ARCHETYPE: Equivalent-Resistance Puzzles

There is a moment in every serious circuits course when the comfortable rules stop working. You know that series resistances add and parallel conductances add — Resistors in Series & Parallel drilled that. Then an examiner hands you twelve equal resistors soldered into the edges of a cube and asks for the resistance between opposite corners. Try to find two resistors in series: there aren't any. In parallel: none either. Every reduction move you own has nothing to grab. Yet the answer — $5R/6$ — takes about three lines, if you know the move the puzzle was built to reward. That is what this article trains: the small stable of "impossible-looking" networks examiners have recycled for a century, and the four moves that dissolve them. Symmetry folding merges points at the same potential and turns the cube into a trivial chain. Balance lets you erase — or short — the middle branch of a Wheatstone-shaped network. Self-similarity cracks the infinite ladder with one quadratic. And for the rare network resisting all three, a triangle-to-star exchange rewires the diagram into something reducible. None of these puzzles requires new physics — every move is licensed by the two laws you already trust, applied with a geometer's eye. Learn the four moves and the classic answers, and an entire genre of JEE and board questions turns from ambush into routine.

Key takeaways

Equivalent-resistance puzzles are networks where nothing is in series or parallel — cube, bridge, ladder — built to reward a move, not brute force.
Move 1 — symmetry folding: nodes that a symmetry exchanges sit at the same potential and can be merged; no potential difference, no current, so the merge changes nothing.
The cube folds into three parallel groups in series: $R/3 + R/6 + R/3 = 5R/6$ between opposite corners.
Move 2 — balanced bridge: the middle branch carries no current, so erase it or short it — both give the same answer.
Move 3 — self-similarity: an infinite ladder equals itself minus one cell; the consistency quadratic's positive root is $R_{\text{eq}} = \frac{1+\sqrt{5}}{2}R \approx 1.618R$ .
Move 4 (rare) — wye–delta: exchange a blocking triangle for an equivalent star and the network becomes reducible.
Key idea: when series-parallel fails, don't compute harder — look for equal potentials, balance, or self-similarity, and let the structure of the network do the work.

The core insight

Nodes held at the same potential by symmetry can be merged without changing any current, a balanced bridge branch can be erased or shorted at will, and a self-similar ladder must equal its own tail — three structural facts that reduce every classic "impossible" resistor network to ordinary series-parallel arithmetic.

Where this lives today

The infinite ladder is not a curiosity: inside nearly every digital-to-analog converter sits an R–2R ladder, a finite cousin whose self-similar structure splits currents in exact binary ratios using only two resistor values — precision by topology, immune to manufacturing tolerance because only ratios enter. Symmetry folding, meanwhile, is how engineers sanity-check chip and grid simulations: a mirror-symmetric layout must report matched node voltages, and any asymmetry flags a netlist bug before fabrication. And the genre still recruits — the internet's most famous physics nerd-snipe, the xkcd infinite-grid-of-resistors problem, is Move 1 and Move 3 escalated to graduate level. The cube, ladder and bridge are training wheels for a style of argument professionals never stop using.

The setup

The archetype here is not one circuit but one situation: a network in which no pair of resistors shares a single current path (series) or a common pair of nodes (parallel). The flagship is the resistor cube — twelve equal resistors $R$ on a cube's edges, resistance asked between diagonally opposite corners. Each corner joins three resistors, every resistor ends at a three-way junction, and the reduction vocabulary of Resistors in Series & Parallel cannot even begin.

Figure

The resistor cube — twelve equal edges, fed at opposite corners, folds into three layers by symmetry

Why do such networks exist on exam papers? Because they test whether you understand why the reduction rules work. The series rule is a same-current fact; the parallel rule a same-voltage fact — Resistors in Series & Parallel establishes both, and Kirchhoff's Rules gives you the two laws behind them. Each move below makes those hidden facts visible. One habit before any move: redraw the network with named nodes. Trace along bare wire through no components and mark everything you reach as one node; repeat until every junction has a letter. Half of these puzzles die at the redrawing stage — a circuit drawn as a trapezium is often a plain parallel pair in disguise. (The technique module Redraw the Circuit, coming up in this chapter, trains exactly this.)

The master solution pattern

Move 1 — symmetry folding. Suppose some rotation or reflection maps the network onto itself and keeps the two feed terminals fixed (or swaps them). Nodes that the symmetry exchanges must sit at the same potential — the network cannot tell them apart. Now the license:

V_P = V_Q \;\;\Rightarrow\;\; I_{PQ} = \frac{V_P - V_Q}{R_{PQ}} = 0

where: $V_P, V_Q$ — the potentials of two symmetry-related nodes; $I_{PQ}$ — the current in any resistor joining them. Zero potential difference drives zero current whatever the resistance, so you may delete a resistor connecting equal-potential nodes, or solder the nodes into one — either way, no current anywhere changes. Merging is the productive choice: it manufactures parallel groups out of thin air.

Move 2 — the balanced bridge, erase or short. Five resistors in a diamond — arms $P, Q$ on top, $S, T$ below, a bridge resistor across the middle — is the Wheatstone geometry. The Wheatstone Bridge establishes the balance condition; this gym spends it. At balance the two middle nodes share a potential, the bridge resistor carries nothing, and Move 1's license applies both ways:

\frac{P}{Q} = \frac{S}{T} \;\;\Rightarrow\;\; R_{AB} = (P+Q)\,\|\,(S+T) \;=\; (P\,\|\,S) + (Q\,\|\,T)

where: the first form comes from erasing the dead bridge (two series chains in parallel), the second from shorting it (two parallel pairs in series). Both are exact, and computing the answer both ways is the fastest self-check in this gym. The bridge resistor's value never enters — a favourite distractor.

Move 3 — self-similarity for infinite ladders. Chop the first cell off an infinite ladder and the remaining tail is still the same infinite ladder — so if the whole presents $R_{\text{eq}}$ , the tail does too. Writing "first cell + tail" in terms of $R_{\text{eq}}$ gives one self-consistency equation, and the infinity is gone. For the classic ladder whose every cell adds one series $R$ and one parallel $R$ :

R_{\text{eq}} = R + \frac{R\,R_{\text{eq}}}{R + R_{\text{eq}}}

where: the first $R$ is the series element of the cut-off cell, and the fraction is that cell's parallel $R$ combined with the tail (which presents $R_{\text{eq}}$ again). Multiplying out gives a quadratic whose positive root is the golden ratio times $R$ . The same move powers the capacitor ladder in ARCHETYPE: Capacitor Networks & Equivalent Capacitance — the identical quadratic in a different currency.

Move 4 — wye–delta, the lite version. Occasionally a triangle of resistors blocks every reduction and no symmetry helps (the unbalanced bridge is the standard case). The exchange: a triangle (delta) of $R_{ab}, R_{bc}, R_{ca}$ between nodes $a, b, c$ behaves identically to a star (wye) of three resistors to a new centre node, with

R_a = \frac{R_{ab}\,R_{ca}}{R_{ab} + R_{bc} + R_{ca}}

where: $R_a$ — the star resistor attached to node $a$ — is the product of the two delta resistors touching $a$ , over the sum of all three; $R_b$ and $R_c$ follow by rotating the labels. Sanity anchor: a delta of three $3\ \Omega$ resistors becomes a star of three $1\ \Omega$ . Shown once, drilled no further — JEE rewards recognising when it is needed, and if all else fails, the branch-current method of ARCHETYPE: Multi-Loop & Two-Battery Circuits always works.

Derivation AThe resistor cube folded by symmetry

We want the resistance of the twelve-edge cube between opposite corners $A$ and $G$ — the flagship puzzle, in which no two resistors are in series or in parallel anywhere. The strategy is Move 1: find the symmetry, identify the equal-potential nodes, merge them, and watch the cube collapse into a series chain of parallel groups. The thing to watch is the counting — after the merge, every one of the twelve edges must be accounted for, no more and no fewer.

Step 1. Feed current $I$ in at corner $A$ and out at the opposite corner $G$ . Rotating the cube by a third of a turn about the long diagonal $AG$ maps the cube onto itself and keeps both terminals fixed. The three corners adjacent to $A$ cycle into one another, so all three sit at one common potential — layer 1. The three corners adjacent to $G$ likewise: layer 2. Why does "the network cannot tell them apart" force equal potentials? A node's potential is determined by the network and the feed points; the rotation reproduces both identically, so each node's potential must equal that of the node it lands on.

Step 2. Merge each layer into a single node (no current would flow in wires joining equal-potential points, so adding them changes nothing — and once added, the nodes are one). Count edges: $A$ to layer 1 — three in parallel, $R/3$ . Layer 1 to layer 2 — the remaining six edges all run between these layers, $R/6$ . Layer 2 to $G$ — three in parallel, $R/3$ . Total $3 + 6 + 3 = 12$ . All present.

Step 3. The folded network is a plain series chain:

R_{AG} = \frac{R}{3} + \frac{R}{6} + \frac{R}{3} = \frac{2R + R + 2R}{6} = \frac{5R}{6}

Step 4. The same folding with mirror symmetries cracks the cube's other two questions: a face diagonal gives $3R/4$ , a single edge $7R/12$ . Deriving those is the best practice exercise on offer — each is a rerun of Steps 1–3 with a different symmetry. Verified: solving the full eight-node network numerically gives 0.8333R, 0.7500R and 0.5833R for the three terminal pairs — exactly 5R/6, 3R/4 and 7R/12. The fold is exact, not an approximation.

Go deeper · Derivation B · 1 minThe infinite ladder and its self-similarity quadratic

Derivation BThe infinite ladder and its self-similarity quadratic

We want the equivalent resistance of an endless ladder in which every cell contributes one series resistor $R$ and one parallel resistor $R$ , repeating forever. Brute force is literally impossible — there are infinitely many elements — so the strategy is Move 3: the ladder minus its first cell is the same ladder, and that self-similarity is one equation in the one unknown $R_{\text{eq}}$ . Watch for the quadratic's two roots: only one can be a resistance.

Step 1. Let the whole ladder, seen from its input terminals, present resistance $R_{\text{eq}}$ . Remove the first cell (one series $R$ , one parallel $R$ ). What remains is an infinite ladder identical to the original, so the tail also presents $R_{\text{eq}}$ .

Step 2. Rebuild the ladder from the front: the input sees the series $R$ , then the parallel $R$ side by side with the tail's $R_{\text{eq}}$ . Using the product-over-sum rule from Resistors in Series & Parallel:

R_{\text{eq}} = R + \frac{R\,R_{\text{eq}}}{R + R_{\text{eq}}}

Step 3. Multiply through by $(R + R_{\text{eq}})$ and collect terms:

R_{\text{eq}}(R + R_{\text{eq}}) = R(R + R_{\text{eq}}) + R\,R_{\text{eq}} \;\;\Rightarrow\;\; R_{\text{eq}}^2 - R\,R_{\text{eq}} - R^2 = 0

Step 4. Solve the quadratic with the standard formula:

R_{\text{eq}} = \frac{R \pm \sqrt{R^2 + 4R^2}}{2} = \frac{1 \pm \sqrt{5}}{2}\,R

The minus root is negative — not a resistance — so it is discarded on physical grounds, leaving

R_{\text{eq}} = \frac{1 + \sqrt{5}}{2}\,R \approx 1.618\,R

Why this bound makes sense: the answer must exceed $R$ (current crosses the first series resistor regardless) and stay below $2R$ (shorting the entire tail would still leave $R + R$ , an easier network than the truth). The golden ratio sits obediently between 1 and 2. *Verified: iterating the cell formula numerically converges to 1.6180R from any start; a twelve-cell finite ladder already presents 1.61807R — within a hundredth of a percent of the infinite answer.*

The variants examiners actually pose

Figure

The balanced bridge collapses two ways — erase the dead branch or short it, and the answers must agree

network	the move	answer
cube, opposite corners	fold by the diagonal rotation	$5R/6$
cube, face diagonal	fold by the mirror through that diagonal	$3R/4$
cube, one edge	fold by the mirror through that edge	$7R/12$
square of four $R$ plus one diagonal, fed across the other diagonal	mirror symmetry: diagonal is dead — erase it	$R$
balanced bridge $P/Q = S/T$	erase or short the bridge branch	$(P{+}Q)\,\\|\,(S{+}T)$
infinite $R$ -per-cell ladder	self-similarity quadratic	$\frac{1+\sqrt5}{2}R \approx 1.618R$
unbalanced bridge	wye–delta one triangle, then reduce	case by case

The square-with-diagonal is the gentlest introduction to Move 1: four equal resistors on a square's sides, a fifth across one diagonal, feed across the other. The mirror through the feed diagonal swaps the two remaining corners, so the diagonal resistor is dead and the network is two 2-resistor chains in parallel: $2R \,\|\, 2R = R$ . The fifth resistor was scenery.

Worked examples

Board level — mixed network with full back-substitution (CBSE/ISC pattern, 3 marks)

A $12\ \mathrm{V}$ battery of negligible internal resistance drives a $6\ \Omega$ resistor in series with a parallel pair of $4\ \Omega$ and $12\ \Omega$ . Find the equivalent resistance, the battery current, and the current in the $4\ \Omega$ .

Step 1. Reduce the innermost group first. The parallel pair, by product over sum: $R_p = \dfrac{4 \times 12}{4 + 12} = 3\ \Omega$ — smaller than the smaller member ✓.

Step 2. The chain: $R_{\text{eq}} = 6 + 3 = \boxed{9\ \Omega}$ .

Step 3. Battery current by Ohm's Law & Resistance: $I = 12/9 = 4/3 \approx \boxed{1.33\ \mathrm{A}}$ .

Step 4. Back-substitute. The parallel group's shared voltage: $V_p = (4/3)(3) = 4.0\ \mathrm{V}$ , so $I_4 = 4.0/4 = \boxed{1.0\ \mathrm{A}}$ and $I_{12} = 0.33\ \mathrm{A}$ . Junction audit: $1.0 + 0.33 = 1.33\ \mathrm{A}$ ✓. The marks-scheme rhythm: reduce inward, walk back out, audit at a junction.

Board level — the balanced bridge, both routes (CBSE/ISC/NEET pattern)

Arms $AB = 2\ \Omega$ , $BC = 4\ \Omega$ , $AD = 3\ \Omega$ , $DC = 6\ \Omega$ , with a $5\ \Omega$ resistor between $B$ and $D$ . Find the equivalent resistance between $A$ and $C$ .

Step 1. Test balance first — always: $\dfrac{P}{Q} = \dfrac{2}{4} = \dfrac{1}{2} = \dfrac{3}{6} = \dfrac{S}{T}$ . Balanced, so $B$ and $D$ share a potential and the $5\ \Omega$ bridge carries zero current.

Step 2. Route 1, erase the bridge: chains $2 + 4 = 6\ \Omega$ and $3 + 6 = 9\ \Omega$ in parallel: $\dfrac{6 \times 9}{15} = \boxed{3.6\ \Omega}$ .

Step 3. Route 2, short the bridge: $2 \,\|\, 3 = 1.2\ \Omega$ plus $4 \,\|\, 6 = 2.4\ \Omega$ in series: $3.6\ \Omega$ ✓.

Step 4. State the reason for full marks: "at balance $V_B = V_D$ , so the $5\ \Omega$ resistor carries no current and may be removed." The $5\ \Omega$ never entered — if your answer contains it, the balance test was skipped.

Competitive level — the cube with numbers (JEE-Main pattern, ~3 minutes)

Twelve resistors of $6\ \Omega$ each form the edges of a cube. A $10\ \mathrm{V}$ ideal battery is connected across a main diagonal. Find the battery current and the current in each edge meeting the positive terminal.

Step 1. Between opposite corners the fold gives $R_{AG} = \dfrac{5R}{6} = 5\ \Omega$ . (In an exam, quote the fold — three in parallel, six in parallel, three in parallel — in one line.)

Step 2. Battery current: $I = 10/5 = \boxed{2.0\ \mathrm{A}}$ .

Step 3. By the three-fold symmetry, the current entering at $A$ splits equally into the three edges leaving it: each carries $\dfrac{2.0}{3} = \boxed{0.67\ \mathrm{A}}$ . (Each middle-layer edge carries $2.0/6 = 0.33\ \mathrm{A}$ — the standard follow-up variant.) Timing note: a formula-recall question in costume. After this article it should take 60 seconds — the only work is $5R/6$ and two divisions. If you find yourself writing loop equations for a cube, stop: that route costs fifteen minutes.

Competitive level — dead diagonal, then the infinite ladder (JEE-Advanced stretch, ~4 minutes)

(a) Four $2\ \Omega$ resistors form a square $PQRS$ ; a fifth $2\ \Omega$ joins $Q$ to $S$ . Find the resistance between $P$ and $R$ . (b) An infinite ladder is built from $2\ \Omega$ resistors, one series and one parallel per cell. Find its input resistance.

Step 1. Part (a): the mirror through $PR$ swaps $Q$ and $S$ while fixing $P$ and $R$ , so $V_Q = V_S$ and the diagonal is dead. Erase it: two chains of $4\ \Omega$ in parallel give $\boxed{2\ \Omega}$ .

Step 2. Part (b): self-similarity with $R = 2\ \Omega$ : $R_{\text{eq}} = 2 + \dfrac{2R_{\text{eq}}}{2 + R_{\text{eq}}}$ .

Step 3. Clear the fraction: $R_{\text{eq}}^2 - 2R_{\text{eq}} - 4 = 0$ , so $R_{\text{eq}} = 1 \pm \sqrt{5}$ .

Step 4. Discard the negative root (a resistance cannot be negative): $R_{\text{eq}} = 1 + \sqrt{5} \approx \boxed{3.24\ \Omega}$ . Bound check: between $R = 2$ and $2R = 4$ ✓ — and $3.24 = 1.618 \times 2$ , the golden ratio's cameo. Timing note: part (a) is 45 seconds once the mirror is spotted; part (b) is 2 minutes of algebra. The marks die in the discriminant — write it explicitly. If the paper offers $1 - \sqrt5$ among the options, it is fishing for the sign slip.

Exam traps

Erasing without earning it. The biggest mark-loser in this genre: deleting a diagonal or bridge branch because the picture looks symmetric. The license requires a named symmetry fixing or swapping the feed terminals, or a verified balance ratio. Feed the square across a side instead of the diagonal and the "dead" resistor comes alive.

Using the bridge resistor's value. In a balanced bridge the middle resistance never appears in $R_{\text{eq}}$ — options depending on it are distractors. And if the bridge is not balanced, no shortcut exists: check $P/Q$ against $S/T$ before anything else.

Product-over-sum with three resistors. $\dfrac{R_1 R_2 R_3}{R_1 + R_2 + R_3}$ is not the parallel formula for three — the reciprocal rule is. Product over sum is a two-resistor special case only.

Keeping the wrong root. The ladder quadratic always has one negative root; writing $R_{\text{eq}} = \frac{1 - \sqrt5}{2}R$ is an instant zero despite 95%-right algebra. State the discard reason — "resistance cannot be negative" — for the reasoning mark.

No bounds check. Every answer here has cheap bounds: a cube answer must lie between $R/3$ (the easiest start, three parallel edges) and $R$ ; a ladder answer between $R$ and $2R$ . Thirty seconds of bounding catches most slips.

Fraction discipline. Keep $5R/6$ , $4/3\ \mathrm{A}$ , $54/15$ as fractions until the last line. Rounding $4/3$ to $1.3$ before back-substitution makes the junction audit fail cosmetically and costs the check mark; CBSE, ISC and JEE all favour exact fractions.

Origins · 1 minJames Clerk Maxwell and how this was discovered

James Clerk Maxwell

1831–1879 · Edinburgh, Aberdeen, Cambridge

Maxwell is remembered for the field equations, but his century met him first as the great systematiser of circuits. His 1873 Treatise turned network analysis into a discipline — loop currents solved wholesale by determinants — and its exercises set the style for a century of examination questions on symmetric resistor networks; the resistor cube descends directly from that Treatise tradition, which Cambridge examiners (Maxwell among them, as a Tripos setter) made a rite of passage. Fittingly, the man who asked students to fold networks by symmetry spent his final years building the Cavendish Laboratory, where precision resistance measurement became an industrial art. He died at 48, before the cube he helped canonise became every student's favourite ambush.

History

1827

— Ohm publishes the law of the single conductor: every network puzzle will be built from V = IR

1845

— Kirchhoff, aged 21, states the junction and loop rules — networks become solvable in principle

1873

— Maxwell's Treatise systematises network analysis and seeds the exam tradition of symmetric-network puzzles

1899

— Kennelly publishes the delta-wye transformation, the escape hatch for unbalanced bridges

1923

— Zobel's wave-filter theory at Bell Labs makes repeating ladder networks an engineering staple — the infinite ladder gets a day job

Going further

Texts. NCERT Physics Class 12, Ch. 3 · HC Verma, Concepts of Physics Vol. 2, Ch. 32 (the cube and ladder live in the exercise sets) · Halliday, Resnick & Walker, Ch. 27.

Provenance. All results are classical. A verifier should recompute: the cube fold $5R/6$ (with $3R/4$ and $7R/12$ — all confirmed against a full eight-node numerical solution); the bridge both ways ( $6 \| 9 = 3.6\ \Omega$ ; $1.2 + 2.4 = 3.6\ \Omega$ ); the mixed-network audit ( $9\ \Omega$ , $4/3\ \mathrm{A}$ , $1.0\ \mathrm{A}$ ); the ladder roots $1 \pm \sqrt5$ for $R = 2\ \Omega$ and 12-cell convergence to $3.236\ \Omega$ ; delta 3,3,3 → wye 1,1,1.

Prerequisites. Resistors in Series & Parallel · The Wheatstone Bridge. Leads to. ARCHETYPE: Multi-Loop & Two-Battery Circuits (when no shortcut survives) · Redraw the Circuit (technique, coming up) · Series and Parallel Networks (UG deep dive) — the university treatment, with the Thévenin machinery this article holds back.

How do I spot the symmetry quickly in an exam?

Look at the two feed terminals first, not the resistors. Is there a mirror line or rotation axis through the feed points that maps the network onto itself? For the cube fed at opposite corners it is the rotation about that diagonal; for the square fed across a diagonal, the mirror through the feed. Nodes that swap under the operation are your equal-potential pairs. If no operation fixes (or swaps) the terminals, symmetry will not help — test for bridge balance next.

Why exactly can I merge two nodes at the same potential? It feels like cheating.

Current between two points needs a potential difference: $I = (V_P - V_Q)/R$ , and $V_P = V_Q$ makes that zero for any $R$ . A wire added between equal-potential nodes would carry nothing, and a wire carrying nothing cannot disturb any current anywhere. Once the wire exists, the two nodes are one node, and resistors that were not in parallel before suddenly are. Nothing was approximated — the symmetry revealed structure that was always true.

Erase or short the balanced bridge — which should I choose?

Whichever makes the arithmetic cleaner — they are equally legal and must agree. Erasing leaves two series chains in parallel: nice when the arm sums are round. Shorting leaves two parallel pairs in series: nice when the side pairs are round. The professional habit is to do both — agreement is a free, complete check of balance test and arithmetic in twenty seconds.

What if the bridge is not balanced?

Then no branch is dead and the five-resistor network genuinely has no series-parallel reduction. Two honest routes: swap one triangle for a star (Move 4) and reduce, or set up branch currents — the method drilled in ARCHETYPE: Multi-Loop & Two-Battery Circuits. An unbalanced bridge with kind numbers usually signals wye–delta; a bridge with a battery and internal resistance signals branch currents.

Why does the golden ratio appear in the ladder? Is that a coincidence?

It is the fingerprint of the particular cell — one series R, one parallel R — whose self-similarity condition is the same quadratic, $x^2 = x + 1$ , that defines the golden ratio. Change the cell and you get a different, non-golden root. What is not a coincidence is that some fixed number exists: any repeating ladder must present a resistance unchanged by deleting one cell, and that consistency always closes the infinity into one algebraic equation.

How do I know a puzzle needs a special move rather than plain reduction?

Run the two honest tests from Resistors in Series & Parallel: series needs a shared node with nothing else attached; parallel needs both ends on the same two nodes. If no pair passes either test, plain reduction is over and you are in this article's territory. The diagnosis takes fifteen seconds and prevents the classic disaster of "finding" a series pair that isn't one — redraw with named nodes first, always.