RC Snubber Resistors: Count Both Edges and the Repeating Capacitor State

System boundary

A series resistor-capacitor branch connected across a switching voltage. The model uses an ideal two-level source and constant components to calculate resistor energy; it is not a universal snubber design or printed-resistor pulse rating.

Integration interfaces

Integration risks

Identify the exact series RC branch and driving waveform

Draw the resistor and capacitor in series between the switching node and the stated return. This is different from a clamp with a diode, a capacitor discharging through a separate path or an actively regulated load. Name the voltage measured across the complete branch and the voltage across the resistor. The latter determines its instantaneous electrical loss, while the capacitor voltage records the state carried into the next transition.

Begin with the actual high and low durations rather than a frequency label alone. A burst sequence, pulse skipping or variable duty cycle may contain unequal intervals and long pauses. Record startup separately from the repeating condition. A snubber chosen to reduce a ringing peak can also respond to the main switching excursion, so its resistor load is not necessarily limited to the small ring visible after an edge.

A complete charge-discharge cycle contains two losses

For an ideal voltage step of magnitude V into an initially discharged series RC branch, the source supplies C times V squared during complete charging. Half remains in the capacitor and half is dissipated in the resistor. When the source returns to zero and the capacitor completely discharges through that resistor, the stored half is dissipated as well. Thus the complete settled cycle deposits C times V squared in the resistor.

For assumed capacitance of 10 nanofarads and a 10 V excursion, each completely settled transition deposits 0.5 microjoule and the pair deposits 1 microjoule. Multiplying the single-edge result by the cycle frequency would miss half the load. These are ideal-circuit energy values, not permissible pulse energy for a film, a terminal or a selected snubber component.

Solve the capacitor state when settling is incomplete

Let the source alternate between zero and V, remaining high for th and low for tl. Define a as exp(−th/RC) and b as exp(−tl/RC). In the repeating condition, the capacitor rises from Vlo to Vhi during the high interval and falls back during the low interval. Solving those two recurrence relations gives Vhi equal to V times one minus a divided by one minus ab, and Vlo equal to b times Vhi.

The denominator accounts for memory across successive cycles. Assigning zero before every rising edge violates that recurrence unless the low interval actually provides adequate discharge. The equations also distinguish the first startup pulse from later pulses. Use the component values and waveform for the state being studied; a variable switching sequence requires propagating each interval or directly measuring the resulting trajectory.

Calculate a short-interval example from that repeating state

Continue with 10 nanofarads and 10 V, now assuming 100 ohms, a one-microsecond high interval and a one-microsecond low interval. RC is one microsecond, so both decay factors are approximately 0.367879. The capacitor alternates between about 2.6894 and 7.3106 V rather than reaching zero and 10 V. The periodic energy is approximately 0.462117 microjoule per cycle.

The two-microsecond period corresponds to 500 kilohertz, giving approximately 0.231059 W average resistor loss. The complete-settling approximation would give 0.5 W at that cycle frequency, but complete settling does not occur under these assumed intervals. Peak current immediately after either ideal edge is about 73.1 milliamperes. A lower calculated average loss does not establish that this branch adequately damps the real switching circuit.

Check the energy balance without double counting the capacitor

In a periodic cycle the capacitor ends with the same stored energy with which it began. For the ideal source alternating between zero and V, source work during the high interval is V times the charge added, or V C times Vhi minus Vlo. The zero-level interval supplies no source work in this convention. The net source work is therefore exactly the resistor's cycle loss and produces the stated energy formula.

Alternatively, integrate resistor power over both intervals using the exponential currents and confirm the same result. This second calculation is a useful check on signs and event counting. If a real circuit has diode paths, source resistance, nonlinear capacitance or appreciable parasitic inductance, keep the energy in those elements and losses separate. The ideal balance should then be replaced by the actual circuit, not adjusted by an unexplained empirical multiplier.

Use an event ledger for burst and asymmetric operation

A waveform with several edges per control cycle needs a ledger of state changes, not one universal half-CV-squared entry repeated blindly. Track the capacitor voltage before and after each interval and add the resistor energies over the time horizon used for the thermal average. Keep peak voltage and current with each event because equal total energy can be delivered through very different local stresses.

For a long pause, the initial state of the next burst may approach the low source level more closely than during continuous switching. The first edge can therefore have a larger current than later edges. Treat that edge as its own condition. The following table identifies which simplifying calculation fits the known waveform and which additional information is needed.

Verify damping before interpreting a cooler resistor as improvement

The purpose of a snubber is usually to modify a particular transient or resonance. Reducing capacitance may reduce loss while leaving a larger switching spike or more ringing. Changing resistance alters damping and the time-dependent current, even when a complete-settling energy approximation happens to be independent of resistance. Evaluate the switching-node requirement alongside the resistor load rather than optimizing average watts alone.

Use appropriate probes and a controlled power-electronics test setup. The measurement loop itself can produce or hide high-frequency ringing, so retain its connection and bandwidth with the waveform. A custom fired thick-film resistor is not interchangeable with a packaged snubber resistor on the basis of nominal ohms. Its geometry, trim region, conductor transitions and mounted heat path require their own review under the actual repetitive stress.

Specify the full repetitive electrical load

Provide the source levels, edge waveform, switching sequence, capacitor data and resistor-terminal current and voltage. Include the calculated cycle energy, average power, peak quantities and starting state after an idle interval. Keep those electrical outputs separate from any subsequent endurance test. This allows the drawing-specific resistor request to represent the load it will actually experience.

Recalculate when duty cycle, pulse skipping, capacitor technology or the source excursion changes. An unchanged switching frequency does not imply unchanged stress when the capacitor no longer resets in the same way. The resulting event ledger gives engineering and production a reproducible definition of the repetitive load while preserving the separate decisions about damping effectiveness, insulation and component life.