Capacitors in Series and Parallel – Equivalent Capacitance, Charge and Voltage
Combining capacitors is one of the first things anyone learns in circuit theory, and one of the easiest to get backwards. Capacitors combine opposite to resistors: in parallel they add directly, C_eq = C₁ + C₂ + … + Cₙ, while in series their reciprocals add, 1/C_eq = 1/C₁ + 1/C₂ + … + 1/Cₙ. This capacitors in series and parallel calculator evaluates either rule for any number of capacitors, mixes units freely, and then works out how charge, voltage and stored energy divide across the network.
Why the two rules point in opposite directions
Both follow from the plate picture. Capacitance rises with plate area and falls with plate separation, C = ε·A/d. Wiring two capacitors in parallel is electrically the same as widening the plates, so the values add and the result is always larger than the largest member. Wiring them in series stacks the gaps instead, increasing the effective separation, so the result is always smaller than the smallest member. For exactly two in series the reciprocal identity collapses into the familiar product over sum shortcut, C_eq = C₁·C₂ / (C₁ + C₂), and for N identical parts it reduces further to C/N in series and N·C in parallel.
How charge and voltage divide
This is where the two topologies genuinely differ in behaviour. In a parallel bank every capacitor sees the full supply, so the voltages are equal and the charge splits in proportion to capacitance: Qᵢ = Cᵢ·V. In a series string the charge is common to every member — the charge pushed onto one plate is pulled straight off the plate facing it — so Q = C_eq·V and the voltage divides inversely with capacitance, Vᵢ = Q/Cᵢ. The stored energy follows from E = ½CV² for the network and Eᵢ = ½CᵢVᵢ² for each member, and the parts must always sum back to the whole.
Vᵢ = Q/Cᵢ, the smallest capacitor in a series string carries the largest voltage. Two parts rated 25 V do not make a 50 V string unless they are closely matched: a 1 µF and a 10 µF in series across 30 V put 27.27 V on the small one and only 2.73 V on the large one. The safe supply limit here is V_max = min(Cᵢ × V_ratingᵢ) / C_eq, which works out at 27.5 V — well under the 50 V the ratings alone suggest. Real high-voltage strings add balancing resistors across each capacitor for exactly this reason.Mixed networks and design work
Ladder networks reduce in stages: collapse each group with its own rule, then combine the collapsed values with the opposite rule. Two parallel pairs of 10 + 10 µF and 22 + 47 µF become 20 µF and 69 µF, and putting those in series gives 15.506 µF. The calculator records every intermediate value so the reduction can be followed step by step. Running the problem backwards is just as common: given a target and the parts already in hand, C_missing = C_target − ΣC_known in parallel and 1/C_missing = 1/C_target − Σ(1/C_known) in series, with the nearest E12 and E24 preferred values suggested so the answer maps onto a part you can actually buy.
AC behaviour and tolerance
Once the equivalent capacitance is known, its capacitive reactance at any frequency follows from X_C = 1/(2πfC) — 6.875 µF presents about 231.5 Ω at 100 Hz, and less as frequency rises. That single number links the combination result to filter corners, coupling impedance and supply decoupling. Component tolerance matters just as much: recomputing the network with every value at its low and high extremes gives the true envelope, since both combination rules increase monotonically with each capacitance.
The combination arithmetic is exact, but real parts are not. Class-2 ceramics lose a large fraction of their capacitance under DC bias and with temperature, electrolytics routinely carry −20/+80 % tolerances and drift as they age, and every capacitor has equivalent series resistance and inductance that dominate its behaviour at high frequency.
Use the calculated value as the nominal design point, and measure the parts when the tolerance actually matters.
Where this comes up in practice
Beyond coursework, the same arithmetic covers a lot of bench work: building an odd capacitance out of stock parts, sizing a decoupling bank where several ceramics sit in parallel with a bulk electrolytic, stacking capacitors in series to raise the working voltage across a DC bus, trimming an RC or 555 timing network, and tuning an RF circuit where two series capacitors make a value no single part offers. In every case the questions are the same ones the calculator answers: what is the equivalent value, how much charge and energy does it hold, and which component is closest to its limit.