Stacked Ceramic Capacitors’ Impact on Switch Mode Power Supplies

­Not all modern electronics require increased power density. But it’s safe to say that getting more power out of power supplies with equivalent or smaller weights and volumes is of near universal interest in the design world, regardless.

Switch Mode Power Supplies

Switch mode power supplies (SMPS) are one way design engineers can deliver higher power densities. SMPS utilize low-loss switching transistors, which enables high conversion efficiency. As such, they dissipate less energy than linear regulators, which continuously dissipate energy. This efficiency advantage enables SMPS weight reductions, because less transistor heating translates into smaller heatsink needs. Since SMPS switch at such high frequencies, they also use smaller, lighter-weight, and lower-value inductors. Additionally, as SMPS continue to achieve ever-higher switching frequencies, engineers can further reduce the values and sizes of the passive components (e.g., inductors and capacitors) utilized in these designs to help shrink the size and weight of the overall form factor while delivering more power than competing solutions, like linear power supplies. Linear power supplies vary greatly in design in application, so there are outliers, but typical efficiency for the product category as a whole hovers around 60%. SMPS, on the other hand, tend to have typical efficiencies starting around 70%, and many offer efficiencies of 90% or more.

However, despite all these clear advantages, SMPS also have their share of cons — particularly a complex set of design requirements and, if left unchecked, disruptive electromagnetic interference (EMI). Luckily, newly improved, higher-performance versions of several different capacitor technologies can help alleviate these few SMPS disadvantages to satisfy higher power density demands.

Capacitors and SMPS

Capacitors play many crucial roles in SMPS, including input/output filtering, integrating voltage spikes on transistors and diodes, control loop stabilization, and decoupling/filtering power at switching transistors. Of these, the output filter capacitor role is the most impactful. So, let’s dive deeper into SMPS output capacitors.

Equivalent series resistance (ESR) and inductance (ESL) have historically been acceptable for primary output filter capacitors in SMPS, even though they both influence output ripple voltage. To meet target ESR values, design engineers typically just connected multiple capacitors in parallel to reduce the overall ESR of the bulk capacitance. This resulted in the total capacitance exceeding the minimum requirements by 10 times or more. But while reduced ESR and ESL at the cost of excess capacitance wasn’t an electrical disadvantage, connecting several to many parallel capacitors while maintaining a small volume presented a real packaging challenge — especially since they could be the size of a mini soda can.

Fast forward to the new and improved high-frequency SMPS devices. As switching frequencies progressed from 20kHz in the early days of SMPS to 1mHz and higher, capacitor ESL became a primary parameter of concern. The ideal bulk capacitors for these SMPS exhibit high capacitance density, high reliability, and low ESR and ESL and are quick and easy to install.

Typical choices for SMPS output capacitors include aluminum electrolytic capacitors, stacked multilayer ceramic capacitors (MLCCs), and tantalum or tantalum polymer capacitors — the first two of which represent the dominant technologies.

Aluminum Electrolytics

Aluminum electrolytic capacitors have a well established history of successful use in SMPS. These capacitors offer high capacitance — up to thousands of microfarads (µF), moderate ESR, and an acceptably moderate lifespan and are available in three variants that correspond to the electrolyte used in manufacturing. The three types of aluminum electrolytic capacitors are wet, polymer, and hybrid.

Wet aluminum electrolytic capacitorshave a liquid-electrolyte-soaked separation paper between the anode foil and cathode foil. This design approach offers the lowest cost solution, but at the expense of relatively high ESR, useful lifetimes truncated by evaporation-based failures, and poor low temperature performance given that the liquid electrolyte can freeze and fail.

Polymer aluminum electrolytic capacitorsemploy a solid electrolyte instead of a liquid electrolyte. But, otherwise, their construction is essential the same as wet aluminum electrolytic capacitors. The only difference is that the soaked separator paper is replaced with a solid polymer layer. However, compared to wet aluminum electrolytic capacitors, aluminum electrolytic capacitors exhibit substantially reduced ESR, approximately five to 10 times higher reliability, and better low-temperature performance. Their disadvantages are higher costs and a reduced capacitance range.

Hybrid aluminum electrolytic capacitors — a newer development — combine a liquid electrolyte with a conductive polymer layer to overcome the capacitance range limitations of aluminum polymers and the reliability and ESR limitations of wet electrolytic aluminum capacitors. And while they succeed, they’re the most expensive aluminum electrolytic capacitor technology.

Click image to enlarge

Figure 2: Performance benefits and considerations for wet, polymer. and hybrid aluminum electrolytic capacitors

Regardless of the exact technology used, aluminum electrolytic capacitors exhibit relatively high inductance compared to stacked MLCCs, which impacts their switching frequency range and limited their use in the highest frequency SMPS.

Stacked MLCCs

MLCCs have multiple performance advantages in high-frequency output filter applications, the top three of which are low ESL, low ESR, and high reliability.

Stacked MLCCs were initially developed for use in military SMPS circuits, but their use has since expanded to the industrial, enterprise, and communication sectors and beyond — wherever the high-efficiency power conversion benefit outweighs the higher costs of SMPS designs. The technological improvements behind this expansion include higher volumetric efficiency — higher capacitance values in the same or smaller packages and stacks, new stack configurations with the electrodes perpendicular to the PCB, which reduces stack inductance, and new stacked capacitor sizes that increased the number of available solutions. 

Stacked capacitors are commonly available in versions with the MLCCs stacked parallel to the PCB or at a 90° angle to the PCB, as shown in Figure 3.

Click image to enlarge           

Figure 3: A horizontal stack typical for MIL-PRF-49470 applications (left), a vertical stack for COTS and industrial applications (center), and a two-chip horizontal stack for transportation and industrial applications

When you compare the ESR of wet aluminum electrolytic capacitors to that of stacked MLCCs over a range of common SMPS switching frequencies, the MLCCs exhibit much lower ESR across the RF spectrum, which directly corresponds to reduced output ripple voltage across frequency.

Click image to enlarge

Figure 4: ESR values for horizontally stacked MLCCs and wet aluminum electrolytic capacitors over a range of common SMPS switching frequencies

This performance advantage is also illustrated in the following real-world comparison of 2200µF wet aluminum electrolytic capacitors and a 200µF MLCC. This example also shows the effect of 10 220µF wet aluminum electrolytic capacitors in parallel and one 200µF stacked MLCC. And though the parallel wet aluminum electrolytics can be used successfully, the stacked MLCCs deliver better performance in a smaller, lighter package ideal for helping design engineers achieve higher power densities in smaller, lighter SMPS devices.

Click image to enlarge

Figure 5: This graph compares the output ripple voltage of multiple 220µF wet aluminum electrolytic capacitors in parallel to that of a single 200µF stacked MLCC, the latter of which can help design engineers deliver higher power densities in smaller, lighter SMPS

Summary

Although both aluminum electrolytic capacitors and stacked MLCCs — the two most prevalent SMPS output capacitor technologies — have evolved in terms of structure and performance in recent years, testing has proven that stacked MLCCs reliably outperform wet aluminum electrolytics in SMPS, as well as enable smaller, lighter packages critical to the higher power density designs needed to satisfy customer demands.

Additional testing is underway to compare stacked MLCCs to polymer and hybrid aluminum electrolytic capacitors. In these cases, stacked MLCCs won’t have quite as large of an advantage in terms of ESR. However, stacked MLCCs are still expected to exhibit four or five times better ESR performance and five to 10 times better ESL performance, depending on the case sizes, capacitance values, and electrolytic technologies of the aluminum electrolytic capacitors.

Further, both stacked MLCC and aluminum electrolytic capacitor technologies are still rapidly evolving. Next-generation aluminum electrolytic capacitors are being designed with dramatically lower inductance than existing technology, and reliability is improving through improved material system purity and manufacturing innovation. Stacked MLCC advances are being driven by new and optimized ceramic formulations, new stack configurations and case sizes, and alternative terminations designed to further reduce ESR and inductance values.So, their performance will continue to improve across the board, giving design engineers even better output capacitor options for developing small, lightweight, and high-power-density SMPS.

KYOCERA AVX Corporation