Node-Voltage & Mesh-Current Analysis
Series/parallel reduction and the dividers are pattern-matching: they work when the circuit happens to have the pattern. The moment a bridge, a second source, or a dependent source appears, the patterns run out. Node-voltage and mesh-current analysis are the two methods that never run out — they turn any resistive circuit into a linear system you crank through mechanically. Everything in the rest of this course (Thévenin, superposition, phasor AC analysis) is built on top of one of these two. Budget about 90 minutes here; it is the single densest scoring block on the exam.
Node voltage picks unknowns that automatically satisfy KVL, then imposes KCL. Mesh current picks unknowns that automatically satisfy KCL, then imposes KVL. Each method gets one law for free and pays for the other. That symmetry is the whole chapter.
1 · Node voltage — where it comes from
Start from the thing we actually want: every element's current and voltage. That's a lot of unknowns. The trick is to notice that voltage is a property of a node, not of a wire. Every point connected by perfect wire sits at the same potential. So if a circuit has nodes, there are only distinct potentials in the whole thing — and since only differences matter, we can nail one of them to zero (the reference node, or ground) and be left with unknowns.
Why does this get KVL for free? KVL says the voltage drops around any closed loop sum to zero. But if we define every element's voltage as a difference of node potentials, , then going around a loop the potentials telescope: identically. KVL can't be violated by construction. That leaves KCL as the only equation we still have to write.
ME 123 writes KCL as all currents leaving the node sum to zero. Every resistor term is therefore with this node first, and a current source pointing into the node enters the equation with a minus sign. Pick this convention once and never mix it — sign errors here are the single most common way to lose an otherwise perfect exam answer.
The procedure
Choose the reference node. Pick the node with the most connections, or the negative terminal of the main source. Mark it with the ground symbol; it is V by definition.
Label the remaining nodes — one unknown each.
Harvest the free constraints. Any node tied to ground through a voltage source is known: . Do not write KCL there — you'd be introducing the source's unknown current for nothing.
Write KCL at every remaining node, all currents leaving, each resistor term as .
Solve, then back out whatever the question actually asked for (a current, a power) from the node voltages.
Find and in the circuit above.
The source sits between and ground, so V for free — no KCL at that node. Two unknowns remain.
KCL at (all currents leaving: through the 4 Ω back toward , down the 12 Ω to ground, right through the 6 Ω toward ):
KCL at (left through the 6 Ω, down the 3 Ω, and the 2 A source pushes current into the node so it carries a minus sign):
Substituting into the second equation gives .
Cross-check by mesh (three clockwise meshes, the 2 A branch forcing A): A, A. Then V and V. ✓ Both methods agree.
Node connects to node through a 6 Ω resistor, to ground through a 3 Ω resistor, and a 2 A current source pushes current into from ground. Using the course's "all currents leaving" convention, which is the correct KCL equation at ?
2 · Supernodes — when a voltage source gets in the way
Node analysis needs to express every branch current in terms of node voltages. A resistor obliges: . A voltage source does not — it fixes its voltage and lets its current be whatever the rest of the circuit demands. If that source sits between ground and a node, no problem: the node voltage is simply known. But if it floats between two non-reference nodes, we can't write KCL at either one without inventing an unknown source current.
The fix comes straight from KCL's real statement. KCL isn't about nodes specifically — it's about any closed surface: charge doesn't pile up anywhere, so the total current crossing any boundary is zero. So draw a boundary around both nodes and the source between them. The source's mystery current is now entirely inside the boundary and never crosses it — it cancels out before we ever have to name it. That enclosure is a supernode.
One enclosure gives one KCL equation but covers two unknowns, so we're one equation short. The missing one is free: the source itself is a KVL statement about those two nodes.
Find , , above.
KCL over the enclosure. Four things cross the dashed boundary: the 4 A source (entering, so ), the 5 Ω to ground, the 20 Ω to ground, and the 4 Ω heading off to :
Notice what is not in that equation: any current through the 10 V source. It never crossed the boundary.
KVL constraint (the + terminal is at ):
KCL at — an ordinary node, nothing special:
Substituting both into the enclosure equation and multiplying by 20: , i.e. .
Cross-check by mesh (three clockwise meshes; the 4 A branch forces A): solving gives A and A. Then V, V, and V. ✓ Both methods agree.
3 · Step through it yourself
The supernode below is a smaller cousin of the one above — the same structure the Spring 2025 midterm leaned on. Click through the steps and watch the equations accumulate; the last step re-solves the identical circuit by mesh analysis so you can see the two methods land on the same numbers.
A 3 A source pushes current into node from ground. A 4 Ω resistor runs from to ground. A 6 V source sits between and with its + terminal at . A 4 Ω resistor runs from to ground. What is ?
4 · Dependent sources
A dependent source's value is set by some other voltage or current in the circuit — the control variable. It is not a new unknown, and it is not a known number either. The rule is mechanical: write the node equations exactly as usual, treating the dependent source's value as a symbol, then add one extra equation expressing the control variable in terms of node voltages and substitute it away. The system stays square.
Never "turn off" a dependent source, and never leave the control variable in the final system. A dependent source is not an independent source — it does not get zeroed in superposition, and it does not get deactivated when finding . It just sits there obeying its controller.
Find , and above.
Control-variable equation — is the current in the 5 Ω branch, defined flowing from toward :
KCL at (6 A entering ⇒ ; out through the vertical 5 Ω and the horizontal 5 Ω):
KCL at (out through the 5 Ω back toward and down the 10 Ω; the dependent source drives into the node, so ):
Substitute the control equation, , and multiply by 10:
Then , so .
Cross-check by element bookkeeping (no node equations at all): everything arriving at leaves through the 10 Ω, and arrives twice — once from the 5 Ω, once from the dependent source — so . Walking back through the horizontal 5 Ω, . KCL at : , so A and V, V. ✓ Agrees.
5 · Mesh current — the mirror image
Now flip the whole idea around. Instead of assigning a potential to each node, assign a circulating current to each mesh — a mesh being a loop with nothing inside it, like a single windowpane. The physical current in any branch is then the sum of the mesh currents flowing through it. ME 123 draws every mesh current clockwise, always; the consistency is what makes the sign bookkeeping automatic.
Why does this get KCL for free? Because a mesh current is, by definition, a loop — it enters every node on its path exactly as much as it leaves. Sum any number of loops at a node and you still get zero net accumulation. KCL is satisfied identically, which leaves KVL as the only equation left to write. Exactly the mirror of node analysis.
The procedure
Draw a clockwise current in every windowpane and label them
Harvest the free constraints. A current source touched by only one mesh fixes that mesh current outright — write (sign from whether the source's arrow agrees with the clockwise direction) and skip that mesh's KVL.
Walk each remaining mesh clockwise, summing voltage drops to zero. Through a resistor in the direction of your walk: . Through a source from to : a drop of ; from to : .
Solve, then recombine. A branch's real current is the mesh current, or the difference of two of them.
Find and above.
Mesh 1, walking clockwise from the bottom-left: up through the 24 V source ( to , a rise, so ), right through the 3 Ω with , then down the shared 6 Ω with :
Mesh 2, clockwise: up through the shared 6 Ω — now it's my current first, — right through the 6 Ω, then down through the 12 V source from to , a drop of :
Adding the two equations: .
Cross-check by node analysis. With the bottom rail as reference, the top-left node is V, the top-right is V, and the middle node is the only unknown: V. Then the 3 Ω carries A , the middle 6 Ω carries A, and the right 6 Ω carries A . ✓ Both methods agree.
Meshes 1 and 2 share a resistor , and both mesh currents are drawn clockwise as the course requires. You are traversing mesh 1 clockwise and summing voltage drops. Which term does contribute?
6 · Supermesh — the mirror of the supernode
Mesh analysis has the exact dual of the supernode problem. A current source refuses to tell you its voltage. If it's on the outer edge of one mesh, fine — it fixes that mesh current and you skip the KVL. But if it sits on a branch shared by two meshes, you can't write either mesh's KVL without inventing the source's unknown voltage.
Same fix, dualized: walk a loop that goes around the outside of both meshes, avoiding the offending branch entirely. The source's unknown voltage is never traversed, so it never enters the equation. That loop is a supermesh, and it gives one equation for two unknowns — with the source itself supplying the missing constraint, since it dictates the difference of the two mesh currents.
Find and above.
The constraint. Clockwise runs down the shared branch; clockwise runs up it. So the upward current there is , and the source's arrow points up at 2 A:
The supermesh KVL, walking the dashed loop clockwise from the bottom-left: up through the 32 V source (a rise, ), right through the 4 Ω carrying , jump the gap, right through the 2 Ω carrying , down the 4 Ω carrying :
Substituting : , so .
Cross-check by node analysis. Call the top-middle node and the top-right node ; the top-left is V. KCL at gives , and KCL at (2 A entering) gives:
So V and V. Then the 4 Ω on the left carries A and the 2 Ω carries A . ✓ Both methods agree.
7 · Which method should you actually use?
On an exam the question usually names the method. When it doesn't, count before you write:
| Node |
|
| Mesh |
|
| Either |
|
Node ↔ mesh · KCL ↔ KVL · node voltage ↔ mesh current · voltage source is the awkward one ↔ current source is the awkward one · supernode ↔ supermesh · "all currents leaving" ↔ "all drops clockwise". Learn one method properly and the other is the same method with the words swapped.
8 · Practice
Every answer below was solved twice, by node analysis and by mesh analysis, and shipped only because both agreed. Do the same: solve it one way, then check it the other.
Find and , and the total power delivered by the 28 V source.
Node analysis. The source pins the left node at 28 V. Two unknowns.
So , giving V and V. The source current is the current in the 4 Ω: A, so W delivered.
Cross-check by mesh. Two clockwise meshes: and . The second gives ; substituting, , so A and A. Then V, V, and W. ✓
Third sanity check — power balance. W dissipated, exactly what the source delivers.
Find and , then the power delivered by the 18 V source. Finally, verify that total power delivered equals total power dissipated.
Node analysis. The 18 V source floats between two non-reference nodes, so enclose it. Crossing the boundary: the 6 A source (entering, ), the 12 Ω, the 6 Ω.
Substituting and multiplying by 12: , so and V, V.
Source power. Everything arriving at comes through the source, and it all leaves down the 6 Ω: that current is A, flowing from through the source to — i.e. entering the terminal and leaving the terminal. That's delivery: W delivered.
Cross-check by mesh. Two clockwise meshes; the 6 A branch is touched only by mesh 1, so A. Mesh 2, walking clockwise: up the 12 Ω (a rise of ), right through the source to (a rise of 18), down the 6 Ω (a drop of ): A. Then V, V, and the source current is A. ✓
Power balance. Delivered: W by the current source, plus 90 W by the voltage source W. Dissipated: W. ✓
The dependent voltage source on the right is worth volts, where is the current in the 5 Ω resistor (rightward, as drawn). Find , and , and say whether the dependent source is absorbing or delivering power.
Control equation first. The 5 Ω sits only in mesh 1, and points the same way clockwise runs along the top, so .
Mesh 1, clockwise: up through the 30 V source (rise, ), right through the 5 Ω, down the shared 20 Ω:
Mesh 2, clockwise: up the shared 20 Ω as , right through the 10 Ω, then down through the dependent source from to — a drop of :
Substituting: , so A and A.
Cross-check by node analysis. Bottom rail is reference, top-left is 30 V; call the middle node and the right node . Now and the dependent source forces . KCL at :
So V, V, A ✓, the 10 Ω carries A ✓, and the shared 20 Ω carries A ✓.
Power. The source is worth V and carries A flowing downward, i.e. into its terminal — passive sign convention says it is absorbing W.
Both mesh currents are clockwise, as always. The 3 A source points downward. Find and , and the voltage across the current source (top terminal positive).
Constraint. Clockwise runs down the shared branch and clockwise runs up it, so the downward current there is . The arrow points down at 3 A:
Supermesh KVL, walking the outer loop clockwise and skipping the source branch: up through the 30 V source (), right through the 6 Ω with , across the gap, right through the 4 Ω with , down the 2 Ω with :
With : A, A.
Cross-check by node analysis. Top-left is 30 V; call the top-middle node and the top-right . KCL at gives , i.e. . Now KCL at — the 3 A source leaves this node going down, so it enters with a :
Substituting : , so V and V. Then the 6 Ω carries A ✓ and the 4 Ω carries A ✓.
The source voltage is just the node voltage at its top terminal (its bottom is grounded): V. That was the quantity mesh analysis refused to give us directly — which is exactly why the supermesh existed.
All three mesh currents are clockwise. Find , , and . (They will not come out to round numbers — that's fine; carry fractions and check them exactly.)
Mesh 1, clockwise: rise through the 30 V source, drop across the 2 Ω, drop across the shared 4 Ω:
Mesh 2, clockwise: the shared 4 Ω on the left as , the 2 Ω exclusive to mesh 2, the shared 4 Ω on the right as :
Mesh 3, clockwise: the shared 4 Ω as , the 2 Ω exclusive to mesh 3, then the 10 V source — its faces mesh 2, so clockwise (top to bottom through it) is a drop:
Solve. From the first equation, . Substituting into the second and combining with the third eliminates and down to a single equation in : , so . Then and .
Cross-check by node analysis. The 30 V source pins the top-left corner at 30 V, and the 10 V source pins the top-right corner at 10 V — both directly, no resistor in the way of either source's own terminal — so only the two internal rung-top nodes and are unknown:
Solving gives V, V. Then A, A, and A — all three match the mesh solution exactly. ✓
The dependent source is worth volts, oriented on top / on bottom as drawn. Both mesh currents are clockwise. Find , , and the voltage delivered by the dependent source.
Mesh 1, clockwise: rise through the 60 V source, drop across the 10 Ω, drop across the shared 20 Ω:
Mesh 2, clockwise: the shared 20 Ω as , then the dependent source — clockwise crosses it to , a rise, contributing :
Substituting into the first equation: A, so A.
Cross-check by node analysis. The 60 V source pins the top-left corner at 60 V. Because the dependent source has no series resistor, it pins the middle node directly: , and .
Then A ✓. The rung current is A downward, which is , giving A ✓ — both match the mesh solution.
Power. The source is worth V and V confirms the polarity drawn ( on top). It carries A flowing from its (bottom) terminal out into the external circuit — that's delivery: W.
All three mesh currents are clockwise; mesh 2 has no exclusive element of its own. The 4 A source points downward. Find , , , and the voltage across the current source (top terminal positive).
Constraint. Clockwise runs down the shared branch and clockwise runs up it, so the downward current there is . The arrow points down at 4 A:
Mesh 1, clockwise (unaffected by the supermesh): rise through the 50 V source, drop across the 5 Ω, drop across the shared 10 Ω:
Supermesh (meshes 2 and 3 combined), clockwise, skipping the current-source branch: the shared 10 Ω as (mesh 2's own contribution), then straight through to the 8 Ω as (mesh 3's own current — mesh 2 has nothing else to contribute):
Combine all three equations: from the constraint, . Substituting into the supermesh equation: . With from mesh 1, this reduces to , so A, A, and A.
Cross-check by node analysis. The 50 V source pins the top-left corner at 50 V. Mesh 2's top is a plain wire, so the rung-top node and the current-source-top node are the same node ; the right edge is grounded directly, so the far-right corner is 0 V. One unknown, :
Then A ✓, the rung current A A ✓, and A ✓ — all three match, and ✓.
The source voltage is V, top positive. Its 4 A flows out of that terminal into the rest of the circuit, so it is delivering W.
If you can do four things without hesitating, this chapter is done: (1) write a KCL equation with every term as this node minus that node; (2) recognize a floating voltage source and draw the enclosure without being told; (3) walk a mesh clockwise writing ; (4) spot a shared current source and reach for the outer loop. Everything else in this chapter is arithmetic.