Physics and Measurement

Every equation in physics carries a hidden passport — its dimensions — and checking that passport takes just seconds but can catch errors that hours of algebra miss. Dimensional analysis is one of the oldest tools in the physicist's kit, used by Newton, Fourier, and Rayleigh to deduce laws before they were fully derived.

Definition — Dimension: the expression of a physical quantity in terms of the base quantities: mass M, length L, time T, electric current A, temperature K, amount of substance mol, and luminous intensity cd.

Definition — Dimensional formula: the algebraic expression showing the powers of base dimensions in a derived quantity. For example, velocity = L T⁻¹, force = M L T⁻², energy = M L² T⁻².

Definition — Dimensionless quantity: a quantity whose dimensional formula is M⁰ L⁰ T⁰ — it is a pure number. Examples: angle (radians), refractive index, strain, relative density, Reynolds number, π.

Dimensional Formulas of Common Quantities

Memorising these is mandatory for JEE, NEET, and RRB competitive exams.

Quantity	Formula	Dimensions
Velocity	distance/time	L T⁻¹
Acceleration	velocity/time	L T⁻²
Force	mass × acceleration	M L T⁻²
Work / Energy	force × distance	M L² T⁻²
Power	work/time	M L² T⁻³
Pressure	force/area	M L⁻¹ T⁻²
Momentum	mass × velocity	M L T⁻¹
Angular momentum	mass × velocity × radius	M L² T⁻¹
Torque	force × arm	M L² T⁻²
Gravitational constant G	from F = Gm₁m₂/r²	M⁻¹ L³ T⁻²
Planck's constant h	from E = hf	M L² T⁻¹
Coefficient of viscosity	from F = η A dv/dx	M L⁻¹ T⁻¹
Surface tension	force/length	M T⁻²

What Dimensional Analysis CAN Do

1. Check the Dimensional Consistency of an Equation

A physically correct equation must be dimensionally homogeneous: every additive term must have the same dimensions, and the two sides must match. If they don't, the equation is definitely wrong. If they do, it is possibly correct (you still need physics to confirm).

Example — check v = u + at:

• [v] = L T⁻¹
• [u] = L T⁻¹
• [at] = (L T⁻²)(T) = L T⁻¹ ✓

All terms have dimensions L T⁻¹. The equation passes the dimensional check.

Example — check KE = ½mv²:

• [½mv²] = M (L T⁻¹)² = M L² T⁻² = [Energy] ✓

2. Convert Units Between Systems

The value of any physical quantity in a new unit system equals its value in the old system multiplied by the ratio of the old unit to the new unit for each dimension.

Example — convert 1 joule to CGS (ergs):
1 J = 1 kg·m²·s⁻²

In CGS: 1 kg = 10³ g, 1 m = 10² cm, 1 s = 1 s.

So 1 J = (10³ g)(10² cm)²(s)⁻² = 10³ × 10⁴ g·cm²·s⁻² = 10⁷ ergs.

3. Derive Relationships (Up to a Constant)

This is the most powerful application and the one most tested. If you know which physical quantities govern a phenomenon, you can set up a power-law ansatz and solve for the exponents from dimensional consistency.

Classic example — time period of a simple pendulum:

Assume T depends on length L, mass m, and g:

T = k · Lᵃ · mᵇ · gᶜ

Writing dimensions on both sides:

[T¹] = [L]ᵃ [M]ᵇ [L T⁻²]ᶜ

T¹ M⁰ L⁰ = Mᵇ · L^(a+c) · T^(-2c)

Equating exponents:

• T: 1 = −2c → c = −½
• M: 0 = b → b = 0
• L: 0 = a + c → a = ½

Result: T = k √(L/g). The constant k = 2π emerges only from the full Newtonian analysis; dimensions cannot find it.

Why this matters: If an exam gives you four quantities and asks "which combination has dimensions of energy?" you can evaluate each using this method in under a minute.

What Dimensional Analysis CANNOT Do

1. Find Dimensionless Constants

The formula T = 2π√(L/g) contains 2π. Rayleigh's method finds the √(L/g) part but is blind to 2π. Similarly, the ½ in KE = ½mv² and the factor 4 in Coulomb's law are invisible to dimensions.

Implication: dimensional analysis gives proportionalities, not equations. Never claim a derived proportionality is a complete formula without experimental confirmation.

2. Distinguish Between Quantities of the Same Dimension

Work and torque both have dimensions M L² T⁻². They are physically different (work is a scalar, torque is a pseudo-vector), but their dimensional formulas are identical. A dimensional check cannot tell them apart.

Other pairs with shared dimensions:

• Pressure and bulk modulus (M L⁻¹ T⁻²)
• Angular momentum and Planck's constant (M L² T⁻¹)
• Frequency and decay constant (T⁻¹)

3. Handle Transcendental Functions

Arguments of sin, cos, ln, eˣ, and exponentials must be dimensionless. You cannot apply dimensional analysis to separate parts of an equation like x = A sin(ωt + φ), because the argument (ωt + φ) must itself be dimensionless. If you try to balance dimensions, the function-within-an-equation blocks you.

Rule: if any term in an equation involves a trigonometric, logarithmic, or exponential function, the argument must be dimensionless. This is a quick way to catch errors — e.g., writing sin(velocity) in an equation is dimensionally impossible.

4. Determine Which Physical Quantities Govern a Phenomenon

Dimensional analysis assumes you already know the relevant variables. If you leave out the correct variable or include an irrelevant one, the derived relationship will be wrong. This requires physical intuition, not algebra.

Worked Examples

Question: Using dimensional analysis, find the velocity of a wave on a string if it depends only on tension F (dimension M L T⁻²) and linear mass density μ (dimension M L⁻¹).

Step 1: Assume v = k Fᵃ μᵇ.

Step 2: [L T⁻¹] = [M L T⁻²]ᵃ [M L⁻¹]ᵇ = M^(a+b) · L^(a−b) · T^(−2a)

Step 3: Equating:

• T: −1 = −2a → a = ½
• L: 1 = a − b = ½ − b → b = −½
• M: 0 = a + b = ½ − ½ ✓

Conclusion: v = k √(F/μ). The true formula is v = √(F/μ) — dimensional analysis got the structure exactly right (here k = 1 happens to be confirmed by full analysis).

---

Question: Check whether the equation P = ρgh is dimensionally correct (P = pressure, ρ = density, g = acceleration due to gravity, h = depth).

Step 1: [P] = M L⁻¹ T⁻²

Step 2: [ρgh] = (M L⁻³)(L T⁻²)(L) = M L⁻¹ T⁻² ✓

Conclusion: The equation is dimensionally consistent.

Common Misconceptions

Misconception 1: "If an equation is dimensionally correct, it must be physically correct."
Not true. The equation v = u − at is dimensionally correct (same dimensions as v = u + at) but physically wrong for uniform acceleration in the direction of motion. Dimensional analysis is a necessary but not sufficient condition.

Misconception 2: "Dimensionless means zero value."
No. Dimensionless means the dimension is M⁰L⁰T⁰ = 1 (the number 1, not zero). Angles, ratios, Reynolds numbers — all have definite non-zero values; they just don't carry units.

Misconception 3: "You can always use dimensional analysis for any formula."
Only when the relationship is a power law. Formulas involving sums/differences of different functional forms (like position x = ½at² + vt) can be individually checked term by term but cannot be derived as a whole by this method.

:::compare "Powers of Base Quantities — Quick Lookup"

Quantity	M	L	T
Force	1	1	−2
Energy / Work / Torque	1	2	−2
Power	1	2	−3
Pressure / Stress	1	−1	−2
Momentum	1	1	−1
Gravitational constant G	−1	3	−2
Planck's constant h	1	2	−1
Charge	0	0	1 (× A)
:::

:::keypoints Key points

• Dimensions express a quantity in terms of M, L, T, A, K, mol, cd.
• A valid equation must be dimensionally homogeneous — all additive terms share the same dimensions.
• Dimensional analysis derives power-law relationships up to a dimensionless constant.
• It cannot find pure numbers like 2π, ½, or 4π.
• Work and torque share M L² T⁻² — dimensions alone cannot distinguish them.
• Arguments of sin, log, and exponential functions must always be dimensionless.
• Dimensional analysis checks consistency and guides derivations but does not replace full physical reasoning.
• Converting between SI and CGS uses the ratio of old-to-new base units raised to the power given by the dimensional formula.
  :::

:::memory
"MLCT" — Mnemonic for what dimensional analysis Can and cannot do:
Match equations (yes) · Limit: no constants (no) · Convert units (yes) · Transcendental functions blocked (no).
Also: "Dimensions are the skeleton — flesh (constants, signs, functions) must come from physics."
:::

:::recap

• Dimensional formula expresses any derived quantity in powers of M, L, T (and other base dimensions).
• Homogeneity check: every term in an equation must have identical dimensions.
• Pendulum derivation: T ∝ √(L/g) — mass drops out because b = 0; constant 2π not obtainable.
• Two limits: (1) dimensionless constants invisible, (2) same-dimension quantities indistinguishable.
• Transcendental-function arguments must be dimensionless — this is a quick error-trap.
• Dimensional analysis is a filter, not a proof: passing it is necessary but not sufficient for correctness.
  :::

Significant figures (sig figs) express measurement precision. Rules: (1) All non-zero digits are significant. (2) Zeros between non-zeros are significant (e.g. 1002 has 4). (3) Leading zeros are NOT significant (0.0025 has 2). (4) Trailing zeros after a decimal point ARE significant (2.300 has 4). (5) Trailing zeros in a whole number without a decimal are ambiguous (1500 is treated as 2 sig figs unless written 1.500x10^3). Memory aid: 'Sandwiched zeros count, leading zeros don't, trailing-after-decimal do.' Scientific notation removes ambiguity: only the mantissa digits count. In JEE, exact constants (like the 2 in 2pi) have infinite sig figs and never limit precision. Always identify sig figs before applying rounding rules in numerical problems.

In a JEE problem, you can measure length to the nearest millimeter and time to the nearest hundredth of a second, but the moment you start combining these into a derived quantity like acceleration or density, those tiny measurement errors propagate. Knowing exactly how they propagate — and which formula to use when — is one of the cheapest topics to ace in the Physics and Measurement chapter.

Definition: An error in measurement is the difference between the measured value and the true value of a physical quantity.

Definition: Error propagation is the calculation of how errors in measured quantities (A, B, C ...) carry through to give the error in a derived quantity (Z) that depends on them.

The two big rules at a glance

Across all of error analysis, there are only two basic patterns to remember. Memorise these and you can derive everything else.

• For addition and subtraction (Z = A + B or Z = A - B): the absolute errors add.
• For multiplication and division (Z = AB or Z = A/B): the relative (fractional) errors add.

These two ideas drive every error MCQ in JEE Main.

Rule 1 — Addition and Subtraction (absolute errors add)

If Z = A + B or Z = A - B, then:

delta Z = delta A + delta B.

The plus sign is always used. Even for a subtraction, the maximum possible error is obtained by adding the magnitudes — this is the worst-case scenario.

Why we add and not subtract: if A is too high by delta A and B is too low by delta B, then A + B can be off by delta A + delta B in the same direction. If we wrote a minus sign, we would be hoping that the two errors cancel — that is a lucky case, not the maximum possible error.

Rule 2 — Multiplication and Division (relative errors add)

If Z = AB or Z = A / B, then:

delta Z / Z = delta A / A + delta B / B.

Again, the plus sign is always used, regardless of whether it is multiplication or division.

Definition: The relative (fractional) error in a quantity A is delta A / A — it is a dimensionless ratio. Multiplying by 100 gives the percentage error.

So a 2% error in A combined with a 3% error in B gives a 5% error in Z = AB or Z = A/B.

Rule 3 — Powers (errors multiply by exponents)

If Z = A^p B^q / C^r, then:

delta Z / Z = |p| (delta A / A) + |q| (delta B / B) + |r| (delta C / C).

The exponents come out as multipliers of the relative errors, and we use absolute values of the exponents because we are after the maximum possible error.

This is the rule that creates "shock" answers. A quantity raised to a high power dominates the total error. If your formula is Z = A^3 B, a 2% error in A becomes 6% in Z, while a 2% error in B stays at 2% — the error in A is three times more important than the error in B.

Worked example — density measurement

Question: The density of a metal cube is to be found from its mass M and side L. The relation is rho = M / L^3. The percentage errors are 1% in M and 2% in L. What is the maximum percentage error in rho?

Solution:
Step 1: Write rho in power form — rho = M^1 L^(-3).
Step 2: Apply the powers rule: delta rho / rho = 1 x (delta M / M) + |-3| x (delta L / L).
Step 3: Substitute: delta rho / rho = 1 x 1% + 3 x 2% = 1% + 6% = 7%.
Conclusion: The maximum percentage error in the density is 7%, dominated by the 6% contribution from L.

This single example captures the central JEE insight: in a derived quantity, the largest power-times-error term dominates, so reducing error in that variable is the highest-payoff design choice in an experiment.

Why we always use the worst case

A natural question is: why must we always add the magnitudes and not let them partly cancel? The answer is that the maximum permissible error is a guarantee — it tells the examiner (and the experimenter) what the worst possible deviation is, regardless of the signs of individual errors. Probabilistic cancellation is treated separately as statistical error, not here. JEE Main always asks for the maximum error unless explicitly stated otherwise.

Rounding off the final answer

After applying the rules, the final answer must be rounded sensibly:

• For addition and subtraction results: round to the least number of decimal places among the inputs.
• For multiplication and division results: round to the least number of significant figures among the inputs.

So if you add 12.3 m and 4.56 m, you write 16.9 m (one decimal place), not 16.86 m. If you multiply 1.234 cm by 2.0 cm, you write 2.5 cm^2 (two significant figures), not 2.468 cm^2.

Real-world example

Real-world example: In a JEE Main lab to find the value of g using a simple pendulum, the formula g = 4 pi^2 L / T^2 involves T raised to the power -2. If a student's stopwatch error is 0.1 s in a 20 s period (about 0.5%), it contributes 1% to the error in g. A 1 mm error in a 1 m length (0.1%) contributes only 0.1%. Conclusion: a more accurate stopwatch matters far more than a more accurate ruler in this experiment — straight out of error propagation.

Common misconception

Common misconception: "For subtraction, absolute errors should be subtracted to find the error." Wrong. Absolute errors are always added, even for subtraction. The worst case is when both errors push in the same direction.

Another wrong idea: "Higher powers reduce errors because the result becomes precise." Wrong — exactly the opposite. Higher powers multiply the relative error of that quantity, making them the dominant error contributors.

A third trap: "Percentage error = absolute error." Wrong. Percentage error = (delta Z / Z) x 100 — it is a relative quantity. Absolute error is in the same units as the quantity itself.

Worked example — combining rules

Question: A physical quantity P is given by P = (A^2 B^3) / (sqrt(C) D). The percentage errors are: A = 1%, B = 2%, C = 4%, D = 3%. Find the maximum percentage error in P.

Solution:
Step 1: Identify exponents — A: 2, B: 3, C: 1/2, D: 1 (all in absolute value).
Step 2: Apply the powers rule: delta P / P = 2(delta A / A) + 3(delta B / B) + (1/2)(delta C / C) + 1(delta D / D).
Step 3: Substitute: delta P / P = 2(1) + 3(2) + (1/2)(4) + 1(3) percent = 2 + 6 + 2 + 3 = 13%.
Conclusion: The maximum percentage error in P is 13%, with B (6%) the dominant contributor.

This is the structure of nearly every JEE Main error MCQ — assign each variable an exponent, multiply by its percentage error, then add.

:::compare

Operation	Form	Error rule
Addition	Z = A + B	delta Z = delta A + delta B
Subtraction	Z = A - B	delta Z = delta A + delta B
Multiplication	Z = AB	delta Z / Z = delta A / A + delta B / B
Division	Z = A / B	delta Z / Z = delta A / A + delta B / B
Powers	Z = A^p B^q / C^r	delta Z / Z = |p|(dA/A) + |q|(dB/B) + |r|(dC/C)
:::

Why it matters

Why it matters: Error propagation is the most predictable scoring topic in the Units and Measurement chapter — JEE Main asks one question on it almost every year, and it follows one of the patterns above. Practising 5-10 power-rule problems builds enough speed to solve any such MCQ under 60 seconds.

:::keypoints

• For addition and subtraction, absolute errors add.
• For multiplication and division, fractional (relative) errors add.
• For powers, multiply each fractional error by the absolute value of its exponent and then add.
• Always use worst-case (sum), never subtraction of errors.
• Percentage error = (delta Z / Z) x 100.
• Round addition/subtraction answers to the least decimal places; multiplication/division answers to the least significant figures.
• A high power makes that variable's error dominate the total — focus measurement effort there.
  :::

:::memory
"Add, Add, Multiply" — Addition rule adds absolute errors; multiplication rule adds relative errors; power rule multiplies relative errors by the exponent.
"Power dominates" — wherever the exponent is biggest, the error from that variable is biggest.
:::

:::recap

• Two clean rules cover almost everything: sums use absolute errors, products use relative errors.
• Powers turn into multipliers on the relative error — design your experiment to keep the high-exponent variable as precise as possible.
• Always use the maximum-error (worst-case) version, then round the final answer using the right significant-figure rule.
• One question, one rule, one minute — the highest reward-per-effort topic in the Measurement chapter.
  :::

A small cube on a lab bench seems like an innocent object — until the question asks for the percentage error in its density. This is one of the cleanest places to see why JEE Main keeps recycling error analysis: a tiny relative error in the side length quietly becomes the biggest contributor to the final uncertainty, simply because volume depends on the cube of length.

Definition: Absolute error in a measurement x is the smallest amount by which the measured value can be wrong, written as Δx. Relative error is Δx/x, and percentage error is (Δx/x) × 100%.

Definition: Propagation of error is the rule that tells you how uncertainty in input quantities flows into a quantity computed from them. For a product or quotient like q = A^p · B^q · C^r, the relative errors add — and each is multiplied by the power.

Setting up the problem

We are given a cube with measured mass m = 10.0 g carrying an absolute error Δm = 0.1 g, and measured side L = 2.00 cm with ΔL = 0.01 cm. We need the density ρ = m/V and the percentage uncertainty in it.

Since the cube has side L, its volume is V = L³, so density becomes:

ρ = m / L³ = m · L^(-3)

This is a product-of-powers expression, exactly the form where the propagation rule applies cleanly.

Step-by-step computation

Question: Find ρ and its percentage error for the cube above.

Solution:

Step 1 — Compute the nominal density.
ρ = 10.0 / (2.00)³ = 10.0 / 8.00 = 1.25 g/cm³.

Step 2 — Apply the propagation rule. Because ρ = m · L^(-3), the magnitudes of the relative errors add, with each multiplied by the absolute value of its exponent:

Δρ/ρ = Δm/m + 3 · (ΔL/L)

Step 3 — Plug in the numbers.
Δm/m = 0.1 / 10.0 = 0.010
3 · ΔL/L = 3 × (0.01 / 2.00) = 3 × 0.005 = 0.015
Δρ/ρ = 0.010 + 0.015 = 0.025

Step 4 — Convert to percentage and find Δρ.
Percentage error = 2.5%.
Δρ = 0.025 × 1.25 g/cm³ ≈ 0.03 g/cm³.

Conclusion: ρ = 1.25 ± 0.03 g/cm³, i.e. a 2.5% uncertainty.

Why the length error dominates

Notice the punchline: ΔL/L (0.5%) was actually smaller than Δm/m (1%). Yet because the side enters as L³, the side's relative error gets a factor of 3, growing it to 1.5% — now larger than the mass contribution.

This is the heart of the power-multiplication rule: whenever a quantity appears raised to a high power, its relative error matters disproportionately. In a sphere, V = (4/3)πr³, so radius error is again tripled. In kinetic energy K = (1/2)mv², the velocity error is doubled. Question-setters love this asymmetry because it punishes students who add relative errors blindly without weighting by the power.

Why it matters: In every experiment in your lab manual — from simple pendulum's g to Young's modulus to refractive index — the final reported error is dominated by the quantity that is either raised to the highest power or measured with the coarsest instrument. Identifying that dominant term tells you which instrument to upgrade first if you want a better result.

Real-world example: A jeweller verifying that a small gold cube is solid (not plated) measures its mass on a digital balance and its side with vernier calipers. Pure gold density is about 19.3 g/cm³. To distinguish, say, 19.3 from 18.5 confidently, the jeweller needs the combined relative error well under 2%. A 0.1 mm error on a 5 mm side is already 6% in density — so for small samples, the calipers, not the balance, set the trust limit.

Common misconception: Many students think the absolute errors add: Δρ = Δm + 3ΔL, or that the side's error contributes the same way as mass's. Both are wrong. It is the relative (percentage) errors that add, and only when the formula is a product of powers. For sums and differences, the absolute errors add instead — never mix the two rules.

:::compare

Operation	Error rule
q = A + B or q = A − B	Δq = ΔA + ΔB (absolute errors add)
q = A · B or q = A / B	Δq/q = ΔA/A + ΔB/B (relative errors add)
q = A^n	Δq/q = n · (ΔA/A)
q = A^p · B^q · C^r	Δq/q = p·(ΔA/A) + q·(ΔB/B) + r·(ΔC/C)
:::

Reading the result correctly

Once we know Δρ ≈ 0.03 g/cm³, the answer must be reported with the right number of significant figures. The rule of thumb: round the absolute error to one significant figure, and round the value to the same decimal place. Here Δρ has one sig fig in the hundredths place, so ρ is reported to the hundredths place: 1.25 ± 0.03 g/cm³. Reporting "1.2500 ± 0.0312" is technically incorrect — the trailing digits pretend to a precision the data does not support.

If the question changes the numbers slightly (say L = 2.0 ± 0.1 cm), reapply the formula — do not memorise just the 2.5% answer. The whole point is the method.

:::keypoints

• For ρ = m / L³, relative errors add: Δρ/ρ = Δm/m + 3 · (ΔL/L).
• The exponent in the formula multiplies that variable's relative error.
• Δm/m = 1%, 3·ΔL/L = 1.5%, total = 2.5%.
• ρ = 1.25 g/cm³, Δρ ≈ 0.03 g/cm³.
• The smaller-looking ΔL/L can dominate because of the power 3.
• Absolute errors add for sums/differences; relative errors add for products/quotients.
• Report value and error to the same decimal place; round error to 1 sig fig.
  :::

:::memory
"Power pumps the percent." Whenever a quantity is raised to a power n, multiply its percentage error by n before adding. Volume (n = 3) is the classic culprit.
:::

:::recap

• Density formula: ρ = m / L³.
• Error formula: Δρ/ρ = Δm/m + 3(ΔL/L).
• Numerical answer: ρ = 1.25 ± 0.03 g/cm³, 2.5% error.
• The cube of length amplifies the side's tiny error into the dominant term.
  :::