Chapter 15: Mechanical Primitives: Wheel, Axle, Screw, Gear, Bearing

Era span: ~3500 BCE → Renaissance mechanism culture · Difficulty: mid
Requires: Ch 12, Ch 14
Unlocks: Ch 16, Ch 18, Ch 20, Ch 23, Ch 29, Ch 45

Most machines can be understood through six classical simple machines plus transmissions and motion converters. This chapter is a working mechanical vocabulary, not a claim that every device is literally reducible to six parts. Fluency here lets Part III move quickly instead of re-deriving basics.

15.1 The Six Classical Simple Machines

1. Lever — trade distance for force around a fulcrum; F₁·d₁ = F₂·d₂. Crowbars, oars, pliers.
2. Wheel and axle — a rotary lever with an adjustable radius ratio. Windlasses, capstans, grindstones.
3. Pulley — changes force direction; a block-and-tackle trades pull distance for force, with real efficiency reduced by friction and line mass.
4. Inclined plane — a ramp trades travel distance for reduced force. Stairs, loading ramps, and shallow grades are applications.
5. Wedge — two inclined planes back to back; converts force into splitting, clamping, or lifting action. Wedges, knives, and splitting tools.
6. Screw — an inclined plane wrapped around a cylinder; converts rotation into large axial force or fine displacement. Presses, clamps, fasteners, and lathe leadscrews.

Derived mechanisms: gears transmit ratio and torque, cams shape motion, ratchets store motion intermittently, and crank-connecting-rod systems convert rotation to reciprocation. They are essential, but they are not substitutes for the six classical categories.

Mechanical advantage arithmetic is Ch 12 proportion in action; teach apprentices to compute before they build.

15.2 Wheels Done Right

• Rolling resistance beats dragging by ~10–50× on prepared surfaces — the entire economic case for roads (Ch 24).
• Spoked wheels cut rotating mass ~5× versus solid discs: faster animals, less axle wear, responsive war chariots. Build with felloes (rim segments), spokes mortised into hub, rawhide-bound while damp for shrink-fit tension.
• Axle bearings: plain bronze or iron bushings in hardwood hubs, greased with tallow. Friction loss in a decent plain bearing runs 2–5 % — acceptable until ball bearings below.
• Caster/roller intuition: logs under sledges are the zero-tech ancestor; keep rollers in the heavy-movement toolkit forever.

15.3 Power Transmission Elements

• Rope/cord drives: cheap, quiet, tolerant of misalignment; stretch limits precision.
• Flat belts (leather, later rubberized fabric): medium distances, moderate power; crowned pulleys self-track. Belting becomes THE factory transmission for a century once Ch 16/Ch 26 power sources exist.
• Gears: wooden lantern pinions (trundle wheels) meshing with wooden-cogged wheels serve mills for centuries; all-iron cut gears arrive with clockmakers' skill. Tooth profile perfection is NOT required early — crude gears run noisily but work if teeth are evenly spaced and mesh depth consistent.
• Crank and connecting rod: converts rotation↔reciprocation. Historically underexploited for a millennium after its first appearance; DO exploit it immediately — it is the key motion converter between waterwheels (rotary) and hammers, pumps, and saw blades (reciprocating).
• Escapements (clockwork): release stored energy in counted ticks — mechanical counting, previewing computation (Ch 36).

15.4 The Lathe: Mother of Machine Tools

Safety warning: rotating work grabs hands, hair, and rags and does not let go — belts, chucks, and flywheels maim on contact, and long cuts whip. Tie back hair and clothing, use guards and tool rests, stop the spindle before measuring or clearing swarf with hooks (never fingers), and prove chuck keys are out before starting.

A lathe spins workpiece against a cutting tool. With it you make: round shafts, pulleys, wheels, screws, cannon bores, telescope parts, engine cylinders. Without it, everything above stays artisan-lumpy.

Build order: 1. Pole/bow lathe (cord-wrapped spindle, spring return): turns wood well, metal poorly. 2. Treadle lathe with flywheel: continuous rotation, feet free hands. 3. Metal-turning lathe with leadscrew and slide rest: Maudslay's ~1800 configuration — tool fed by a screw against a rigid carriage. This one machine births interchangeable parts and every precision industry.

Key threshold: a lathe that holds 0.5 mm accuracy upgrades your whole civilization's artifacts; one holding 0.05 mm makes precision screws, valve gear, and interchangeable parts possible. Large cylinders are a boring-mill problem, not a lathe problem: Watt's engines (Ch 23) became practical once Wilkinson's 1774 boring mill held a bore of more than a metre true to roughly a millimetre — coarse by later standards, but enough to seal a piston. Precision propagates: lathes make better lathes.

15.5 Screws and Standard Threads

Hand-cut threads (chase over a turned mandrel, or die stocks) suffice early. Standardize thread pitches EARLY across workshops — bolts that fit only their mate are a tax on every future repair. Historical detour compressed by fiat: pick three coarse pitches (large/medium/small), publish them, enforce them. Whitworth did exactly this nationally in 1841 — constant-pitch series on a 55° thread flank angle, the first national screw-thread standard — and standard threads plus standardized measures made his later "millionth of an inch" measuring machines meaningful (Ch 20).

The gauge system that makes interchangeability real (§15.7 states the doctrine; this is its hardware):

1. Master artifacts first. Preserve one known-good example of each critical part, handled only with gloves and never used — it is the shop's reference against which everything else is checked.
2. Limit gauges per critical dimension. For each dimension that matters functionally, make a hardened-steel GO gauge at the maximum-material limit and a NO-GO gauge at the minimum-material limit. Plug gauges for holes, ring/snap gauges for shafts, thread plug/ring gauges for screws. A worker who cannot measure with rules can still sort perfectly: GO enters, NO-GO refuses.
3. The 10 % budget. Total gauge tolerance ≤10 % of the workpiece tolerance; reserve another ~10 % of the tolerance band as wear allowance on the GO member (it rubs every part; the NO-GO should barely ever touch). Audit gauges monthly against masters; retire GO members past their wear budget — a worn GO gauge passes loose parts forever after. Threads get their own GO/NO-GO ring (external) and plug (internal) gauges — they check fit, the only thing that matters on a bolted joint, while pitch itself is verified once at gauge manufacture against a screw-pitch gauge.
4. When gauge blocks become affordable: Johansson's wrung-stack gauge blocks (Sweden, 1896) gave absolute reference lengths in hardened lapped sets (47–87 blocks covering any dimension in steps; blocks adhere when wrung because the intervening ~25 nm fluid film lets molecular attraction act); they became ordinary toolroom equipment once lapped-hardened-steel production scaled, roughly by the 1920s–30s. Until then, master artifacts plus disciplined limit gauges carry a rebuild economy's entire metrology — which is precisely how the historical armory system ran.

Worked dimensioning example (one critical fit, done once properly, copied forever):

Fit: Ø25 mm running fit, hole in bracket Ø25.000–25.100 mm (+0.10/−0). Work tolerance W = 0.10 mm.

Gauge tolerances here are 0.004 wide — inside the 10 %-of-W budget (≤0.010) — with a further 0.010 wear allowance on the GO member. Both gauge zones sit inside the work tolerance (Taylor's principle): a few good holes right at the limits are rejected, deliberately, so that no bad hole is ever accepted. Gauges protect the assembler, not the machinist. The shaft mirrors this with snap or ring gauges at reversed limits. Make both members in hardened steel, grind, lap to blue-fit against the masters, and log monthly audits on the same charts as production (Ch 47 §47.13 treats gauge dimensions exactly like product dimensions).

Bootstrap problem, solved the historical way: your first limit gauges must be better than your first products, so make them from your best current output — turn soft blanks on the best lathe setting available, harden (water quench plain-carbon steel, then temper straw-color per Ch 14 §14.5), grind the measuring surfaces, and lap against fine abrasive until a master part blue-fits evenly. The gauge then defines production, not vice versa. Store gauges oiled and temperature-stable — flash rust ruins lapped reference surfaces faster than wear does — and rebuild the gauges whenever process capability improves (Ch 47 §47.13 capability arithmetic tells you when): each gauge generation tighter than the last is precision propagating upward through hardware.

Documented payoff, the armory precedent: John H. Hall at Harpers Ferry's Rifle Works (1820s) tooled every rifle part to limit gauges staffed by unskilled hands; in January 1826 he demonstrated before U.S. Ordnance officials that rifles could be disassembled, parts interchanged among them, reassembled, and fired — field repair by part-swap, no artisan required. Springfield adopted the practice, and the "armory system" diffused into sewing machines, clocks, bicycles, and automobiles. Retrofitting gauge discipline onto a craft industry took generations historically; adopt it while your industry is still small enough to enforce one standard.

Key threshold: when more than ~1 % of production fails a NO-GO gauge, the fault sits in machine wear or process drift, not worker sloppiness — fix the machine, chart the dimension (Ch 47 §47.13), and require tolerance ÷ 6σ̂ ≥ 1.33 before promising interchangeable parts at all.

15.6 Bearing and Friction Management

Friction engineering = free efficiency:

• Lubrication: tallow/vegetable oil films cut plain-bearing friction roughly in half and halve it again under pressure feeds. Grease cups (simple reservoirs dripping onto journals) automate care.
• Journal geometry: longer, larger-diameter bearings carry more load at lower unit pressure; keep surfaces smooth and aligned.
• Ball/roller bearings: rolling contact can reduce friction sharply, but manufacture, loading, lubrication, seals, and failure analysis are demanding. Once steel quality and precision support them, a design may skip crude rolling-element experiments—but it may not skip plain-bearing competence or containment.

15.7 Machine Tools and Interchangeable Parts

The lathe births a FAMILY of machines, each removing metal with controlled precision:

• Drill press (rotating bit fed vertically), milling machine (rotating cutter over fed workpiece — flat surfaces, slots, gears), planer/shaper (linear cutting strokes), grinder (abrasive finishing to a few micrometres).
• Each machine tool can make PARTS FOR BETTER MACHINE TOOLS — the self-improving loop that separates toolmaking cultures from craft cultures. Fund it continuously (Ch 23's cylinder tolerances depend entirely on this base).

Interchangeability doctrine: parts made to gauge-checked tolerance fit ANY assembly of their type without filing.

1. Define functional gauges (go/no-go rings and plugs) per dimension.
2. Machine to gauge; workers never measure with rules, only check against limits.
3. Repair becomes part-swap instead of artisan refit — armies and railways ran on this arithmetic.

Historical note: interchangeable parts originated in arms contracts (Gribeauval, Springfield) because militaries needed battlefield repair by unskilled hands; civilian industry adopted it once gauges standardized (Whitworth's thread standards, §15.5, which also gives the gauge hardware and the Harpers Ferry precedent).

15.8 Design Doctrine

1. Support geometry: a rigid body on a plane is statically determined by three non-collinear support points. A three-legged table does not rock under a level load, but uneven loads, joints, friction, and deformation can still move it; “three points” is not a universal stability theorem.
2. Triangulate loads — triangles resist deformation (Ch 12); rectangles rack.
3. Design for manufacture: if your shop can't hold the tolerance, the design is wrong regardless of elegance.
4. Fail visibly and safely: shear pins, breakaway mounts — machines should die loudly where they can't kill.
5. Size with a stated safety factor: stress is force ÷ area, and a part's working stress must sit below the material's tested strength by a factor that covers uncertain loads, variable material, wear, fatigue, and the consequence of failure. Historical machine-design practice ran from about two for well-characterised steel under steady load to six or more for brittle cast iron or shock loading; the applicable code, not habit, sets the number for anything that can hurt people (Ch 27 §27.11).

The civilization that internalizes these sections has absorbed, in one generation, the mechanical vocabulary that took history five millennia to compile.

15.9 The Ancient Engineering Record

Precision mechanics is far older than clocks:

• Antikythera mechanism (~150–100 BCE): at least 30 interlocking bronze gears computing eclipse cycles and calendrical dials. Derek de Solla Price's 1974 reconstruction began the field; 2005–06 CT imaging revealed hidden gear trains and month names. One surviving artifact implies an unbroken craft tradition behind it — survivors' bias runs backwards in mechanism history.
• Hellenistic automata workshop: Ctesibius (water clocks, force pump), Philon, Hero of Alexandria (aeolipile reaction sphere, coin-operated holy-water dispenser) — pneumatics and hydraulics engineered for temples and wonder; the same physics later powers turbines (Ch 16, Ch 26).
• Vitruvius (~25–15 BCE) documents treadwheel cranes, hodometers, and hoists in De architectura — single-treadwheel cranes lifted on the order of 6 tonnes; the Baalbek trilithon blocks (~800 t each) moved WITHOUT cranes, by sledges, rollers, levers, and brute coordination — matching tool class to load, not forcing one tool everywhere.
• Su Song's astronomical clock tower (1092 CE) integrated a water-driven chain drive with an escapement-like regulator — mechanical timekeeping sophistication centuries before Europe's weight-driven clocks, built for state astronomy rather than commerce.
• Bearings history honestly told: roller-and-sledge transport is documented in Egyptian tomb scenes; true ball bearings await machined races (Leonardo sketched them, 1490s); antiquity ran on plain journals + grease because manufacturing tolerance, not insight, was the binding constraint (§15.6).

Lesson for rebuilders: mechanical genius recurs across civilizations whenever machining culture exists; the machine-tool chapter is where ancient genius and industrial output actually connect.

15.10 Engineering Drawings: The Shop's Common Language

Interchangeable parts (§15.5, §15.7) need one more standard beside threads and gauges: a drawing that every shop reads the same way.

• Orthographic projection: each view (front, top, side) shows the part square-on, aligned so that every feature lines up from view to view, with sections cut through to show hidden interiors. Gaspard Monge's descriptive geometry, taught at the new École Polytechnique from the 1790s, turned this into a teachable discipline. Choose first-angle (European custom) or third-angle (American custom) projection once, and print its symbol on every sheet.
• Dimensions and tolerances: every functional size carries its limits (Ø25.000–25.100 in §15.5). Dimension from the datum surfaces the part is actually machined and inspected from; never chain critical dimensions end to end, where the errors accumulate.
• Title block and revision control: part name, material, finish, scale, drawn-by and checked-by names, and a revision letter. The shop builds only to the current revision, and superseded prints are withdrawn and archived, not left on benches (Ch 11 §11.3).
• Copies that match the master: hand tracings were multiplied by blueprint (cyanotype) from the 1870s (Ch 21 §21.9), so one checked master could reach every shop unchanged.

Key threshold: two shops, working only from the same drawing, make parts that pass the same gauges and assemble without filing. Until that happens, the real specification lives in one machinist's head.