What Is a Circuit?

Intuition

Why it has to be a loop

Think about the chain on a bicycle. It is a closed ring: the pedals push one part of it, and because the chain has nowhere to go except round, the push arrives at the back wheel. Break a single link and the whole chain goes slack. The pedals spin freely, the wheel does nothing, and the links furthest from the break are as useless as the break itself.

Electricity behaves the same way, and for the same reason. Charge has to have somewhere to go and somewhere to come back from. A battery does not squirt electricity out of its positive terminal into the world; it pushes charge round a ring that starts and ends at its own two terminals. Complete that ring and current flows in it. Interrupt it at any point at all — a cut wire, a switch turned off, a burnt-out filament, a battery lifted out of its holder — and the current everywhere in the ring falls to zero.

Not every ring is a useful one. The push has to come from somewhere: a battery, a socket, a solar panel. The charge needs a conductor to travel along, which might be wire, track on a board, or the metal inside a switch. And somewhere in the ring there has to be a lamp, a motor, a buzzer or a chip, something that turns the energy into an outcome you wanted. Most real circuits add a fourth part as well, a switch, so that the loop can be opened and closed on purpose.

The word circuit is worth taking literally. It shares a root with "circuitous" and "circular", and the thing it names is the trip round, not the parts that make the trip possible.

Practitioner

The four parts, and what sets the current

A working circuit is a source, a conducting path, a load, and usually some control. The source maintains a voltage, the path carries the current, and the load converts electrical energy into heat, light, motion or computation. The control — a switch, a relay, a transistor — decides when the loop is complete.

The current is a joint result rather than a property of any one part. It appears when a voltage is applied across a resistance, so the source and the load settle it between them:

Rearranged for current, that says current equals voltage divided by resistance. Halve the load resistance and the source has to supply twice the current. That is why "what current does this draw?" is a question about the load at least as much as about the supply.

Adding a second component means choosing where to put it. Wire it into the same loop and both parts carry the same current, which makes a series circuit. Give it its own loop back to the source and each part has its own current while both see the same voltage, which makes a parallel circuit. Almost every real circuit mixes the two arrangements.

When a circuit fails, the loop has either broken where it should be complete or closed where it should be open. Those are the open and short circuits behind most of the faults you will ever chase.

On paper the loop is drawn as a schematic, which records what connects to what and deliberately says nothing about where the parts physically sit. Build the same loop on a breadboard or a printed board and the physical arrangement matters again.

Engineer

The loop as a graph, and the wire as an approximation

Formally a circuit is a graph: components are edges, and the points where they join are nodes. Any closed walk through that graph is a loop. Conservation applied twice — once to charge, once to energy — is enough to solve any circuit made of lumped components. Charge cannot accumulate at a node, so the currents entering a node equal the currents leaving it — Kirchhoff's current law. Energy per unit charge is path-independent, so the voltage changes round any closed loop sum to zero — Kirchhoff's voltage law.

Between them, those two laws are why a single-loop circuit has exactly one current, and why the voltages across its parts must add up to the source voltage. They also expose an assumption that Layer 1 left unstated: the connecting wire was treated as a component with no properties at all. Real wire has resistance, and once it goes back into the drawing it is simply another element in the loop.

The lumped-element model underlying all of this says that a component's behaviour can be collapsed to a point, and that the wire between components is instantaneous and property-free. That holds while the circuit is small compared with the wavelength of the signals running through it. At mains frequency a building is small; at a few gigahertz a centimetre of track is not, and the loop has to be treated as a transmission line instead. The changeover is abrupt: the model applies while the geometry of the layout is irrelevant and stops applying once that geometry matters, so recognising which regime a design sits in is part of the engineering.

Because the graph is the circuit, a circuit is defined by its connections and not by its drawing. Two schematics that look nothing alike are the same circuit if the same terminals meet at the same nodes, which is why redrawing a confusing schematic is a legitimate and often decisive troubleshooting technique.

Professional

The loop you drew and the loop that exists

Every current has a return path, and that return path is a real conductor. A schematic shows a symbol for ground and stops there, while the board has to carry the current back through copper of finite width and finite resistance. The area enclosed between the outgoing and the returning conductor is itself a loop: it radiates when the current changes, and it picks up interference when something else radiates. That is where EMI and EMC begin. Boards laid out by someone thinking about the loop rather than about the wire come out quieter.

Contacts are components you did not draw. Connectors, switch contacts, sockets and solder joints all sit inside the loop and all have resistance.

A contact resistance of 20 mΩ carrying 3 A loses 60 mV and dissipates 0.18 W inside the contact itself. That stays negligible until the contact oxidises or loosens. The current through it is set by the rest of the circuit, so when the contact resistance rises the power lost in the contact rises with it, in proportion. The extra heat speeds the oxidation along, the resistance rises again, and the process feeds back on itself, which is why this kind of failure creeps rather than arrives. Intermittent faults in old equipment are very often this loop quietly running away with itself, and connectors and switches carry current ratings for exactly this reason.

Loops nobody intended are still circuits. Leakage across a dirty board, a chassis that bonds two supposedly separate returns, a probe ground clip that closes a path through the mains earth: each completes a loop that appears on no schematic, and each obeys the same laws as the loop that was designed. Ground loops are the classic case.

Where you choose to interrupt the loop is a design decision in its own right. A switch in the high side leaves the load sitting at ground potential when it is off; a switch in the return leaves the load floating at supply potential instead. Low-voltage work can be made to work either way, but on mains-connected equipment the safety standards decide, and the disconnect goes in the live conductor. Fuses are the deliberate weak link in the loop, sized to open before the wiring does.

Fault-finding takes the same shape. A dead circuit is a loop with a break somewhere in it, and the systematic method is to walk round the loop with a meter, halving the remaining suspects at each measurement instead of probing wherever seems promising. DC troubleshooting sets that method out.