The Screw Gauge

The micrometer screw refuses to retire. Aerospace shops and EV battery lines still reach for a C-frame micrometer — now digital, resolving $1\ \mathrm{\mu m}$ — because a rigid frame, a hardened anvil and a precision thread remain the cheapest trustworthy path to the hundredth of a millimetre, each instrument tracing its honesty up a calibration chain to gauge blocks: "check your zero error" scaled to a national metrology institute. The screw's deeper legacy is as an actuator. Micrometer heads on optical tables, microscope fine-focus, the coarse approach that lowers an atomic-force microscope's tip without crashing it — all are Gascoigne's idea in reverse: a big rotation commanding a tiny, controlled advance. Even in 2026, with fabs measuring transistors at nanometre scales, the stages carrying those wafers are positioned by descendants of the laboratory screw gauge.

Hold a screw gauge: a C-shaped steel frame with a measuring axis across its mouth. On the left tip sits the anvil, a polished fixed face; facing it is the spindle, the moving end of a precision screw. The screw threads through a nut inside the sleeve, which carries the main scale — millimetre marks above an engraved reference line, half-millimetre marks below it. Around the sleeve turns the thimble, whose bevelled edge carries the circular scale, usually fifty divisions. At the far end sits the ratchet, a cap that clicks and slips when the gripping force exceeds a set value.

Everything the instrument can do follows from one machined promise: the thread is uniform. Equal turns produce equal advances; one-fiftieth of a turn produces one-fiftieth of a pitch. That proportionality lets us subdivide a rotation instead of a length — engraving fifty ticks around a $56\ \mathrm{mm}$ rim is easy, while engraving hundredth-of-a-millimetre ticks along a line would be unreadable. Feel the amplification: a thimble of radius $r \approx 9\ \mathrm{mm}$ sweeps $2\pi r \approx 56.5\ \mathrm{mm}$ of rim per revolution while the spindle advances $p = 0.5\ \mathrm{mm}$ — a factor of about $113$. The circular divisions sit a luxurious $1.1\ \mathrm{mm}$ apart on the rim while each stands for just $0.01\ \mathrm{mm}$ of advance.

Before any measurement, an examiner (and a good experimenter) establishes the instrument's two defining constants. The pitch is the distance the spindle advances in one complete rotation. You never trust the label: you measure it, by giving the thimble a known number of full rotations and reading the advance on the main scale. Five rotations carrying the thimble edge $2.5\ \mathrm{mm}$ down the sleeve mean

where: $p$ — the pitch, the spacing between adjacent turns of the helical thread; a spindle is a helix fed through a fixed nut, so one revolution advances it by exactly one thread spacing. Laboratory gauges have $p = 0.5\ \mathrm{mm}$ or $1\ \mathrm{mm}$ — mixing them up is a classic ambush. The least count is the smallest advance the instrument can resolve — the linear worth of one circular-scale division. Since one full turn ($N$ divisions passing the reference line) advances the spindle by $p$, each division corresponds to

where: $N$ — the number of divisions around the thimble (usually 50, sometimes 100). Read it physically: the screw divides the millimetre mechanically (into pitches), the circular scale divides the pitch optically (into divisions) — resolution manufactured in two stages. Pitch $1\ \mathrm{mm}$ with $N = 100$ gives the same $0.01\ \mathrm{mm}$; with $N = 50$, a coarser $0.02$. Compute LC from the stem's numbers, never from memory.

A reading is assembled from two scales, then corrected. Close the spindle onto the object on the ratchet, two or three clicks, and read:

where: MSR — the last sleeve graduation fully visible at the thimble's edge, including the half-millimetre marks below the line; $n$ — the circular division coinciding with the reference line. The half-mm clause deserves respect: record MSR $= 6$ when the edge has passed the $6.5$ half-mark and your answer is wrong by exactly $0.5\ \mathrm{mm}$ — the most common screw-gauge blunder in board practicals.

Zero error is the systematic offset the instrument carries when it should read zero. Bring anvil and spindle into contact (on the ratchet). If the circular zero sits exactly on the reference line, the gauge is honest. If not:

Positive zero error — the circular zero has sunk below the reference line, which now crosses the scale at some small division $n_0$ (say the 3rd). The count started early: the gauge reads high. Zero error $= +\,n_0 \times \mathrm{LC}$.

Negative zero error — the circular zero rides above the line, which coincides with a division $n_c$ near the top (say the 46th). The gauge reads low by the shortfall. Zero error $= -\,(N - n_c) \times \mathrm{LC}$. Both cases collapse into one universal correction — commit only this to memory:

where: the sign does the thinking — subtracting a positive error removes the head start; subtracting a negative error adds back the shortfall. A gauge with zero error $-0.04\ \mathrm{mm}$ reading $1.78\ \mathrm{mm}$ means the wire is truly $1.82\ \mathrm{mm}$.

Two habits separate a trustworthy reading from a lucky one; both exist because real screws are imperfect.

Backlash. The screw must slide inside its nut, so there is a whisker of clearance between their threads. Turning steadily one way keeps the thread faces pressed together and the proportionality of the derivation holds exactly. Reverse direction and the thimble rotates through the slack before the opposite faces engage — the scale turns while the spindle stands still, shifting readings by a few hundredths. The cure is procedural: always approach the final setting rotating the same way; if you overshoot, back off well past the mark and come in again.

The ratchet. Squeezing changes what you measure — a brass wire compresses by measurable hundredths under an enthusiastic thumb. The ratchet is a torque limiter: beyond a set gripping force it slips with a click instead of passing your impatience to the wire. Advance with the ratchet only, stop after two or three clicks, every time. Same force, every reading — that is what makes measurements comparable.

The third row is the chapter's quiet lesson: least count is absolute resolution, but experiments live by relative error — $0.01\ \mathrm{mm}$ is superb on a $5\ \mathrm{mm}$ rod ($0.2\%$), mediocre on a $0.1\ \mathrm{mm}$ foil ($10\%$).

Board level — pitch, least count, and one corrected reading (ICSE/CBSE practical pattern)

The thimble of a screw gauge is given 5 complete rotations, during which its edge moves through $2.5\ \mathrm{mm}$ on the main scale. The circular scale has 50 divisions. With the faces closed, the circular zero sits below the reference line and the 3rd division coincides with it. When a wire is gripped (ratchet clicking), the main scale reads $2.5\ \mathrm{mm}$ and the 43rd circular division coincides. Find the pitch, the least count, and the corrected diameter of the wire.

Step 1. Pitch from the rotation test: $p = \dfrac{2.5\ \mathrm{mm}}{5} = 0.5\ \mathrm{mm}$.

Step 2. Least count: $\mathrm{LC} = \dfrac{p}{N} = \dfrac{0.5}{50} = 0.01\ \mathrm{mm}$.

Step 3. Zero error: zero below the line, line at division 3 → positive error, $+3 \times 0.01 = +0.03\ \mathrm{mm}$.

Step 4. Observed reading: $\mathrm{MSR} + n \times \mathrm{LC} = 2.5 + 43 \times 0.01 = 2.93\ \mathrm{mm}$.

Step 5. Correct with the universal rule: corrected $= 2.93 - (+0.03) = \boxed{2.90\ \mathrm{mm}}$. The marks-scheme sentence worth writing: "Positive zero error means the instrument reads high, so the error is subtracted."

Competitive level — the negative zero error (JEE-Main pattern)

A screw gauge has pitch $0.5\ \mathrm{mm}$ and 50 circular divisions. With the flat faces in contact, the 46th division coincides with the reference line, the zero lying above it. The gauge then reads $\mathrm{MSR} = 1.5\ \mathrm{mm}$ with the 28th division coinciding when a wire is gripped. The wire's diameter is…

Step 1. Identify the error type before computing: zero above the line → negative error, and the gauge reads low.

Step 2. Magnitude from the shortfall, not the coinciding division: zero error $= -(N - n_c)\,\mathrm{LC} = -(50 - 46)(0.01) = -0.04\ \mathrm{mm}$. (Writing $-0.46$ here is the classic wrong option.)

Step 3. Observed reading: $1.5 + 28 \times 0.01 = 1.78\ \mathrm{mm}$.

Step 4. Correct: $d = 1.78 - (-0.04) = \boxed{1.82\ \mathrm{mm}}$. Timing note: 60–90 seconds. The only decision is Step 2 — drill "above the line → count from 50" to reflex; papers love quoting the error in cm ($-0.004\ \mathrm{cm}$) to buy a free unit slip.

Competitive level — from least count to the error budget (JEE-Main/Advanced pattern)

A wire's diameter, zero error corrected, is $d = 0.500\ \mathrm{mm}$ (LC $0.01\ \mathrm{mm}$). Its resistance is known to $1\%$, its length to $0.2\%$. Find the diameter's contribution to the cross-sectional area's uncertainty, and the maximum percentage error in $\rho = \dfrac{\pi d^2 R}{4 L}$.

Step 1. The instrument's contribution to the diameter: $\dfrac{\Delta d}{d} = \dfrac{0.01}{0.500} = 0.02 = 2\%$.

Step 2. Area goes as $d^2$, so its relative error doubles the diameter's: $\dfrac{\Delta A}{A} = 2\dfrac{\Delta d}{d} = 4\%$.

Step 3. Maximum error in resistivity adds the independent contributions: $\dfrac{\Delta \rho}{\rho} = 2\dfrac{\Delta d}{d} + \dfrac{\Delta R}{R} + \dfrac{\Delta L}{L} = 4 + 1 + 0.2 = \boxed{5.2\%}$.

Step 4. Read the budget like an experimenter: the screw-gauge term is four-fifths of the total. Halving $\Delta d$ helps enormously; halving $\Delta L$ is pointless. The diameter — squared in the formula and the smallest length measured — is the experiment's soul. Timing note: about 90 seconds; the physics is Step 2's factor of 2. Powers multiply relative errors — $\rho \propto d^2 \Rightarrow 2\,\Delta d/d$; the rest is addition.

The sign of the zero-error correction. One rule — corrected $=$ observed $-$ signed error — and both wrong signs sit among the options. Positive (zero below the line): subtract. Negative (above): add.

The hidden half-millimetre. The thimble edge can sit just past a half-mm mark while concealing it; reading $6.0$ for $6.5$ is wrong by exactly $0.50\ \mathrm{mm}$. Two students differing by exactly half a millimetre is this trap's signature.

Assuming the standard pitch. $p = 0.5$, $N = 50$ is common, not law: papers set $p = 1\ \mathrm{mm}$ with $N = 100$ (same LC) or $N = 50$ (LC $= 0.02$!). Compute LC from the stem, never recall it.

Unit mischief. Zero errors quoted in cm ($-0.004\ \mathrm{cm} = -0.04\ \mathrm{mm}$), areas wanted in $\mathrm{m^2}$ from readings in mm. Convert once, at the start, in writing.

Counting the coinciding division for a negative error. The 46th division coinciding does not make the error $0.46\ \mathrm{mm}$; the error is the gap between zero and the line, $(50-46)$ divisions $= 0.04\ \mathrm{mm}$.

Reporting beyond the least count. An LC-$0.01$ instrument yields $1.82\ \mathrm{mm}$, never $1.8200\ \mathrm{mm}$. And in the practical hall: approach from one direction, finish on the clicks — backlash and squeeze are invisible in a write-up but fully visible in the result.

Texts. NCERT Physics Laboratory Manual (Class XI), screw-gauge experiment — the canonical CBSE procedure · Selina Concise Physics Class 9, Ch. 1 — the ICSE treatment of pitch, least count and zero error · JEE-Main archives, Experimental Skills section (screw-gauge numericals in most papers, 2011–2025).

Provenance. Old-HS source mined for narrative; all conventions per NCERT lab manual and ICSE Class 9 canon. A verifier should recompute: $p = 2.5/5 = 0.5\ \mathrm{mm}$; $\mathrm{LC} = 0.5/50 = 0.01\ \mathrm{mm}$; amplification $2\pi \times 9 / 0.5 = 113$; rim spacing $56.55/50 = 1.13\ \mathrm{mm}$; board example $2.5 + 43(0.01) = 2.93 \to 2.93 - 0.03 = 2.90\ \mathrm{mm}$; negative zero error $-(50-46)(0.01) = -0.04\ \mathrm{mm}$ and $1.78 - (-0.04) = 1.82\ \mathrm{mm}$; error budget $2(0.01/0.500) = 4\%$ and $4 + 1 + 0.2 = 5.2\%$; quiz 5: $6.5 + 0.32 + 0.02 = 6.84\ \mathrm{mm}$.

Prerequisites. Least Count & Instrument Errors · The Vernier Calipers. Leads to. Error-Question Patterns (where zero errors become 30-second MCQs) · deep dive: Accuracy, Precision & Uncertainty.