Ferrite bead

A bead of ferrite material with a conductor running through it. On a schematic it looks like an inductor, and at low frequency it behaves like one. At DC it is a few tens of milliohms, near enough to a piece of wire that the rail through it does not care. Somewhere in the tens of megahertz it becomes a resistance of tens or hundreds of ohms, and that resistance is lossy on purpose. Energy arriving there is dissipated as heat instead of being reflected back down the line, which is what a plain inductor would do with it.

Impedance of a ferrite bead against frequency on a logarithmic axis: near zero ohms at 1 MHz, rising through about 100 ohms at 100 MHz, broadly flat to 1 GHz impedance frequency 0 Ω 50 Ω 100 Ω 150 Ω 1 MHz 10 MHz 100 MHz 1 GHz a near short at DC the noise turns into heat about 100 Ω at 100 MHz

The usual job is to stop one part of a board polluting another. A switching regulator makes a rail carrying tens of millivolts of ripple at its switching frequency and every harmonic above it, which is fine for logic and not fine for an ADC reference. A bead in series between the two, with decoupling on each side of it, gives that noise somewhere to go while the quiet side keeps the DC it needs.

Two things about a bead catch people out.

It is not protection. A bead passes DC, and it passes a fault current just as happily, so one sitting between a connector and a regulator input protects nothing. That case is what input-protection exists to flag. The rule walks out from each connector net through series pass elements looking for a fuse or a clamp, and a path whose only inline part is a bead leaves the finding standing.

It splits one logical rail into two nets. The noisy side and the quiet side have different names after the bead, so a rule reasoning about one net at a time cannot see across it. The engine answers that with a reachability walk that crosses two-terminal series parts (resistor, inductor, ferrite, fuse) rather than stopping at each one. That walk is how profile-termination still finds a split terminator that no single net touches directly. The same boundary is why floating-input goes quiet on any net carrying a bead. Claiming that an input floats when a passive sits beside it is guesswork rather than a finding.

Where the course teaches it: chapter 1 puts a ferrite in the table of what a two-terminal part is doing, and the recurring jobs files it under taming a fast edge.