Modern Buck Converter Features for Engineers

In the March 2023 Issue of IEEE Journal of Solid-State Circuits (JSSC) 🏷️https://lnkd.in/gqfy3H_R, researchers from UC San Diego and University of Colorado Boulder reported a step-down (buck) DC-DC converter fabricated in 130nm Bipolar-CMOS-DMOS (BCD) technology. This “Integrated Transformer-less Stacked Active Bridge (ITSAB)” voltage converter requires only a couple of 10nH IPD inductors, while operating at a modest switching frequency (2-5MHz). The measured peak efficiency is at 91.2% in stepping from 9.6~12V down to 2.15~3.3V. With two #IPD inductors, this converter achieves a peak power density of 1.36W/mm3. Excerpts (edited): 📝Though active devices benefit from transistor scaling, passive devices, particularly inductors, do not enjoy the same rate of miniaturization. Consequently, inductors dominate the size of modern converters. 📝With the help of capacitors, inductors can be magnetized with much lower voltages such that smaller inductors can be used. But most of the previously proposed hybrid (i.e., LC) converters are based on discrete implementations, thereby requiring inductances in the range of ≥100nH. 📝The ITSAB converter in this work strives to achieve a superior power density with nH-scale inductors, while operating at switching frequencies in the MHz range. The use of IPDs (along with an advanced-packaging effort) helps further increase the power density. 📝The operation of capacitor charge transfer in an ITSAB converter resembles that of a Dickson-star SC converter. The key difference is in that, the charge transfer between flying capacitors (C1-C3) and the output in a Dickson-star SC converter is of the hard-charging type, while in an ITSAB converter it's instead a soft-charging operation enabled by IPD inductors (L1 and L3). 📝The ITSAB converter exhibits an effort to overcome challenges in other converter types that require much larger inductors (conventional hybrid converter), have difficulty in achieving fine regulation together with high efficiency and compact size (resonant converter), or suffer from lower efficiency and limited regulation (SC converter). 🔍Observations: With the reported input/output profile (i.e., 9.6~12V/2.15~3.3V) and form factor, this ITSAB converter design is applicable to power-management operations for consumer (e.g., portable battery) and automotive (e.g., infotainment). It may be worthwhile to explore the FOWLP solution as an alternative for optimizing IPD integration. In a perfect world, an ideal converter would be not only (input) voltage- but also (load) current-agnostic, while achieving a ~100% efficiency. In reality, optimizations of PWM algorithms are carried out on a case-by-case basis. Further reading: 🏷️Full JSSC article: https://lnkd.in/g5V5gdJm 🏷️400V-to-48V GaN: https://lnkd.in/gykzvAQs 🏷️48V-to-12V MOSFET: https://lnkd.in/gKqYXGaH FOWLP: Fan-Out Wafer-Level Package IPD: Integrated Passive Device PWM: Pulse-Width Modulation SC: Switched-Capacitor ➟ To be continued.

Can GaN Really Shrink the Capacitor Bank? Past experience with power electronics engineers and their design edicts / challenges, one truth has always held: passives dominate board real estate in buck converter designs. The question I keep hearing now is—can Gallium Nitride finally change that equation? Here’s my perspective: Where GaN delivers: • Higher switching frequencies → less output capacitance for ripple. • Faster transient response → less energy buffer needed. • Multiphase at high efficiency → ripple cancellation, smaller cap banks. Where GaN doesn’t move the needle: • Hold-up and ride-through (you can’t cheat the physics of energy storage). • Surviving harsh input transients. • Ultra-low-frequency DC stability on sensitive rails. My take? GaN isn’t just a faster switch—it’s an enabler to redesign the power stage. If you simply swap Si for GaN, nothing changes. If you re-architect around higher frequency, faster loops, and multiphase, then yes—you can genuinely reduce those bulky MLCC arrays and polymers. But that’s just my view from the field. I’d love to hear yours: Have you successfully shrunk capacitor banks by moving to GaN? What trade-offs (EMI, thermals, cost) did you encounter? Do you see GaN as a capacitor reducer or just a silicon replacement? #Engineers #PowerElectronics #GalliumNitride #GaN #BuckConverter #DCDC