DC-Link Capacitors: Sizing, Ripple and Selection

DC-Link capacitors form an essential stage in power conversion for many applications, including three-phase Pulse Width Modulation (PWM) inverters, photovoltaic and wind power inverters, industrial motor drives, automotive onboard chargers and inverters, medical equipment power supplies, etc. DC-Link capacitors buffer the instantaneous power difference between a source and a converter, limiting DC bus voltage ripple and supplying local switching current. This article explains their operation, technology choices and selection requirements for design engineers working on inverters, onboard chargers, industrial drives, renewable energy converters and power supplies.

Key Takeaways

A DC-Link capacitor connects across the DC bus rails. Its main tasks are energy buffering, ripple-current handling and reduction of the impedance seen by the switching converter.
Required capacitance and required ripple-current capability are separate constraints. A bank that satisfies voltage ripple or hold-up requirements can still overheat.
Aluminum electrolytic capacitors are often effective for bulk energy storage; metallized film capacitors are strong candidates for high ripple-current duty; ceramic capacitors can provide compact local decoupling.
Compare effective capacitance, frequency-dependent ESR and impedance, permissible voltage, ripple-current capability and hotspot temperature at the actual operating conditions.
Fast SiC and GaN switching increases the importance of the complete commutation-loop inductance. Adding ceramic capacitance to a bulk bank also requires an anti-resonance assessment.
Validate the assembled bank over its electrical and thermal mission profile, including startup, regenerative operation, cooling conditions and safe discharge.

Basic Function of DC-Link Capacitors

A capacitive DC link provides an intermediate energy buffer in a voltage-source power converter. It may sit between a rectifier and an inverter, between an AC/DC front end and a DC/DC stage, or directly across the battery-fed input of a traction inverter. A preceding rectifier is therefore not essential to the definition. [1] The capacitor charges when incoming current exceeds outgoing current and discharges when the converter draws more current than the source supplies. This reduces the resulting bus-voltage variation and prevents the upstream source and its wiring from having to supply every switching-current pulse directly. Let i_C be capacitor current, i_source the current entering the DC link from the source, i_converter the current leaving towards the converter, C the capacitance and v_C the voltage across the ideal capacitive element. With positive capacitor current defined as charging current, Eq. (1) describes the instantaneous current balance.

i_C(t)=i_{\mathrm{source}}(t)-i_{\mathrm{converter}}(t)=C\frac{\mathrm{d}v_C(t)}{\mathrm{d}t} \tag{1}

Over an interval from t₁ to t₂, the change in capacitor voltage, Δv_C, follows Eq. (2). These relationships assume approximately constant capacitance and do not include the additional terminal-voltage contributions from ESR and inductance.

\Delta v_C=\frac{1}{C}\int_{t_1}^{t_2}i_C(t)\,\mathrm{d}t \tag{2}

Three duties should be distinguished. Bulk energy storage supports slower power imbalance or hold-up. Ripple-current handling accommodates repetitive current exchanged by the converter. Local decoupling supplies rapid commutation current through a short, low-inductance path. One capacitor assembly may contribute to all three, but the requirements should be checked separately. [1, 2]

Types of DC-Link Capacitors

Aluminum electrolytic, metallized film and ceramic capacitors offer different combinations of energy density, current capability, loss, size and reliability. Table 1 summarizes their usual roles; it is not a ranking that applies to every voltage, package or application. [1]

Technology	Main advantages	Selection limits	Typical roles
Aluminum electrolytic	High bulk capacitance and attractive cost for stored energy	Polarized; ESR, temperature, ripple current and electrolyte-related wear-out require attention	Rectified supplies, general-purpose drives, UPS and hold-up storage
Metallized polymer film	Low ESR, high ripple-current capability and available self-healing constructions	Lower capacitance density than electrolytics in energy-storage-led comparisons; temperature, humidity and terminal layout remain important	Traction inverters, renewable energy converters and high-current DC links
Ceramic	Compact low-inductance decoupling and useful high-frequency performance	Dielectric-specific DC bias, temperature and ageing behaviour; mechanical stress and current capability require verification	Local power-stage decoupling, distributed DC links and selected compact converter designs

Table 1. Main DC-Link capacitor technologies and their application-dependent trade-offs.

Bulk energy storage versus ripple-current capability

For a design dominated by stored energy or hold-up time, an electrolytic solution can be smaller and more economical than a film solution. In a design dominated by ripple current, an electrolytic bank may require additional parallel units simply to meet its thermal limit. A film bank can then become competitive in total size and cost even if it contains less nominal capacitance. [1]

Compare complete solutions at the same bus voltage, permissible ripple, thermal conditions and lifetime target. Capacitance alone does not describe equivalent functionality.

Ceramic dielectric choice and effective capacitance

Class II ceramic capacitors, including X7R parts, can lose substantial capacitance under DC bias. Nominal capacitance measured under small-signal catalog conditions may therefore overstate the capacitance available in a high-voltage DC link. Temperature, tolerance and ageing also need to be included in the effective-capacitance assessment. [2]

Class I ceramics offer more stable capacitance but generally lower capacitance density. Specialized power ceramics can have different bias and temperature characteristics from conventional Class II MLCCs. Use the curves for the selected dielectric and part rather than applying one ceramic derating assumption to all technologies. [2]

Applications

The DC-link duty depends on the source, load, converter topology and control method. The same capacitor technology can face very different current spectra in a mains-fed drive, a battery-fed inverter and an onboard charger. [1]

Renewable energy: photovoltaic and wind converters use DC links to buffer power differences between conversion stages and limit bus-voltage variation.
Electric vehicles: traction inverters use DC-link capacitance to supply switching current locally, while onboard chargers require energy buffering between their AC/DC and DC/DC functions.
Industrial drives: the bank handles current exchanged between the rectifier or active front end and the motor inverter, including operating conditions associated with acceleration and braking.
Power supplies: UPS, switched-mode and medical equipment power supplies may require both ripple filtering and a defined hold-up interval. Hold-up requirements can dominate the required capacitance.

Design Considerations

Begin with the converter requirements, not a preferred capacitor family. Establish the bus-voltage range, allowable ripple, operating modes, current spectrum, environmental conditions and required service life before comparing parts. A reliability-oriented design evaluates electrical stress, thermal stress and lifetime together. [1]

Capacitance: determine the minimum effective value required by power imbalance, voltage ripple, transient response and any hold-up requirement.
Voltage: include the highest normal bus voltage, superimposed ripple, startup behaviour and relevant transient or regenerative conditions. Apply the selected manufacturer’s voltage-temperature limits.
Ripple current: calculate the current through the capacitor bank rather than substituting the load current or average DC input current.
ESR and inductance: compare their frequency dependence and the complete connection geometry, including busbars, PCB traces and terminals.
Temperature and lifetime: assess internal heating, nearby heat sources, cooling, humidity and the duration of each operating condition.

DC-Link Capacitors in Automotive Applications

Automotive DC links must combine electrical performance with mechanical robustness and a suitable thermal mission profile. The onboard charger and traction inverter share the need for energy buffering, but their voltage, capacitance and ripple-current requirements should not be assumed identical. [1, 2]

Power conversion configurations using an intermediate DC-link capacitor — Figure 1. Common uses for DC-link capacitors in power conversion systems. Original source credit: Center of Reliable Power Electronics, Aalborg University.

Onboard charger

In the conventional two-stage onboard charger represented in Figure 2, the AC/DC front end rectifies the mains input and performs power factor correction. An intermediate DC link buffers energy before a downstream DC/DC converter regulates the charging output. This description applies to that architecture; it does not imply that every charger uses the same arrangement.

The bulk capacitor, C_bulk, connects across the intermediate DC rails. Its voltage rating follows the actual regulated bus and transient limits, not simply the nominal traction-battery voltage. Its capacitance follows the permitted bus-voltage excursion and the energy that must be buffered.

For a single-phase sinusoidal mains input with approximately unity power factor and approximately constant output power, instantaneous input power contains a component at twice the line frequency. The energy buffer must accommodate that power pulsation unless another part of the architecture actively decouples it. This is distinct from switching-frequency ripple and is not a universal description of every three-phase charger. [1]

A practical bank may combine bulk electrolytic or film capacitance with smaller local ceramic capacitors. Adding ceramics does not by itself remove the bulk energy requirement. The array must be evaluated using effective capacitance, current distribution, losses and possible resonances. [1, 2]

Traction inverter

The traction inverter converts battery DC into controlled three-phase motor power. During regenerative operation, power flows back towards the DC bus and battery. Motor-position sensing, current sensing and gate control belong to the wider drive system; they are not functions of the DC-link capacitor.

Battery-fed motor inverter with a smoothing capacitor connected across the DC input rails — Figure 3. Simplified block schematic for an inverter in a motor-drive circuit.

The smoothing capacitor, C_s, connects across the inverter’s positive and negative DC rails. It does not connect between a DC terminal and an AC motor terminal. Its local current path reduces the switching-current demand on the battery interconnect, whose parasitic inductance would otherwise contribute to voltage excursions. [2]

In a battery-fed inverter, high switching-current capability can be more important than long hold-up time. Film capacitors often provide the bulk DC-link function, while appropriately selected ceramic capacitors provide local high-frequency decoupling near the switching devices. The final arrangement must be validated as a combined electrical network. [1, 2]

Electric-vehicle application illustration accompanying the DC-link capacitor discussion — Figure 4. Automotive power conversion is an important application for DC-link capacitors. Original image source credit: Alternative Fuels Data Center, US Department of Energy.

For a given transmitted power, increasing DC voltage reduces the required average current. This can reduce conduction losses in a given interconnect, but also increases insulation, clearance and component-voltage requirements. Higher voltage is therefore a system trade-off rather than an automatic guarantee of higher performance.

Hybrid and electric vehicle powertrain architecture showing interconnected power conversion functions — Figure 5. HEV/EV powertrain architecture. Original source credit: Keysight Technologies.

EV inverter system divided into battery input, DC-link buffer and switching inverter blocks — Figure 6. EV/HEV system block diagram, separating the source, DC-link buffer and inverter functions.

Read Figure 6 as three functional blocks: the battery supplies DC power, the DC link buffers the difference between source and inverter current, and the switching bridge converts DC into motor AC. The capacitor is an energy-storage element, not an additional active conversion stage. The actual inverter topology and its control determine the current waveform that the bank must handle.

Why the DC-Link Capacitor Is Vital

The DC-link capacitor carries the difference between source current and converter input current. Consequently, its RMS ripple current cannot generally be obtained from load current alone: modulation, operating point, source impedance and the switching strategy all influence the waveform. [1]

Ripple current and voltage ripple are different limits

Ripple current is the time-varying current component through the capacitor. For a steady periodic operating condition with no net charge accumulation, the average capacitor current is approximately zero, although its RMS value can be substantial. A ripple-current rating is a thermal operating limit at stated conditions; a voltage rating is a separate electrical limit.

Voltage ripple includes the change in stored charge and voltage contributions from ESR and parasitic inductance. The original duty-cycle-based sizing relationship is retained in Figure 7. Treat such expressions as topology-specific approximations: a three-phase inverter requires the current waveform associated with its actual modulation and operating point rather than an unspecified duty-cycle shortcut. [1]

Duty-cycle-based relationship for minimum capacitance and allowable peak-to-peak ripple voltage — Figure 7. Duty-cycle-based ripple-sizing relationship. Its applicability depends on the converter topology and assumed capacitor-current waveform.

In the original notation, C_MIN is the required minimum capacitance, I_OUT is output current, D_Cycle is duty cycle, f_SW is switching frequency and V_pp(max) is allowable peak-to-peak ripple voltage. For a general charge-balance calculation, use Eq. (2) with the actual capacitor current.

Frequency-dependent impedance

A useful first-order model contains capacitance C, equivalent series resistance ESR and equivalent series inductance L_s. In Eq. (3), Z is impedance, f is frequency and j is the imaginary unit. This lumped model helps explain the main behaviour, but measured or manufacturer-supplied impedance data are preferable when internal resonances become important. [1]

Z(f)=\mathrm{ESR}(f)+j\left(2\pi fL_s-\frac{1}{2\pi fC}\right) \tag{3}

Below series resonance, capacitive reactance usually dominates. Near resonance, the reactive terms cancel in the simple model. Above resonance, inductance dominates and impedance rises with frequency. Self-inductance therefore does not provide unlimited improvement in high-frequency attenuation.

There is no universal switching-frequency threshold at which every film capacitor becomes unsuitable. Suitability depends on impedance, current spectrum, pulse stress, internal construction and the external current loop. The switching repetition frequency and the much faster switching-edge behaviour must both be considered. [2, 3]

Design Considerations in Selecting an Inverter DC-Link Capacitor

Selection requires more than matching nominal capacitance and voltage. A complete assessment checks effective capacitance, losses, current distribution, thermal conditions and lifetime across the intended operating envelope. [1]

Ripple-current spectrum and losses

Calculate or simulate the capacitor current over relevant operating modes, including partial load and regenerative operation where applicable. A rectangular or triangular waveform should follow from the circuit analysis, not from the appearance of a simplified block diagram.

For harmonic components, Eq. (4) estimates resistive loss using the ESR at each frequency. Here P_loss is dissipated power, I_k,rms is the RMS current of harmonic k and f_k is its frequency. The familiar I² × ESR expression is a useful single-frequency or effective-ESR approximation, not a reason to ignore frequency dependence. [1]

P_{\mathrm{loss}}\approx\sum_k I_{k,\mathrm{rms}}^2\,\mathrm{ESR}(f_k) \tag{4}

Use consistent ESR and loss definitions so that dielectric losses are not counted twice. Compare ripple-current ratings only after checking their reference frequency, temperature, cooling assumptions and permitted temperature rise. Current sharing in a parallel bank depends on branch impedance and layout as well as nominal capacitance.

Hold-up energy and capacitance

Hold-up is the ability to continue supplying a load while the DC bus falls after input power is interrupted. It is distinct from suppressing a switching-edge voltage spike. For approximately constant capacitance, the usable stored energy, ΔE, between initial voltage V₁ and minimum operating voltage V₂ follows Eq. (5).

\Delta E=\frac{1}{2}C\left(V_1^2-V_2^2\right) \tag{5}

For constant output power P_out, required hold-up time t_h and approximately constant downstream efficiency η, Eq. (6) gives the ideal minimum effective capacitance. V₁ must exceed V₂. These are actual bus-voltage limits, not automatically fixed fractions of the mains voltage.

C_{\min}=\frac{2P_{\mathrm{out}}t_h}{\eta\left(V_1^2-V_2^2\right)} \tag{6}

For example, a 1 kW output, 20 ms hold-up interval, 90% downstream efficiency and a bus falling from 400 V to 300 V require approximately 635 µF of effective capacitance. The capacitor supplies about 22.2 J while the load receives 20 J. These conditions illustrate the energy-storage calculation; they do not establish a universal DC-link value. [5]

Allow for tolerance, ageing, the lowest starting bus voltage and the actual converter undervoltage threshold. With strongly voltage-dependent capacitance, the constant-C energy expression is insufficient; use the capacitor’s voltage-dependent charge or energy data.

Thermal design and service life

The relevant thermal limit is not simply ambient temperature. Ripple-current losses create internal heating, while terminals and nearby power modules can conduct additional heat into the capacitor. Cooling arrangements, capacitor spacing and position within a bank influence the resulting temperature distribution. [1, 4]

Use the selected manufacturer’s thermal model and lifetime curves with the actual mission profile. A generic temperature-doubling rule is not a substitute for a technology-specific lifetime model. Voltage, humidity and mechanical stress can also affect reliability, and the applicable failure mechanism differs between electrolytic, film and ceramic technologies. [1]

A bank sized for ripple-current capability may contain more capacitance than the minimum obtained from a voltage-ripple calculation. Conversely, a bank selected for hold-up energy may still need additional current capability or local decoupling. Recheck both requirements after the thermal design is established.

Hybrid banks and anti-resonance

Combining bulk film or electrolytic capacitors with local ceramic capacitors can improve the available current path near a switching bridge. However, the interconnect inductance and the different capacitor branches can create anti-resonances, producing impedance peaks and circulating currents rather than uniform broadband improvement. [2]

Include the busbar, PCB and package parasitics in the model. Check impedance peaks against the switching frequency and relevant harmonics, and assess damping where required. Adding more capacitance is not always the correct remedy; reducing loop inductance or changing the distribution of capacitance may be more effective. [2]

Design and selection checklist

Use the following checklist to compare complete candidate solutions. Record the conditions behind each rating so that electrical, purchasing and component-engineering reviews use the same design basis.

Electrical envelope: minimum and maximum DC bus voltage, ripple, transients, current spectrum, overload duration and regenerative conditions.
Capacitance: minimum effective value after tolerance, temperature, DC bias and applicable ageing or end-of-life allowance.
Loss and impedance: ESR versus frequency and temperature, ESL, impedance curves, pulse-current limits and permissible voltage slew rate where specified.
Thermal conditions: internal temperature limit, ambient profile, cooling method, terminal heating and thermal coupling between parallel units.
Bank construction: branch current sharing, mechanical support, connection torque where applicable, insulation distances and vibration constraints.
Series connection: voltage sharing under steady-state and transient conditions, leakage-current variation and any required balancing network.
Qualification: verify the exact orderable part’s qualification, test conditions and documentation. Automotive component qualification does not replace validation of the assembled inverter or charger.
Safety: define precharge or inrush control, discharge provisions, residual-voltage verification, fault protection and access controls for the stored energy.
Catalog versus custom: use a catalog bank where its ratings and geometry meet the requirement. Consider a custom busbar-integrated assembly when connection inductance, packaging, cooling or mechanical interfaces justify the additional qualification and sourcing work.
Procurement: compare approved alternates at operating conditions, document permitted substitutions and retain drawings, models and lifetime assumptions for the released design.

Prototype validation and condition monitoring

Measure bus-voltage ripple, switching overshoot, capacitor current and representative temperatures on the assembled converter. Use appropriately rated measurement equipment and a probing arrangement whose parasitics do not dominate the result. Compare the measured spectrum and temperatures with the electrical and thermal models.

For service-critical equipment, capacitance or ESR trending can support condition monitoring. Interpret changes at comparable temperature and frequency, and use the manufacturer’s end-of-life criteria. A single case-temperature measurement or capacitance reading does not fully describe the internal condition of a bank. [1]

Capacitors for Demanding Inverter Designs

Recent capacitor developments matter when they change the achievable thermal conditions, current path or effective capacitance of the converter. For fast-switching designs, the capacitor and its mechanical integration should be treated as part of the power-stage design rather than as an isolated bulk component. [2, 4]

Metallized film and self-healing

Metallized film capacitors use a thin electrode deposited on a polymer dielectric. Under suitable conditions, a local dielectric breakdown causes nearby metallization to clear, isolating the damaged area. The capacitor can continue operating with a small loss of active area and capacitance. Repeated clearing, electrode corrosion and other degradation mechanisms can eventually reach an end-of-life criterion. [1]

Self-healing is not a guarantee against short circuits, excessive energy, thermal failure or unsuitable operating conditions. Nor should it be assumed for every film construction: metallized and film/foil designs have different characteristics. Pulse duty, permitted overvoltage and environmental limits must come from the selected part’s documentation.

Film capacitors can offer advantageous ripple-current capability and non-polar operation, but they are not universally smaller or longer-lived than electrolytics. Their advantage is strongest where high ripple current, low loss and suitable voltage capability determine the complete bank size. [1]

Wide-bandgap switching and distributed decoupling

SiC and GaN devices can produce rapid current transitions, making the inductance of the complete commutation loop important. A low-inductance capacitor specification alone cannot compensate for a long or poorly arranged external connection. Distributed capacitance places a smaller decoupling portion near the switching devices while retaining the main bulk energy buffer elsewhere. [2]

Recent power-module work also places ceramic decoupling inside or immediately adjacent to the module. This can shorten the current loop, but exposes the capacitor to demanding thermal and assembly conditions. Selection must account for dielectric behaviour, effective capacitance, current capability and mechanical stress, followed by an anti-resonance check of the complete network. [2]

Low-inductance film constructions remain relevant as well. Manufacturers offer DC-link film families intended for fast-switching SiC applications; their terminal geometry and current capability should be evaluated at bank level rather than excluded by an arbitrary switching-frequency cutoff. [3]

Higher-temperature polymer-film dielectrics

Higher-temperature polymer-film developments can change the trade-off between capacitor placement, cooling and bank size. One recent example uses a polypropylene and cyclic olefin copolymer blend in a modular DC-link capacitor. The design significance is increased thermal capability under the manufacturer’s specified conditions, not a general increase in the permissible temperature of all polypropylene capacitors. [4]

When comparing an established film solution with a newer dielectric, check the voltage-temperature operating envelope, ripple spectrum, hotspot limit, lifetime model, self-healing behaviour and mechanical interface. A higher temperature rating is useful only if the complete bank still meets its electrical and reliability requirements.

Conclusion

DC-Link capacitor selection is a combined energy, current, impedance and thermal design problem. Determine the minimum effective capacitance from the required buffering duty, then independently verify ripple-current capability, voltage stress, current sharing and service life.

Electrolytic, film and ceramic technologies each have a useful role. The best solution is the bank that meets the converter’s operating and safety requirements with a validated connection layout and thermal design, not necessarily the one with the highest nominal capacitance or the lowest catalog ESR.

FAQ

What is the function of a DC-Link capacitor?

A DC-Link capacitor buffers the instantaneous power difference between a source and a converter. It limits DC bus voltage variation and supplies local switching current. Depending on the application, it may also provide hold-up energy. It connects across the DC rails and does not require a rectifier to be present upstream.

Where are DC-Link capacitors used?

DC-Link capacitors are used in traction inverters, onboard chargers, photovoltaic and wind converters, industrial motor drives, UPS equipment and switched-mode power supplies. Their duty varies with the architecture. A battery-fed inverter may be dominated by switching-current requirements, while a mains-fed supply may also require substantial energy storage for power pulsation or hold-up.

Should a DC link use film, electrolytic or ceramic capacitors?

Choose the technology according to the dominant requirement. Aluminum electrolytics are often effective for bulk energy storage. Metallized film capacitors are strong candidates for high ripple-current duty. Ceramics can provide compact local decoupling, but their effective capacitance and mechanical limits require attention. A mixed bank can be useful if its current distribution and resonances are verified.

How is the required DC-Link capacitance determined?

Determine capacitance from the permitted bus-voltage change and the charge or energy that must be buffered. Use the actual capacitor-current waveform for ripple calculations and the initial and minimum bus voltages for hold-up calculations. Then apply tolerance and operating-condition allowances. Separately verify ripple-current capability, since the calculated capacitance may not provide sufficient thermal capacity.

Why are DC-Link capacitors important in electric vehicles?

In a traction inverter, DC-Link capacitors provide a local path for switching current and help limit bus-voltage excursions associated with interconnect inductance. In an onboard charger, they buffer energy between conversion functions. The two applications do not necessarily require the same bank. Voltage range, current spectrum, thermal conditions and regenerative operating requirements must be checked separately.

Can ceramic capacitors be added in parallel with a film DC-link bank?

Yes, suitably selected ceramics can supply high-frequency current close to the switching devices while the film bank provides bulk buffering. However, the combination introduces additional resonant behaviour through the interconnect inductance. Check effective capacitance under bias, branch currents, thermal limits and anti-resonance peaks before assuming that more parallel capacitance will improve the design.

How to Select the Right DC-Link Capacitor

Select and validate a DC-link capacitor bank against its electrical, thermal and safety requirements.

Define the converter requirements
Record the topology, operating modes, bus-voltage range, allowable ripple, hold-up requirement, environmental conditions and service-life target.
Calculate the minimum effective capacitance
Use charge balance for voltage ripple and available stored energy for hold-up. Include tolerance, temperature, bias and applicable ageing allowances.
Evaluate current and losses
Obtain the capacitor-current waveform and frequency spectrum. Check ripple-current ratings and estimate losses using ESR at the relevant frequencies and temperatures.
Select the technology and bank layout
Compare electrolytic, film, ceramic and mixed solutions. Include connection inductance, current sharing and anti-resonance in the assembled-bank assessment.
Verify thermal life and safety
Check hotspot temperature and lifetime across the mission profile. Confirm qualification documentation, mechanical requirements, precharge, discharge and fault protection.
Validate the prototype and release the specification
Measure ripple, overshoot, capacitor current and temperatures on the assembled converter. Reconcile the results with the models and document approved parts and substitution limits.

References

The following documents support the operating principles, selection guidance and recent design developments discussed in the article.

H. Wang and F. Blaabjerg, Reliability of Capacitors for DC-Link Applications in Power Electronic Converters: An Overview, IEEE Transactions on Industry Applications, volume 50, issue 5, pages 3569–3578, 2014.
TDK Electronics, CeraLink’s Benefit in Power Module Applications, TDK Electronics, 2025.
TDK Electronics, TDK Offers Robust Capacitors with Low Self-Inductance and High Design Flexibility, TDK Electronics, 2022.
TDK Electronics, The New ModCap Likes It Hot, TDK Electronics, 2026.
Passive Components Blog, Film Capacitors Characteristics and Uses in Power Applications, Passive Components Blog, 2022. Linked in Further Reading.

Latest technical review and revision: 3 November 2026.

Source: KEMET