You know the feeling. Soldering iron's hot, multimeter's beeping continuity, the three-D-printed bracket you spent two hours dialing in is finally in your hand — and you're staring into a drawer of forty-seven nearly identical little silver screws, and not one of them is the M-two-point-five by six millimeter pan-head you need.
The worst part is you can see it in your mind. You know exactly which screw it is. You probably held it three disassemblies ago and thought, I should order more of these.
You didn't. Because you were on a roll. And now the whole project is dead on the bench over something that costs eight cents.
Daniel sent us this one, and I felt it in my bones. He's been doing a lot of electronics repair over the last year, built up an inventory system, put real effort into it — and still hit that wall where the one screw he needs simply isn't there. His question is basically: can we build a field guide? Map the common M-sizes to actual components, decode what UNC six-thirty-two actually is, machine screws versus pointed screws, and give someone a mental map they can use to build a logical storage system that doesn't let them down.
The thing is, he's right that this is one of those domains where if you can learn the system conceptually, everything clicks. Most people never do. They learn by osmosis — this screw came out of a laptop, so it goes back into a laptop. Which works fine until you lose it, or strip it, or need to source a replacement from a supplier who wants an actual spec.
The paradox here is genuinely interesting. Electronics repair demands absurd precision in fastening — torque specs, thread engagement, the right drive type so you don't cam out and destroy a board — but the fastener ecosystem itself is a chaotic mess of overlapping standards. Metric and imperial. Coarse thread and fine thread. Machine screws and self-tapping screws. Phillips, Pozidriv, Torx, hex, and about six others. Nobody teaches this. You just accumulate a jar of mystery screws and hope.
The jar of shame. Every workbench has one.
Every single one. And here's the thing — Daniel's actually ahead of most people because he tried to build a system. Most repairers never get that far. They just raid the jar and pray.
Our job here is to give him, and anyone else staring into that drawer, a taxonomy that actually works. What the numbers mean, which sizes map to which components, how to tell Phillips from Pozidriv at a glance, and what a rational starter inventory looks like.
Thread form — that's your M-two, M-two-point-five, M-three, and the imperial equivalents. Head style — pan, flat, button, socket cap. And drive type — what you stick the tool into. If you can place a screw on those three axes, you can identify it, order it, and know where it lives in your organizer.
The beautiful thing is, once you understand the naming convention, you can look at any small screw and deduce its specs in about thirty seconds. It stops being tribal lore and becomes a system.
Let's build that mental map. We'll start with the M-size hierarchy and what actually uses each one, then get into thread pitch, head styles, drive types, and the inventory system that actually covers ninety percent of what you'll ever touch.
We'll talk about the imperial elephant in the room — UNC six-thirty-two — because that one screw has probably stripped more threads than all the others combined.
The reason this stays fragmented even among people who've been repairing things for years is that nobody sits you down and says "here's how fasteners work." You learn in context — this screw came out of a ThinkPad hinge, so it goes back into a ThinkPad hinge. The knowledge is tied to the object, not to the screw itself.
Which is fine until you're staring at a bag of fifty M-two-point-five screws from AliExpress and trying to remember whether the length you need is measured from under the head or includes the head. And whether M-two-point-five even is the one that mounts an M-dot-two drive.
And that's the core of what Daniel's asking for — a mental model that's portable. Not "the screw from that Dell laptop" but "an M-two by four millimeter flat-head." Once you can name it, you can order it, stock it, and never be dead in the water again.
Let's name the axes. You mentioned three — thread form, head style, drive type. Walk me through what someone actually needs to hold in their head.
Thread form first, because it's the one people mess up most. In electronics, you're almost always dealing with metric coarse thread — that's your M-one-point-six, M-two, M-two-point-five, M-three, and so on. The M stands for metric, the number is the outer diameter in millimeters. Standard pitch is assumed unless specified otherwise — M-three is M-three-by-zero-point-five, M-four is M-four-by-zero-point-seven. Fine pitch exists but it's rare in consumer electronics. You'll see it in camera tripod mounts and some vibration-prone industrial gear, but for a repair bench, coarse pitch is your default.
Then there's the imperial stuff lurking in the drawer.
The imperial stuff is mostly UNC — Unified Coarse. The one that haunts everyone is number-six-thirty-two, which is used in PC power supplies, older drive cages, and a lot of US-standard electrical enclosures. It has a nominal diameter of three-point-five-oh-five millimeters, which is infuriatingly close to M-three-point-five — a metric size that barely exists in the wild. People grab what looks right, force it, and strip the threads.
The rule of thumb is: if it came out of a PC power supply and it's not quite M-three and not quite M-four, it's probably six-thirty-two.
And the other imperial size you'll bump into is number-four-forty, which shows up in some older standoffs. But six-thirty-two is the one that's burned into my memory.
That's thread form. What about head style?
Four types cover almost everything in electronics. Pan head — slightly domed, the default, the workhorse. Flat head — tapered underside, meant to sit flush in a countersunk hole. That's your laptop bottom cover screws, anything where you need a smooth surface. Button head — lower profile than pan, more rounded, decorative. Very common in three-D printer builds where you want things to look clean. And socket head cap screw — cylindrical head with a hex socket, used where you need to apply real torque. That's your structural connections, motor mounts, linear rail brackets.
Then the third axis — what you stick into the screw.
And this is where the Pozidriv trap lives. Phillips is the cross-shaped one everyone knows. It was actually designed to cam out — the driver is supposed to slip out of the screw head when you hit a certain torque, to prevent over-tightening on assembly lines. That's a feature, not a bug. Pozidriv looks almost identical but has those little radial tick marks between the cross slots, and it's designed specifically not to cam out. If you use a Phillips driver on a Pozidriv screw, it'll slip, chew up the head, and leave you with a stripped mess.
Pozidriv shows up where?
IKEA furniture, a lot of European electronics, some laptop internals. The giveaway is those extra lines on the screw head — if you see them, reach for a Pozidriv driver. Then you've got Torx, the six-pointed star, which is increasingly common in three-D printers and RC gear because it handles high torque without stripping. And hex, which is your standard Allen-key socket — that's what socket head cap screws use.
The mental map Daniel needs is: look at a screw, measure the diameter to get your M-size, check the head shape to know what style it is, identify the drive recess, and you've placed it on all three axes. After that, length is just length.
Once you can do that, you're not guessing anymore. You can walk into a hardware supplier or open a catalog and say "I need M-three by ten millimeter pan head Phillips" and get exactly the right thing. That's the system. The rest is just filling in which sizes map to which components — and that's where the real field guide starts.
Let's actually walk the size ladder, because once you can name a screw, the next question is "what uses this size?" And Daniel specifically asked for that mapping. Start at the bottom — M-one-point-six. These are tiny. You'll find them in Raspberry Pi camera modules, some drone arms, and really compact wearable electronics. They're easy to strip just by looking at them wrong.
The dental-floss of the fastener world.
Then M-two, which is your two-point-five-inch hard drive mounting screw, older laptop hinge assemblies, and — this is one people miss — the M-dot-two specification actually calls for M-two by zero-point-four thread for the mounting standoff and M-two by three or four millimeter for the screw that secures the drive itself. So M-dot-two equals M-two. The name is the hint.
Which is either deeply satisfying or mildly infuriating, depending on how long you spent not knowing that.
I choose satisfying. M-two-point-five is the real workhorse of small electronics. It's the dominant size for M-dot-two SSD mounting on most modern motherboards, it's what Raspberry Pi boards use for board mounting and HAT standoffs, and it's all over three-D printer hotend assemblies. If you're doing Pi projects or SSD swaps, M-two-point-five is your size.
M-three is where everything lives.
M-three is ubiquitous. Three-D printer frame joints, Arduino mounting holes, most PC case standoffs, RC car components, camera cage rigs, drone frame arms — if you grabbed a random screw off my bench, there's maybe a sixty percent chance it's M-three. Standard pitch is zero-point-five millimeters, and common lengths run from six to twenty millimeters. For a starter inventory, you want six, eight, ten, twelve, sixteen, and twenty.
M-three is your staple. What about M-four?
M-four steps up into larger three-D printer parts, CNC spoilboard mounting, some power supply terminals, and heavier brackets. Pitch is zero-point-seven. Lengths typically eight to thirty millimeters. You won't need as many of these, but when you need one, a missing M-four stops the whole build.
Then M-five and M-six are basically structural.
Right — extruder mounts, linear rail brackets, some motor mounts. M-five is zero-point-eight pitch, M-six is one-point-oh. At that point you're not doing electronics repair anymore, you're doing mechanical assembly. But if you have a three-D printer or a CNC machine, you'll touch these.
Let's decode the full notation, because Daniel mentioned wanting a conceptual understanding. What does M-two-point-five by zero-point-four-five by six millimeters actually tell you?
Three pieces of information. M means metric thread profile. Two-point-five is the nominal outer diameter — the major diameter of the threads, measured in millimeters. Zero-point-four-five is the thread pitch — the distance from one thread crest to the next. And six is the length in millimeters, measured from under the head for pan heads and button heads, or the total length including the head for flat heads since they sit flush.
That length measurement difference — under-head versus overall — that's the kind of thing that bites you when you're ordering online and the photo is ambiguous.
It bites everyone at least once. You order what you think is a six-millimeter flat-head, it arrives, and it's actually a six-millimeter screw where two millimeters of that is the countersunk head, so you only get four millimeters of thread engagement. Now your laptop bottom cover won't catch.
The thread pitch part is what I think most people gloss over. They see M-three and assume all M-three screws are interchangeable.
They usually are, because coarse pitch is the default and fine pitch is rare in electronics. But "usually" isn't "always," and that's where the mystery failures come from. You grab an M-three screw that looks right, it starts threading in, then binds up after two turns. Congratulations — you just found a fine-pitch screw in your coarse-pitch hole.
What are the standard pitches people should memorize?
For electronics, four numbers cover almost everything. M-two is zero-point-four. M-two-point-five is zero-point-four-five. M-three is zero-point-five. M-four is zero-point-seven. M-five is zero-point-eight, M-six is one-point-oh. Fine pitch variants exist — M-six by zero-point-seven-five is common in automotive and aerospace — but on a repair bench, if you have coarse pitch in each of those sizes, you're covered.
The imperial elephant. Let's talk about six-thirty-two.
Number-six-thirty-two UNC. Nominal diameter is three-point-five-oh-five millimeters, about a hundred and thirty-eight thousandths of an inch. Thirty-two threads per inch, which works out to roughly zero-point-seven-nine-four millimeters between threads. It's used everywhere in PC power supplies, older drive cages, US-standard electrical boxes, and some rack-mount gear.
It looks almost exactly like an M-three-point-five.
M-three-point-five is a ghost. It technically exists — M-three-point-five by zero-point-six is a real metric spec — but you almost never encounter it in consumer electronics. So someone pulls a six-thirty-two screw out of a power supply, loses it, digs through their metric kit, finds an M-three-point-five that's visually identical, and forces it in. The diameters are only about five-thousandths of an inch apart, so it feels like it should work. It doesn't. The thread pitches are different — zero-point-six millimeters versus zero-point-seven-nine-four — so you're cross-threading immediately.
Five-thousandths of an inch. That's the thickness of a sheet of paper.
It's enough to destroy the threads in whatever you're screwing into. The rule is: if it came out of something that plugs into a wall outlet in North America, check whether it's six-thirty-two before you reach for metric.
Let's walk through the identification process. Daniel's got a mystery screw on the bench. What does he do?
Digital caliper — under twenty dollars, no excuse not to own one. Measure the outer diameter of the threads. If it reads two-point-five millimeters, you're looking at M-two-point-five. If it reads three-point-five, pause — that's either M-three-point-five or six-thirty-two, and you need to check the thread pitch next.
Thread pitch without a pitch gauge?
You can count threads over a known distance. Lay the screw against a ruler, count how many thread crests you see in ten millimeters, divide. Ten crests in ten millimeters means one-millimeter pitch — that's M-six coarse. About twenty-two crests in ten millimeters is roughly zero-point-four-five pitch — that's M-two-point-five. It's not as precise as a pitch gauge, but a pitch gauge costs under ten dollars. Just buy one.
Measure diameter, measure length, check pitch, identify the head and drive type. Thirty seconds, you've got a full spec.
Then you write it down and put it in a labeled bin, and you never have to identify that screw again. That's the system. The first time is work. Every time after that is just reading a label.
Now that we can name a screw, let's talk about the other half of the equation — machine screws versus self-tapping screws. This is the distinction that strips threads when you get it wrong.
I've gotten it wrong. The black screws that come with Arduino cases — those have a different feel when they go in.
Those are self-tapping screws, also called thread-forming screws. They're designed to cut their own threads into plastic or soft metal. The tips are often slightly notched or have a cutting edge, and the threads are hardened. You see them everywhere in plastic enclosures — Arduino project boxes, three-D printer printed parts, consumer electronics shells. The mistake is grabbing one and driving it into a pre-tapped metal hole. The hardened threads chew up the existing threads, and now both the screw and the hole are ruined.
The rule is: if the hole already has threads, use a machine screw. If it's bare plastic, use a self-tapper.
That's the rule. Machine screws need a pre-tapped hole or a nut — they're not cutting anything, they're just following existing threads. For Daniel's inventory, I'd stock machine screws as the primary system and keep a small assortment of self-tappers in M-two, M-two-point-five, and M-three for plastic enclosures. You don't need every length — four, six, and eight millimeters cover most plastic case work.
Now head styles. You touched on the four main ones earlier, but let's talk about when you'd actually choose each.
Pan head is your default. It sits proud of the surface, the underside is flat, and it works with almost anything. If you're buying screws to stock, start with pan head. Flat head is for when the screw needs to disappear — laptop bottom covers, panel mounts, anything where you're going to run your hand across the surface. The countersunk angle is standardized at ninety degrees for metric electronics. Button head is the aesthetic choice — lower profile than pan, more rounded. Three-D printer builders love them because they make a machine look intentional rather than assembled from hardware-store parts. The tradeoff is the socket is shallower, so you get less torque before the driver slips. And socket head cap screws are for when you need to really crank something down — motor mounts, structural brackets, anywhere vibration is a concern. The cylindrical head gives you deep hex engagement, so you can apply serious torque without stripping. They're overkill for mounting a circuit board, but for the extruder assembly on a three-D printer, they're exactly right.
Which brings us to drive types. And the Pozidriv situation.
The Pozidriv situation deserves its own warning label. Phillips was invented in the nineteen-thirties and the cam-out was intentional — Henry Phillips designed it so assembly-line workers couldn't over-torque screws on aircraft. Pozidriv came along in nineteen-sixty-six as an improvement specifically to prevent cam-out. It looks almost identical, but there are four extra radial lines stamped between the cross slots — little tick marks at forty-five degrees. If you see those lines, it's Pozidriv. Do not use a Phillips driver.
How do you tell if you don't have great lighting or your eyes aren't what they used to be?
The quick test: try a Phillips driver. If it seats fully and doesn't wobble, it's Phillips. If it feels like it's sitting on top of something and rocks slightly, you're probably in a Pozidriv screw and the driver isn't reaching the bottom of the recess. The Pozidriv driver has a blunter tip and parallel flanks — it'll seat deeper and won't rock.
Then Torx is the star-shaped one that's taking over.
Torx — technically hexalobular — is everywhere now in three-D printers, RC vehicles, and even some laptops. The six-pointed star distributes torque evenly and resists cam-out far better than Phillips. If you're building a Prusa or a Voron, you're using Torx screws. The sizes you'll encounter are T-eight, T-ten, T-fifteen, T-twenty, and T-twenty-five.
For Daniel's inventory, what's the minimum driver set?
A good Phillips set covering PH-zero, PH-one, and PH-two. A Pozidriv set — PZ-zero, PZ-one, PZ-two — which you can get as a single interchangeable bit set. A Torx set from T-eight through T-twenty-five. And a metric hex set from one-point-five millimeter through four millimeter for socket head cap screws. That's six or seven drivers total and it covers ninety-five percent of electronics fasteners.
Now the inventory system itself. Daniel's been trying to build one. What does a rational setup actually look like?
Organize by diameter first, not by project. The temptation is to keep all the screws from one laptop in a bag together — but that's how you end up with seventeen bags of mystery screws. Diameter-first means all your M-threes live together, then subdivided by length. I use small parts organizers with removable bins — the kind with the clear drawers. Each bin gets a label with the full spec: M-three by ten millimeter pan head Phillips.
The starter inventory?
Ten bins gets you operational. M-two-point-five by six, M-two-point-five by ten. M-three by six, eight, ten, twelve, sixteen, and twenty. M-four by ten and sixteen. And one bin for number-six-thirty-two by quarter-inch, which is the standard PC power supply length. Add a small assortment of washers — M-two-point-five, M-three, and M-four flat washers and split lock washers — plus M-three hex nuts and M-four nylock nuts. That's maybe thirty dollars in hardware and it covers eighty percent of what you'll ever reach for.
The other twenty percent?
Buy as needed. You cannot stock every variant — the combinatorics are infinite. M-two by three millimeter flat-head Torx exists, and someday you'll need exactly one. When that day comes, order a pack of ten, use one, label the other nine, and now your inventory grew by exactly the thing you actually use. Trying to pre-buy every possible combination is how you end up with drawers full of screws you'll never touch.
The eighty-twenty rule applied to fasteners. Stock the heavy hitters, let the edge cases find you.
Keep a mystery bin. A single bin, unlabeled, where unidentified screws go. When you have ten minutes and a caliper, work through it. Identify, label, file into the right bins. The mystery bin should be a temporary state, not a permanent feature.
The jar of shame, but with a path to redemption.
And that's the whole system. It's not about having every screw — it's about knowing what you have, where it is, and what to call the one you don't.
How do you turn all this into something that actually works on your bench tomorrow? Three concrete steps.
First one's cheap. Digital caliper, under twenty dollars. Thread pitch gauge, under ten. Those two tools together let you identify any mystery screw in about thirty seconds.
The workflow is dead simple. Measure the outer diameter with the caliper — that gives you your M-size. Measure the length from under the head — that's your screw length. Check the pitch against the gauge — match the teeth until they seat perfectly. Identify the head style and drive type by eyeballing it. Then look it up. That's it. You've gone from "some little silver screw" to "M-three by eight millimeter pan head Phillips" in half a minute.
Once you can name it, you can order it. That's the whole game.
Step two: build your starter inventory around the big three. M-three is your most-used size — stock six, eight, ten, twelve, sixteen, and twenty millimeter lengths, five to ten of each. M-two-point-five is your M-dot-two and Raspberry Pi workhorse — stock four, six, eight, ten, and twelve. And number-six-thirty-two for PC power supplies and US enclosures — quarter-inch and three-eighths-inch lengths cover most of what you'll encounter. Add M-four by ten and sixteen, and M-two by four and six for the odd jobs. That's maybe twelve bins total and it covers eighty percent of electronics repairs.
Step three is the one that costs nothing but pays off every time. Make a reference card.
Physical card, taped to your parts organizer. One side: M-size quick reference — M-two is hard drives and M-dot-two standoffs, M-two-point-five is SSDs and Pi boards, M-three is everything, M-four is structural, six-thirty-two is power supplies. Other side: Pozidriv versus Phillips identification — those radial tick marks — and the standard thread pitches. M-two is zero-point-four, M-two-point-five is zero-point-four-five, M-three is zero-point-five, M-four is zero-point-seven.
You'll reference it constantly for the first month, and then you'll have internalized it and won't need it anymore. That's the point.
Here's the thing Daniel was driving at — the meta-lesson. Fastener knowledge is not tribal lore passed down from some gray-bearded repair wizard. It's a learnable system. Once you understand the naming convention, you can look at any screw and deduce its specs. More importantly, you can buy replacements with confidence instead of squinting at a blurry Amazon photo and hoping.
The difference between "I think this looks right" and "I know this is right" is about thirty dollars in tools and an afternoon of setting up bins. That's it.
There's one question I keep coming back to. We've built this whole system around standardized M-sizes — and it works because the standard is stable. But with three-D printing and desktop CNC becoming accessible, are we going to see a shift toward custom fasteners designed for specific printed parts? Threads modeled right into the print, captive nuts in custom pockets, that kind of thing.
I think we'll see both. The standard M-sizes aren't going anywhere — the supply chain is too deep, and nobody wants to source a proprietary screw for a bracket they could have built with an M-three. But the custom stuff is already happening. Voron printers use heat-set inserts pressed into printed parts — you design the pocket for an M-three insert, melt it in with a soldering iron, and now your plastic part has metal threads. That's the hybrid approach. Standards where they make sense, custom integration where it adds value.
The taxonomy still matters, because even the custom stuff references back to it. You're still specifying M-three inserts. You still need to know the pitch.
The standard is the foundation. Custom is just building on top of it.
Which brings us back to where Daniel started. The next time a project stalls because you're missing one screw, you'll know exactly which screw you need, where to find it, and what to buy as a replacement. That's not magic. It's taxonomy.
Now: Hilbert's daily fun fact.
Hilbert: In the eighteen-sixties, a crofter on the Isle of Lewis in the Outer Hebrides uncovered a stone box containing a set of carved ivory pieces now known as the Lewis Chessmen — one of the few surviving artifact sets that preserves both logographic and phonetic marking conventions side by side, since some pieces bear runic inscriptions while others use purely symbolic identifiers.
...right.
This has been My Weird Prompts. If you enjoyed the episode, leave us a rating or review wherever you listen — it helps. Send your weird prompts to show at my weird prompts dot com. We're back next week.
