How this instrument works
Bandwidth here means the frequency span between the two points where a resonant circuit's output power falls to half its peak value — the −3 dB points, in the shorthand electronics settled on decades ago. BW = f₀ ⁄ Q simply runs the defining relationship backwards: Q itself is usually defined first, as Q = f₀ ⁄ BW, the center frequency divided by that half-power width. A radio tuning coil, a quartz crystal, a laser cavity, and a struck wine glass all obey the same ratio, because Q is a dimensionless measure of how much energy a resonant system stores compared with how much it loses each cycle, and bandwidth is simply what that ratio looks like laid out on a frequency axis.
The shape of the formula follows from a second-order resonant response. Near its peak, a single tuned circuit or a single-pole filter traces a curve that falls to half power almost symmetrically about f₀, and the width of that lobe scales inversely with Q: double the Q and the lobe becomes half as wide for the same center frequency. That inverse relationship is why bandwidth divides rather than multiplies — a quality factor of 10 leaves ten times less spectral room than a loosely coupled resonance, and a factor of a few thousand, typical of a good quartz crystal, can carve out a sliver only a few hundred hertz wide even at megahertz frequencies.
The formula assumes one clean, well-behaved resonance, and it stops being reliable once Q drops much below about 0.5, where the half-power points no longer sit symmetrically either side of f₀ and the peak itself shifts. It also describes a single isolated pole: a multi-stage filter built from several coupled resonators gets its overall bandwidth from the filter's design topology, not from dividing one component's Q into one center frequency. Reach for BW = f₀ ⁄ Q for a single tuned circuit or a single spectral line; use filter-design tables once several stages start interacting.
- Enter Center frequency in whichever unit fits the circuit — Hz for audio, kHz or MHz for an IF or RF tuning stage, GHz for a microwave cavity.
- Enter Q factor, the dimensionless ratio of stored to dissipated energy per cycle; a loosely coupled tank circuit might sit near 10, a good quartz crystal in the thousands.
- Read Bandwidth — the width of the resonance between its half-power points, shown in Hz, kHz, or MHz.
- If Q factor is left at zero or entered negative, the instrument flags it: a bandwidth cannot be negative or infinite, so Q must be a positive number.
Worked example — a 100 MHz circuit with Q = 10
A shortwave receiver's front-end tuning circuit is centered at f₀ = 100 MHz with a Q factor of 10 — a modest, loosely coupled resonance typical of a simple LC tank rather than a crystal. Bandwidth works out to BW = f₀ ⁄ Q = 100,000,000 ⁄ 10 = 10,000,000 Hz, a clean 10 MHz. Signals anywhere from roughly 95 MHz to 105 MHz pass through with at least half their power intact, while stations further out are progressively attenuated.
Raise the Q to 100 — swap the loose tank for a well-built, more sharply tuned stage — and the same 100 MHz center frequency narrows to just 1 MHz of bandwidth, ten times more selective at rejecting a neighboring station. Drop Q to 1, an overdamped, barely resonant circuit, and the bandwidth balloons to the full 100 MHz, meaning the circuit hardly discriminates between frequencies at all.
Questions
What does the Q factor actually measure?
Q measures how many radians of oscillation a resonant system completes before it dissipates the energy it started with — formally 2π times the ratio of energy stored to energy lost per cycle. Equivalently, Q is the center frequency divided by bandwidth, f₀ ⁄ BW. A high-Q circuit, like a quartz crystal, rings for thousands of cycles before its energy decays away; a low-Q circuit, damped hard by resistance, loses most of its energy within a cycle or two.
Why are the bandwidth's edges called half-power points?
Because that is exactly what they are: the two frequencies, one below f₀ and one above, where the circuit's output power has fallen to half its peak value — a 3 dB drop, hence the common label '−3 dB bandwidth'. Voltage or current amplitude at those points sits at 1 ⁄ √2, about 0.707, of its peak, not half, since power scales with amplitude squared. Everything inside that span passes with at least half its power; everything outside is progressively attenuated.
Does a higher Q always mean a better circuit?
Not automatically — it trades selectivity for speed and tolerance. A high-Q filter rejects nearby frequencies sharply, which is exactly what a radio tuner or a channel filter wants, but it also responds more slowly to changes and demands tighter component tolerances, since a small drift in f₀ can shift the whole narrow passband off target. A broadband amplifier or an audio crossover often wants deliberately low Q so it responds evenly across a wide range instead.
Can I use this formula for a mechanical resonator, not just an electrical one?
Yes — BW = f₀ ⁄ Q is general to any single, lightly damped resonant mode, mechanical or electrical. A struck wine glass, a tuning fork, and an LC tank circuit all obey it, because Q is defined from energy storage and loss, not from any one physical mechanism. Only the numbers differ: a wine glass might carry a Q in the low thousands, while a car's suspension, heavily damped on purpose, sits closer to 1.
Where does the letter Q come from?
It stands for 'quality factor', a term credited to K. S. Johnson of Western Electric's engineering department in the 1920s, during work on coils and filters for telephone transmission lines. Engineers needed one number to express how close a real coil came to the lossless ideal used in circuit theory, and Q — the ratio of reactance to resistance in its earliest form — became the standard shorthand still used today.
What happens to bandwidth as Q approaches zero?
It grows without bound — BW = f₀ ⁄ Q has no finite limit as Q shrinks toward zero, meaning an extremely lossy, heavily damped system stops behaving like a resonance at all and starts looking flat across frequency. In practice Q rarely goes below about 0.5 for a real second-order system; past that point the half-power points stop framing a symmetric peak around f₀, and one bandwidth number no longer describes the response well.