RC Snubber Calculator

Size the R and C across a switching device or a relay contact from figures you can actually measure: the ringing frequency on a scope, and the same frequency again with a known capacitor added. That gives the parasitic capacitance and the loop inductance, and from those the snubber capacitor, the resistor that damps the ring hardest, the damping the pair achieves, the peak voltage they save and the watts the resistor has to take.

Last updated 24 Sep 2026

RC snubber

Ring frequency → Rs and Cs

What is the snubber across?

How do you know the parasitic capacitance?

Ring frequency, as found (MHz)Measured on the switch node with a scope: count the cycles of the oscillation that follows the transition, or read the period off the screen. A 10:1 probe with a short ground spring, not a ground lead.

Capacitor added across the device (pF)A known film or C0G capacitor soldered across the device, chosen to drop the ring frequency by roughly half. Two to five times the capacitance you expect is about right.

Ring frequency with that capacitor fitted (MHz)

Parasitic capacitance Cpar (pF)Used only when you are entering it directly. It is the device’s own output capacitance at the working voltage plus the mounting and track capacitance. Use the figures from your part’s datasheet; typical values vary widely between manufacturers.

Snubber capacitor as a multiple of CparThe classic range is 4 to 10. Four is the smallest that gives strong damping; eight damps the ring critically; beyond that the resistor only gets hotter.

Voltage across the device when it is off (V)

Current step that excites the ring (A)The current that stops flowing in the loop inductance at the transition: the diode’s peak reverse-recovery current, or the switched current at a hard turn-off. For a relay, the load current being interrupted.

Switching frequency (kHz)Used for the resistor’s dissipation — the snubber capacitor is charged and discharged once per switching cycle.

Load inductance (mH)Relay case only: the inductance of the coil, solenoid or motor winding the contact is breaking.

Highest contact voltage you will allow (V)Relay case only. Below the contact’s rated breakdown and below the voltage at which the gap re-strikes — a few hundred volts on a low-voltage DC contact.

Contact operations per minuteRelay case only: how often the contact opens. This sets the snubber resistor’s average power, which is normally negligible.

An inductive load, the device that interrupts it, and the RC snubber across that device. The parasitic inductance that actually rings is the wiring and package inductance in this loop, not the load inductance drawn here; the ring frequency you measure on the switch node is what tells you its value. The symbol changes to a contact when the relay case is selected. No current dots are drawn — the current that matters here flows for a few hundred nanoseconds after the device opens.

91ΩExample

a switch node ringing at 6.195 MHz that falls to 2.631 MHz when 1,000 pF is added across the device, 4× snubber capacitance, 48 V off state, 1.5 A current step, 100 kHz

From two scope measurements to R and C

f1, f2: the ringing frequency before and after a known capacitor Cadd is soldered across the device
n: the snubber capacitor as a multiple of the parasitic capacitance; n = 8 gives critical damping exactly
ζ: damping factor of the pole pair that is left. ζ = 1 is critical — no overshoot, no ring
Rs: the value that maximises ζ for that n, derived on this page rather than taken as a rule of thumb

Worked example

a switch node ringing at 6.195 MHz that falls to 2.631 MHz when 1,000 pF is added across the device, 4× snubber capacitance, 48 V off state, 1.5 A current step, 100 kHz
C_par = 1,000 pF ÷ ((6.195/2.631)² − 1) = 220.1 pF
L = 1 ÷ ((2π × 6.195 MHz)² × 220.1 pF) = 2.999 µH, so the surge impedance √(L/C_par) is 116.7 Ω
C_s = 4 × 220.1 pF = 880.2 pF → 1 nF (E12, rounded up), which is 4.54× C_par
R_s = 116.7 Ω × (1 + 4.54)^0.75 ÷ 4.54 = 92.82 Ω → 91 Ω (E24, nearest)
Damping ζ = (√(1 + 4.54) − 1) ÷ 2 = 0.677, which leaves 0.31% of the ring after one cycle, at 2.97 MHz instead of 6.195 MHz
Peak voltage falls from 223.1 V to 122.4 V, and the resistor takes 230.4 mW from charging C_s plus 337.4 mW from the ring energy — 567.8 mW in all

What the ratio buys, and what the textbook rule gives instead

The last two columns are the familiar rule of thumb — make the resistor equal to the surge impedance √(L/Cpar) — evaluated against the same network. It is exactly right at a ratio of about 2.4 and increasingly conservative above that: at 8× it gives a damping of 0.53 where 1.00 was available. Every figure in the last column comes from the eigenvalues of the three-state network, not from a formula.
Cs ÷ Cpar	Best damping ζ	Rs for it	Rs = √(L/Cpar)	ζ that rule gives
1×	0.207	196.3 Ω	116.7 Ω	0.162
2×	0.366	133.1 Ω	116.7 Ω	0.354
3×	0.500	110.1 Ω	116.7 Ω	0.493
4×	0.618	97.59 Ω	116.7 Ω	0.535
6×	0.823	83.74 Ω	116.7 Ω	0.536
8×	1.000	75.83 Ω	116.7 Ω	0.529
10×	1.000	70.52 Ω	116.7 Ω	0.524
16×	1.000	61.09 Ω	116.7 Ω	0.515

Measuring a ring, and damping it

Every hard-switched node rings. A loop of copper has inductance, the device and its mounting have capacitance, and when the current in that loop changes abruptly the two exchange energy at a few megahertz until something absorbs it. The ring costs you three things: voltage overshoot that eats into the device’s rating, radiated and conducted emissions at the ring frequency and its harmonics, and occasionally false triggering elsewhere on the board. An RC snubber across the device adds a deliberate, lossy path that absorbs the energy.

Measure, do not calculate. Neither of the two values that matter can be computed from a schematic. Put a scope probe on the switch node — a 10:1 probe with the ground spring, not the long ground lead, or the probe’s own loop will ring too — and read the frequency of the oscillation after a transition. Then solder a known capacitor across the device, two to five times what you expect the parasitic to be, and read the frequency again. The ratio of those two frequencies gives the parasitic capacitance, and either frequency then gives the inductance. This is the method in Nexperia’s AN11160 and it is the only honest way to start.

Where the resistor value comes from. With the snubber fitted the node is a third-order network: the loop inductance, the parasitic capacitance, and the series R–C branch across it. Writing its characteristic equation with C_s = n·C_par and R_s = ρ·√(L/C_par) and normalising the frequency to 1/√(LC_par) gives σ³ + ((1+n)/nρ)·σ² + σ + 1/nρ = 0, and the damping of the surviving pole pair is largest when all three roots have the same magnitude. That happens at ρ = (1+n)^3/4/n, and the damping it gives is exactly ζ = (√(1+n) − 1)/2 — so a ratio of 8 is critical damping, no more and no less. Both results were checked here against the eigenvalues of the network for ratios from 1 to 16, and against a brute-force sweep of the resistor value. The familiar rule R_s = √(L/C_par) is the same answer at a ratio of about 2.4 and a conservative one above that; the table shows the gap.

What the resistor has to survive. The capacitor is charged to the off-state voltage and discharged again once per switching cycle, and both halves happen through the resistor, so it takes C_s·V² joules per cycle — not ½C_sV², which is only one of the two halves. Add the ring energy ½·L·I² that the snubber exists to absorb and you have the average power. It arrives as pulses of tens of nanoseconds, so the pulse rating matters as much as the average: use a thin-film or bulk-metal-foil part, or two in series, and never a wirewound, whose inductance defeats the whole exercise. The capacitor energy calculator gives the stored energy directly, and for the conduction and switching losses the snubber is protecting see the diode power loss calculator and the MOSFET loss calculator.

Relay contacts are a different problem. A contact opening on an inductive load does not ring at megahertz — it arcs, because the inductor will produce whatever voltage it needs to keep its current flowing and the gap between two separating contacts breaks down at a few hundred volts. There is nothing to measure with a scope, so the capacitor is sized from energy instead: it has to absorb the load’s ½LI² without the contact voltage exceeding the limit you set, which gives C_s = L·I² ÷ (V_max − V_supply)². The resistor’s job there is to limit the current that flows back out of the capacitor when the contact closes, and the traditional choice is the supply voltage divided by the load current, so the inrush is no worse than the load itself. On a DC load a flyback diode across the load is simpler and better; the RC is for AC, and for cases where a diode’s slow release is unacceptable. Mains voltage can kill. This page does the arithmetic only; for anything connected to the supply, your local wiring code and a licensed electrician decide what is allowed.

Frequently asked questions

How do you calculate an RC snubber?

Measure the ringing frequency, then measure it again with a known capacitor across the device. Those two numbers give the parasitic capacitance and the loop inductance. Take the snubber capacitor as four to eight times the parasitic capacitance, and the resistor as √(L/Cpar)·(1+n)^0.75/n where n is that ratio.

How do you measure the ringing frequency of a switching node?

With an oscilloscope on the drain or cathode, using a 10:1 probe and its ground spring rather than the flying ground lead — a 6 cm ground lead rings on its own and will give you the probe’s frequency, not the circuit’s. Trigger on the transition and measure the period of the first few cycles of the oscillation that follows it.

How much power does a snubber resistor dissipate?

Cs × V² × fsw from charging and discharging the capacitor once per cycle, plus ½ × L × I² × fsw for the ring energy it absorbs. At 1 nF, 48 V and 100 kHz that is 230.4 mW, and the ring adds 337.4 mW here — so a 1 W part.

What value capacitor should an RC snubber use?

Four to eight times the parasitic capacitance of the node. Below four the damping available is poor; at eight the ring is critically damped and there is nothing left to remove, so more capacitance only heats the resistor.

Does a snubber stop voltage overshoot?

It reduces it, by lowering the surge impedance √(L/C) that the interrupted current has to work against. Going from 220 pF to 1.22 nF of node capacitance cuts that impedance by a factor of 2.4, and the overshoot with it. It does not clamp: for a hard clamp you need an RCD snubber or a TVS.

Can I use an RC snubber across a relay contact?

Yes, and it is the standard fix for contact arcing on an AC load. Size the capacitor from the load’s stored energy rather than from a ring frequency, and choose the resistor so the capacitor’s discharge current at contact closure is no larger than the load current. On DC, a flyback diode across the load is cheaper and works better.

Related calculators

References

Nexperia. AN11160: Designing RC snubbers, Rev. 3.1, 21 October 2024. The two-measurement method: C_LK = C_add ÷ (x² − 1) with x = f_RING0/f_RING1, then L_LK from f = 1/(2π√(L_LKC_LK)); a snubber capacitor 1–2 times C_LK as a starting point and a resistor near the characteristic impedance √(L/C_s).
Severns R. Design of Snubbers for Power Circuits. Cornell Dubilier. The capacitance-addition method for the parasitic inductance, the choice of R_s = E_o/I_o so the capacitor’s discharge current at turn-on matches the load current, and P_diss = f_s × E_stored.
Erickson RW, Maksimović D. Fundamentals of Power Electronics, 3rd ed. Springer, 2020. Switching transitions, parasitic ringing and the energy accounting of a lossy snubber.