcing thermal barrier. The hotter it gets, the better the oxide layer protects. That's why it's the standard coating for stainless and high-temp alloys.
Now the vendor question. Daniel asked where Dremel fits.
Dremel dominates the consumer rotary space. It's the Kleenex of rotary tools. But their carbide bits — the nine ninety series and similar — are underspecified for hardened steel. They don't publish the full grade specifications. They don't tell you the cobalt content, the grain size, the hardness. They sell a bit that says "carbide" and it works fine on soft metals and wood and plastic. For stainless, especially work-hardened stainless, you're gambling.
The industrial suppliers do publish.
Harvey Tool publishes full hardness specs, grain size, cobalt content, coating options. Kyocera and General Carbide do the same. General Carbide's designer's guide has the complete hardness numbers for their grades — you can look up the exact grade, see the Rockwell A hardness, the transverse rupture strength, the grain size classification. That's the difference. A hobby vendor sells you a bit. An industrial supplier sells you a specification.
The unspoken truth you mentioned earlier.
For most engraving tasks on stainless, the limiting factor is not the carbide grade. It's the rigidity of the tool holder and the spindle speed. A fifty dollar industrial carbide bit in a wobbly Dremel collet will fail where a ten dollar bit in a rigid spindle succeeds. The runout — the wobble — means the cutting edge is hammering the workpiece instead of shearing it. That impact load is what chips the edge. And the spindle speed matters because carbide wants to cut fast. Dremel tops out around thirty-five thousand RPM. A proper engraving spindle runs fifty thousand or more.
The committed engraver's formula is: know the workpiece hardness, pick a micrograin carbide with enough cobalt to survive impact, match the geometry to the cut, add a TiAlN coating for heat, and then mount it in something that doesn't wobble.
The order matters. Most people buy the expensive bit first and the rigid spindle last. It should be the other way around.
The bit is the last thing to upgrade, not the first.
Because the bit is the thing everyone can see and hold. The spindle is boring. The collet is boring. But that's where the failures actually happen.
I keep thinking about the menorah. Daniel said it felt satisfying to engrave the underside. And then he thought about durability, about handing it down. And the durability question isn't really about the carbide grade, is it? It's about whether the mark is deep enough and clean enough that it survives decades of handling, cleaning, polishing.
The carbide grade determines whether you can cut the mark at all without destroying the tool. The geometry determines whether the mark is clean. But the depth and the profile of the cut determine whether it survives. A shallow scratch in stainless will last a long time, but a deep, clean V-groove will last centuries. That's the difference between marking and engraving.
Engraving is a physical protest against impermanence. You said that once.
I said something like it. The point stands. You're not just identifying the object. You're asserting that it matters, that it should outlast you.
The material science is how you make that assertion stick.
Hilbert's been sitting at the desk this whole time and he's got that look like he wants to say something about a file set.
Hilbert: The pantograph shop I worked in the summer of eighty-nine had a coffee can full of broken carbide bits and a rule. If it breaks, you were pushing too hard. Not that the bit was bad.
How many bits did you personally break?
Hilbert: I ran a pantograph engraver for a trophy shop. Corporate awards, brass plates, the occasional silver cup. You trace the master template with the stylus and the cutter follows at a reduction ratio. The machine did the geometry. I did the clamping and the swearing.
A pantograph is a beautiful machine. All the skill is in the setup, then it's just tracing.
Hilbert: The skill was in knowing what you were cutting. We had a job once, a batch of plates for some insurance company's annual dinner. Brass, supposedly. The box said brass. The first plate ate a brand-new bit in about four seconds. Turned out someone had mislabeled a batch of steel plates. The hardness was in the label, not the surface.
That lands the whole episode in one sentence.
Hilbert: The shop owner, a man named Rudy, had a set of calibrated files in a leather roll. He taught me to scratch the back of a mystery plate with the files in order until one bit. Then you knew what you were dealing with before you put a bit to it. That file set predates every electronic tester in this conversation and it still works.
The calibrated file set is the oldest indirect method and it's still completely valid. Five HRC increments, you scratch, you feel, you know.
Hilbert: I kept that file set when Rudy retired. It's in a drawer in my flat. If anyone brings an unmarked piece of metal to the studio, I can demonstrate. Takes about thirty seconds.
The files themselves are hardness references. They're the standard you're comparing against.
Hilbert: Rudy's rule was trust the files, not the box. The box lies. The files don't.
That's the practical difficulty in a nutshell. You can't know what you're cutting from the label. You either know the alloy because you bought it, or you test it. And if you test it wrong, you break bits.
Hilbert: I broke a lot of bits before I learned the files. Rudy used to say every broken bit is tuition. I paid a lot of tuition that summer.
The formula Daniel asked for ends up being less about the carbide and more about knowing what's in front of you.
That's the honest answer. The grade selection is the easy part once you know the workpiece. The hard part is the measurement, the identification, the not-guessing.
Which makes me wonder where the craft knowledge is going. As the carbide grades get more specialized and the coatings more sophisticated, does the skill move from the artisan's hands to the material scientist's spec sheet?
I don't think it moves. I think it splits. The artisan still needs to know what a clean cut feels like, what chatter sounds like, when the tool is rubbing instead of cutting. The material scientist gives you the right tool for the job. The artisan still has to run it.
The spec sheet doesn't tell you when to back off the pressure.
It doesn't tell you that the plate you're about to engrave is steel, not brass. That's what the files are for.
Daniel's promised us a geometry episode next — the flute angles, tip shapes, and edge preparations that turn a good grade into a beautiful mark. That's where the art actually lives.
Today gave us the foundation. Hardness is measurable, the measurement is the hard part, and the carbide grade is only as good as the spindle it's spinning in.
The mark on that menorah will outlast all of us, if the depth is right and the cut is clean. That's the whole point.
Thanks to Hilbert Flumingtop for producing, and for the file set that's apparently still in a drawer waiting for a mystery metal.
This has been My Weird Prompts. Email us at show at my weird prompts dot com.
We'll be back soon.
