How this instrument works
Damping ratio weighs the resistance a system actually has against however much would just barely stop it oscillating: ζ = c ⁄ 2√(k·m). Being a pure number, it is what lets a wine glass, a car body and a bridge deck be argued about on one scale. Write its equation of motion in standard form, s² + 2ζω_n·s + ω_n² = 0, and ζ alone fixes the character of that answer — amplitude decays inside an envelope e^(-ζω_n·t) while the ringing underneath runs at ω_d = ω_n·√(1 - ζ²). Shape belongs to ζ; ω_n merely sets the clock speed.
Both the symbol and that 2ζω_n convention were cemented by servomechanism engineers during 1940s war work, when the Radiation Laboratory volume Theory of Servomechanisms (James, Nichols and Phillips, 1947) put second-order standard form in front of an entire generation of control designers, where it has stayed ever since. In practice ζ is rarely computed from c, k and m at all — it gets measured, by striking something and watching how fast its ring dies away. Typical values span five orders of magnitude: silicon MEMS resonators under vacuum sit near 10⁻⁵, welded steel lands around 0.005, bolted steel 0.02, reinforced concrete buildings 0.05 (whatever figure seismic codes assume when publishing response spectra), and sealed loudspeaker boxes tuned to that classic 0.707 alignment.
This formula assumes one lump of mass on one spring with one dashpot, and that dashpot force tracking velocity in strict proportion. Where those fail, ζ stops being a property and becomes a fitted number. Dry rubbing is the clearest case: Coulomb friction decays a ring-down along a straight line rather than an exponential, so whatever ζ you extract shrinks with every pair of peaks you pick. Hysteretic loss in elastomers and bolted joints errs the other way, staying roughly constant per cycle regardless of frequency, which is why structural work often quotes a loss factor η instead, with η ≈ 2ζ near resonance. And ζ attaches to a mode, never to a body: a floor slab carries one value in its first bending mode and quite another in its second.
- Enter Damping coefficient (N·s/m) — your dashpot's rating, force per unit of velocity, not force per unit of travel.
- Enter Mass in kilograms; only whatever genuinely moves counts, with grams and tonnes available on its dropdown.
- Enter Spring constant (N/m) as total stiffness felt by that mass — two 50 N/m springs side by side make 100.
- Read Damping ratio: below 1 it overshoots and rings, at exactly 1 it just avoids overshoot, above 1 it returns slowly without ever crossing.
- Check Critical damping coefficient underneath — the N·s/m figure your damper is being divided by.
Worked example — a 10 kg carriage on a 100 N/m spring
A 10 kg carriage rides on a spring rated 100 N/m, with a 40 N·s/m dashpot bridging it to the frame. Enter those three figures: Critical damping coefficient = 2√(100 × 10) = 2√1000 = 63.2455532 N·s/m, and Damping ratio = 40 ⁄ 63.2455532 = 0.632455532. Below 1, so this carriage will overshoot — though arithmetic here is unusually tidy, since ζ² lands on exactly 0.4.
Follow it through and the numbers stay clean. Natural frequency ω_n = √(100 ⁄ 10) = √10 = 3.162 rad/s, with ringing itself slower at ω_d = 3.162 × √0.6 = 2.449 rad/s, a damped period of 2.57 s. Its envelope decays at ζω_n = c ⁄ 2m = 40 ⁄ 20 = exactly 2 per second, so the time constant is half a second and 2% settling arrives near 4 ⁄ 2 = 2.0 s. First overshoot works out at exp(-π × 0.6325 ⁄ 0.7746) = 7.7% past target, and its successor stays under 1%. Pull that carriage 100 mm aside and let go: one visible bounce of roughly 8 mm, then stillness.
Questions
Does damping ratio have units?
No — it is a pure number, which is rather its whole point. Coefficient c carries N·s/m and scales with the size of your hardware; dividing by 2√(k·m), which carries those same units, cancels them and leaves a figure comparable across any second-order system anywhere. You will often meet it written as a percentage instead: '5% damping' and ζ = 0.05 mean precisely the same thing, both being shorthand for five percent of critical.
How is damping ratio measured on real hardware?
By ring-down, usually. Displace your system, release it, record that decaying oscillation. Logarithmic decrement δ = (1 ⁄ n)·ln(x₀ ⁄ x_n) taken over n cycles converts directly, via ζ = δ ⁄ √(4π² + δ²); for lightly damped things plain δ ⁄ 2π is close enough. A frequency-domain alternative is the half-power method — sweep for resonance, locate two frequencies sitting at 1 ⁄ √2 of peak amplitude, then ζ ≈ (f₂ - f₁) ⁄ 2f₀.
What does a damping ratio greater than 1 mean?
Overdamped. Roots of its characteristic equation become real and distinct, so nothing oscillates and nothing overshoots — motion just leaks back toward equilibrium as a sum of two decaying exponentials. It is also slower than ζ = 1, which regularly surprises people. A heavy door closer wound to ζ = 2 takes visibly longer to shut than that identical door at ζ = 1, because whichever exponential decays more slowly dominates the tail.
How does damping ratio relate to Q factor?
Q = 1 ⁄ 2ζ, so they are two readings off one dial. A tuning fork at Q = 1000 has ζ = 0.0005; anything at ζ = 0.5 has Q = 1. Q is the native language wherever resonance is wanted and sharpness matters — filters, antennas, quartz oscillators — whereas ζ gets preferred wherever a transient has to settle down and behave. Note this relation holds exactly only for a linear viscous second-order system; outside that, both turn into fitted numbers.
Can damping ratio be negative?
Yes, and it is worth recognising what that looks like. Negative ζ means energy is being fed into motion rather than drained from it, so the envelope grows as e^(+|ζ|ω_n·t) instead of shrinking. Aeroelastic flutter is the textbook case: airflow over a wing or a bridge deck can return net energy each cycle, driving effective damping through zero at some critical airspeed. Control loops manage it too, whenever excess gain turns a placid plant into an oscillating one.
Why did my damping ratio change when I added mass?
Because ζ describes an entire system, never its damper by itself. It lives in c ⁄ 2√(k·m), so bolting extra mass on while keeping identical dashpots and springs drags ζ down by the square root of your mass ratio — quadruple that mass and ζ halves, turning a well-behaved response into a ringing one. That trips people up more than anything else here: treating some dashpot's N·s/m rating as though a fixed ζ travelled with it. None does. ζ exists only once you have also named both a mass and a stiffness.