This article based on prof. Sam Ben-Yaakov video explains so‑called two‑capacitor paradox theory and practical implications to hardware designers.
The two-capacitor paradox is a classic thought experiment that appears to violate energy conservation when charge is redistributed between capacitors. This article walks through an intuitive resolution of the paradox, based on Sam Ben‑Yaakov’s derivation and explanations, and translates the key ideas into practical insight for design engineers working with real components and circuits.
In the two‑capacitor paradox, a charged capacitor is suddenly connected to an uncharged capacitor of equal value. A straightforward calculation shows that half of the initial stored energy appears to “disappear” after charge sharing. Rather than being a mathematical curiosity only, this situation is a useful lens for understanding where energy really goes in real passive networks, how resistance and inductance shape transient behavior, and why idealized models have hard limits for engineering practice.
Statement of the two‑capacitor paradox
The classical setup assumes two identical capacitors C1=C2=CC_1 = C_2 = C. Initially, capacitor C1C_1 is charged to voltage V1V_1, while C2C_2 is uncharged. When the two capacitors are connected in parallel, charge redistributes and the final voltage on both capacitors is V1/2V_1/2. The energy before and after the connection can be written in terms of the usual capacitor energy formula E=12CV2E = \tfrac{1}{2} C V^2.
- Initial energy (only C1C_1 charged):
Einitial=12CV12E_{\text{initial}} = \tfrac{1}{2} C V_1^2. - Final energy (both capacitors at V1/2V_1/2):
Efinal=2⋅12C(V12)2=14CV12E_{\text{final}} = 2 \cdot \tfrac{1}{2} C \left(\tfrac{V_1}{2}\right)^2 = \tfrac{1}{4} C V_1^2.
Half of the initial energy appears to be missing after the connection. The central question is: where did this energy go, if charge is conserved and no explicit loss elements were included in the ideal circuit model?
Historical resolution: resistance cannot be ignored
A key historical result, credited in the talk to Charles Zucker (1955), is that one cannot meaningfully discuss “two capacitors connected together” without acknowledging resistance somewhere in the circuit. Even if the connection is very low impedance, a real system always includes nonzero series resistance from conductors, capacitor ESR, contacts, or the surrounding environment.
Once a resistance RR is included in series with the two capacitors during charge transfer, the paradox resolves cleanly:
- The “missing” energy is dissipated as heat in the resistance during the transient.
- The interesting and perhaps counter‑intuitive mathematical result is that the total energy dissipated in RR does not depend on the value of RR itself, as long as it is nonzero.
Whether the resistor is small (large initial current, short time constant) or large (small initial current, long time constant), the integral of I2RI^2 R over time yields the same energy loss. In both cases, the system ends in the same steady state: two capacitors at the common final voltage, with half of the initial energy dissipated in the resistor.
Generalized formulation with arbitrary capacitors and voltages
The presentation goes beyond the symmetric case and considers a general scenario with:
- Two capacitors C1C_1 and C2C_2 of possibly different values.
- Initial voltages V1V_1 and V2V_2 across each capacitor.
- A finite resistance RR in the loop when they are interconnected.
To make the analysis transparent, each charged capacitor is modeled as:
- An ideal capacitor with zero initial voltage in series with
- An ideal voltage source equal to its initial voltage.
From the terminal perspective, this series combination is equivalent to a single charged capacitor because it reproduces the same voltage–charge relation at the external terminals. By replacing both charged capacitors in this way and then combining the two series voltage sources, the problem reduces to:
- A single equivalent capacitor (resulting from the series connection of the two capacitors) driven by a single source of magnitude ΔV=V1−V2\Delta V = V_1 – V_2.
- A series resistance RR in the loop.
A crucial insight is that only the voltage difference ΔV\Delta V matters for the subsequent transient; the absolute values of V1V_1 and V2V_2 can be large but do not directly affect the incremental energy exchange. The effective circuit is just a first‑order RC system charged from ΔV\Delta V, which leads to an exponential current and corresponding power dissipation in RR.
Integrating I2RI^2 R over time yields a closed‑form expression for the dissipated energy that is independent of RR. The energy loss depends on the capacitors and the initial voltage difference but not on how large or small the resistance is, as long as it is nonzero. Re‑specializing this result to the symmetric case C1=C2C_1 = C_2, with C2C_2 initially at zero voltage, reproduces the “half of the initial energy is lost” result.
Intuitive time‑domain view: small vs large resistance
From a design perspective, it is useful to complement the mathematics with a qualitative current‑waveform picture. Depending on the value of RR:
- Small RR: The initial current is large and decays quickly with a short time constant. The current waveform is tall and narrow.
- Large RR: The initial current is smaller, but the transient lasts longer. The current waveform is lower but extends over a longer time.
In both situations, the root‑mean‑square (RMS) value of the current over the full transient, combined with the resistance, integrates to the same total dissipated energy. This explains physically why the energy loss does not depend on RR: a higher peak current over a short time or a lower current over a long time both lead to the same integrated I2RI^2 R.
For engineers, this viewpoint connects directly to practical concerns such as peak‑current stress, switch sizing, and thermal management. Even though the total dissipated energy is fixed by the initial capacitor states, the peak current and time constant do depend strongly on RR, which impacts component ratings and EMI behavior.
Adding inductance: oscillatory exchange and damping
If an ideal inductor is added in series with the capacitors and there is no resistance, the system becomes a lossless LC oscillator. Assuming again that only one capacitor is initially charged:
- Energy oscillates between the electric field in the capacitors and the magnetic field in the inductor.
- The voltage on each capacitor and the current through the inductor vary sinusoidally, with current and capacitor voltage shifted by 90 degrees.
- At instants when the inductor current is zero, all energy resides in the capacitors; at current peaks, energy is stored in the inductor.
Because the system is lossless by assumption, the total energy is conserved and no net dissipation occurs. However, once any realistic resistance is included (series resistance, dielectric loss, winding resistance, or radiation), the oscillation decays over time. Eventually the system again settles in the same final state as the purely resistive case: both capacitors at the common final voltage, with the “excess” energy dissipated as heat in the resistive elements.
This LC view highlights that in real hardware, parasitic inductances are always present (traces, leads, wiring). They can significantly shape current waveforms and voltages during transients, but they do not change the fundamental conclusion about where the energy ultimately goes: losses associated with resistance.
The ideal capacitor thought experiment and its limits
The most extreme version of the paradox considers two perfectly ideal capacitors connected by a perfectly ideal conductor:
- Two equal capacitors, one initially at V1V_1, the other at 0.
- Absolutely no resistance anywhere (including ESR, wiring, or contacts).
- No inductance, no radiation, no mechanical motion of plates; in other words, a purely lumped ideal two‑capacitor system.
In such a hypothetical system, formal circuit analysis predicts an infinite current spike of zero duration when the capacitors are connected. Mathematically, this corresponds to a current impulse (delta function) with unbounded frequency content. The standard energy accounting Einitial≠EfinalE_{\text{initial}} \neq E_{\text{final}} then seems to contradict energy conservation, and the simplistic formula “power equals voltage times current” at the instant of switching becomes ill‑defined (0⋅∞20 \cdot \infty^2).
A variety of proposed resolutions in the literature invoke additional mechanisms such as radiation, stray inductance, or mechanical motion to absorb the missing energy. However, each of these implicitly relaxes the strict ideality assumptions and includes extra physical elements beyond just two perfect capacitors and a perfect conductor. In the framework presented here, once ideality is enforced rigorously—no resistance, no inductance, no radiation, no motion—the paradox has no physically meaningful resolution simply because the assumed system cannot exist in the real world.
From a practical engineering standpoint, this is not a problem. Any realizable circuit will always include some nonzero resistance and often non‑negligible parasitic inductance and radiation. For such real systems, the two‑capacitor paradox disappears, and standard energy conservation holds unambiguously with dissipation in resistive elements.
Practical implications for passive component design‑in
Although the paradox is a theoretical construct, it has several practical takeaways for engineers working with capacitors in power and signal circuits.
Energy redistribution and surge currents
Whenever charge sharing occurs between capacitors at different voltages, the same mechanisms are at play as in the paradox:
- Connecting a charged bulk capacitor to a partially discharged bus redistributes energy, and part of the initial energy will be dissipated as heat in resistances.
- Hot‑plugging boards or sub‑systems with large capacitance onto live backplanes can create significant inrush currents and heating in connectors, traces, and protection components.
- Series resistances (dedicated NTCs, inrush resistors, active current limiting) tune the peak current and time constant but do not change the fact that the “excess” energy will be dissipated somewhere in the resistive path.
Understanding that energy loss depends on the initial voltage difference and capacitances, not on the specific value of RRR, helps in estimating worst‑case thermal loads and designing safe current‑limiting networks.
Role of ESR and parasitics
In real capacitors:
- ESR and lead/winding resistance act as the RR in the paradox and convert part of the energy into heat during sudden transients.
- Parasitic inductance, especially in larger or leaded parts, can introduce ringing and overshoot if not critically damped by resistance.
From the paradox analysis, designers can appreciate that:
- Lower ESR reduces instantaneous heating in the capacitor but can increase initial current peaks and stress external series resistance or switches.
- Any realistic attempt to approach “zero resistance” conditions will run into layout‑induced inductance and radiation paths that reintroduce losses via different mechanisms.
Implications for “lossless” switching concepts
The thought experiment also touches on the idea of switching capacitive nodes without loss. Even if a switch is modeled as an ideal short, the electrodes and conductors form capacitances and inductances. When nodes at different potentials are connected, charge flows, waves propagate, and energy is necessarily redistributed. In practice:
- There is no way to transfer charge between two nodes at different potentials in finite time without dissipating energy in some resistive or radiative mechanism.
- Techniques labeled as “lossless” or “soft” switching generally reduce, but do not eliminate, these dissipations by shaping voltage and current waveforms (for example ZVSZVS and ZCSZCS in power electronics).
The two‑capacitor analysis provides a compact reference case that underscores the fundamental limitation: energy associated with voltage differences cannot be perfectly preserved during reconnection.
Design‑in notes for engineers
Based on the above insights, several concrete design guidelines emerge for those selecting and applying capacitors and associated passives.
Managing inrush and charge‑sharing events
- When connecting large capacitors to a rail, explicitly calculate the initial stored energy E=12CV2E = \tfrac{1}{2} C V^2 and assess how much of that may become dissipated during the connection.
- If a pre‑charged module is connected to a different potential, treat the situation as a two‑capacitor problem and estimate the final shared voltage and the amount of energy to be dissipated.
- Insert controlled resistance or current‑limiting circuitry where necessary to keep peak currents within component ratings, even though the total dissipated energy remains driven by the initial conditions.
Selecting capacitors and resistive paths
- For high‑energy nodes, prioritize capacitors whose ESR and construction can safely handle transient heating due to charge redistribution events, according to manufacturer datasheet limits.
- Use dedicated resistive elements (or active circuits) to offload dissipation from sensitive components such as connectors or relays, ensuring that the resistors are thermally rated for the expected energy pulses.
- Be aware that making interconnects “too ideal” (very low resistance and inductance) can increase stress on other parts of the system by driving very steep current pulses.
Accounting for inductance and layout
- In power and high‑speed circuits, assume that traces and leads add non‑negligible inductance, leading to LC oscillations during fast switching of capacitive loads.
- Apply damping strategies (intentional series resistance, snubbers) to prevent excessive ringing and to control where energy is dissipated.
- Board layout, loop area minimization, and careful routing of high‑di/dt paths are essential not only for EMI control but also for keeping voltage overshoots and current spikes within safe bounds.
Conclusion
Key takeaways for practicing engineers
By reframing the two‑capacitor paradox in terms of realistic circuits with nonzero resistance, the apparent violation of energy conservation disappears: the “missing” energy is simply dissipated as heat in resistive elements during the transient. Generalizing to capacitors with different values and voltages shows that only the voltage difference and equivalent capacitance determine the dissipated energy, and that the result is independent of the exact resistance value. For real designs, the paradox provides a compact mental model for understanding energy redistribution, inrush currents, and the role of ESR and parasitics in capacitor networks.
Engineers can use this understanding to better specify capacitors, resistors, and interconnects in power and signal applications where charge sharing or hot‑plug events occur. Although the idealized limit of zero resistance and zero parasitics remains a mathematically interesting but physically unrealizable scenario, every real system sits comfortably on the side where standard circuit theory and energy conservation agree, provided all loss mechanisms are acknowledged.
Source
This article is based on the educational presentation “Intuitive examination of the two‑capacitors paradox” by Sam Ben‑Yaakov, with interpretations and explanations adapted for a passive component engineering audience.