How this instrument works
Push charge onto a pair of conductors and voltage appears between them. Capacitance is that ratio — how many coulombs each volt buys. It belongs to geometry and materials, not to whatever charge happens to be sitting there: widen the gap or thin the dielectric, and capacitance shifts; but double the charge in an unchanged part, and voltage simply doubles to match. Ewald Georg von Kleist found this storage in 1745 using a nail pushed into a medicine bottle, then Pieter van Musschenbroek repeated it at Leiden the following year, giving electricity its first reservoir.
Scale is where intuition fails. A farad is grotesquely large. Earth itself, treated as an isolated sphere, has self-capacitance of roughly 710 microfarads. Ceramic chip parts hold picofarads through microfarads, electrolytics reach millifarads, and only supercapacitors — carbon electrodes separated by a nanometre-scale electrochemical double layer — reach whole farads. Michael Faraday, whose name rides on that unit after an 1881 vote in Paris, never handled a one-farad component.
C = Q/V presumes linearity, and plenty of real parts refuse to oblige. Class II ceramic dielectrics such as X7R or Y5V are ferroelectric: apply rated DC bias and measured capacitance can collapse by half or more, so a part marked 10 µF might deliver 3 µF where it actually matters. Varactor diodes are deliberately nonlinear, while supercapacitors drift with voltage as well. Treat this ratio as exact for vacuum, air, film and Class I ceramics, and as datasheet-dependent guidance everywhere else.
- Type your coulombs into Charge stored, or switch that field to mC, µC or nC if your figure came off your meter in those units.
- Enter Voltage across the plates in volts, millivolts or kilovolts. Zero is rejected, since dividing by it has no answer.
- Read Capacitance and choose the display unit: pF for radio-frequency work, µF for decoupling, whole farads for supercapacitors.
- Compare against the datasheet. Results far from nominal usually mean charge and voltage were captured at different moments, or under DC bias.
Worked example — the one-farad backup supercapacitor
A supercapacitor sits on the utility meter board, keeping its real-time clock alive through mains outages. During bench testing it accepted one coulomb of charge and settled at one volt across its terminals. Capacitance follows: C = 1 ⁄ 1 = 1 farad, exactly. That answer is not rounded — it is the farad's SI definition restated, because the farad simply is one coulomb per volt.
One coulomb deserves a moment's thought. It amounts to about 6.24 × 10¹⁸ electrons, and the 1 F part holding it at 1 V stores ½ × 1 × 1² = 0.5 joules, roughly what an AA cell supplies in one fifth of a second. Move that same coulomb onto a 100 pF trimmer and arithmetic turns absurd: you would need ten billion volts across millimetres of air.
Questions
Why is the farad such an inconveniently large unit?
Because it was defined from coulombs and volts, not from anything benchtop hardware could plausibly build. One farad means one whole coulomb per volt, and a coulomb is an enormous quantity of charge to park on metal. Parallel plates one millimetre apart in air would need roughly 113 square kilometres of area to hit that mark. Supercapacitors sidestep geometry by separating charge across nanometres at the carbon–electrolyte interface, which is how a thumbnail-sized part can be rated 10 F.
Is a charged capacitor actually carrying net charge?
No. Q here means charge magnitude on one plate; an equal negative charge sits on its partner, so any fully charged capacitor holds no net charge at all and weighs essentially what it weighed empty. What changed is arrangement — charge separated across the gap, with energy living in an electric field between plates. Mistaking plate charge for net charge is a genuinely common early stumble.
How does capacitance relate to permittivity and geometry?
For parallel plates, C = ε₀εᵣA/d: area over separation, scaled by relative permittivity εᵣ of whatever fills that gap. Vacuum permittivity ε₀ is 8.8541878188 × 10⁻¹² F/m. Hence capacitors get wound, stacked or etched — designers chase area while shrinking d. Q/V is the definition and always holds; that geometric form predicts a value before any part exists, but only for idealised plates with fringing fields ignored.
What happens if I enter zero volts?
This sheet stops, because Q/0 has no value. Physically, any capacitor holding charge shows some voltage, so measured readings of one coulomb at zero volts point to a fault or to a probe clipped onto the wrong node. Zero charge at zero volts is perfectly legitimate but says nothing about capacitance; you need both values nonzero to fix the ratio.
Does capacitance change with frequency?
Measured capacitance does, though Q/V itself never breaks. Dielectric polarisation cannot keep up with fast fields, so εᵣ falls as frequency climbs. Above its self-resonance, lead and package inductance dominates, so it behaves inductively — a 100 nF ceramic decoupler typically self-resonates somewhere between 10 and 30 MHz depending on package and mounting. That is exactly why datasheets quote the test frequency, commonly 1 kHz or 1 MHz.
How is the farad realised as a metrological standard?
Through quantum electrical standards rather than plates. Since 2019 the SI fixes elementary charge e and Planck constant h, which pin down Josephson and von Klitzing constants and therefore volt and ohm; capacitance then follows by impedance comparison. National laboratories also build calculable cross-capacitors on Thompson and Lampard's 1956 theorem, whose capacitance per unit length is ε₀ln2/π — about 1.953 pF per metre — independent of electrode size.