Asymmetric Edge Propagation in High Frequency Systems

14 concepts12 papers2 proposals

In elementary circuit design, signals are treated as clean, binary abstractions: a low state becomes a high state, and a high state becomes a low state. We naturally assume that the time it takes for a signal to rise equals the time it takes to fall.

However, as modern electronic systems push into ultra-high frequencies—such as megahertz-range power converters or picosecond-precision timing circuits—this assumption breaks down. In the real world, rising and falling signal edges travel through physical hardware at different velocities, a phenomenon known as asymmetric edge propagation. Left unaddressed, this asymmetry distorts duty cycles, shrinks pulses until they vanish, and degrades energy efficiency. To overcome these bottlenecks, researchers are moving away from symmetric design paradigms and are instead building architectures that natively adapt to edge asymmetry.

The Physical Roots of Edge Asymmetry

Signal propagation asymmetry stems from both semiconductor physics and circuit topology. At the silicon level, the fundamental building blocks of CMOS circuits—PMOS and NMOS transistors—rely on charge carriers (holes and electrons) with inherently different mobilities. As Huawei Li et al. demonstrate in Testable Critical Path Selection Considering Process Variation, failing to distinguish between rising and falling transitions in timing models leads to inaccurate worst-case delay predictions, as manufacturing variations exacerbate these microscopic differences.

Symmetric Ideal:
Input:  └───┘      (Equal rise/fall times)
Output: └───┘

Asymmetric Reality:
Input:  └───┘      (High-side level shifter adds delay to rising edge)
Output:   └───┘    (Pulse width distortion / Duty cycle shift)

At the circuit topology level, asymmetry is frequently introduced by voltage-domain interfaces. In half-bridge power converters, the high-side power switch floats at a high voltage, meaning its control signal must pass through a level shifter to bridge the gap from the low-voltage controller. This extra circuitry introduces a severe propagation mismatch: the high-side turn-on path becomes inherently slower than the low-side path.

As highlighted by the authors of An Active Dead-Time Control Circuit With Timing Elements for a 45-V Input 1-MHz Half-Bridge Converter, this structural imbalance means conventional, uniform delay chains cannot safely optimize the circuit. Furthermore, Qiang Li et al. note in A Novel Floating High-Voltage Level Shifter with Pre-Storage Technique that standard level shifters introduce nanosecond-level propagation delays that differ significantly between rising and falling transitions, directly causing distorted duty cycles.

The High-Frequency Bottleneck

Why has this asymmetry recently become a critical focus? The answer lies in the rise of wide-bandgap (WBG) semiconductors like Gallium Nitride (GaN) and Silicon Carbide (SiC). As explored by Mitiko Miura-Mattausch et al. in Simulation-Based Power-Loss Optimization of General-Purpose High-Voltage SiC MOSFET Circuit Under High-Frequency Operation and Hang Dai in Current source inverter with bi-directional switches, WBG devices allow power electronics to operate at hundreds of kilohertz or even megahertz while maintaining efficiencies above 98%. To achieve this, gate drivers must deliver exceptionally high peak currents and rapid voltage slew rates, as noted by Dong Jiang et al. in IEEE PELS Wuhan Chapter Celebrates PELS Day.

At these extreme switching speeds, a nanosecond-level propagation mismatch is no longer negligible—it represents a substantial fraction of the entire switching period. Dejin Zhou et al. demonstrate in A 20MHz 4A gate driver with 5.5 to 24V output drive voltage for wide bandgap FETs that propagation asymmetries between high-side and low-side driving paths cause severe timing errors. If the high-side switch turns on before the low-side switch has fully turned off, a catastrophic short-circuit (shoot-through) occurs.

This bottleneck extends well beyond power electronics into high-speed signal processing:

Embracing Asymmetry: How Solutions Work

To overcome these limitations, modern architectures do not try to force perfect symmetry onto inherently asymmetric hardware. Instead, they embrace the asymmetry by decoupling the control paths of rising and falling edges.

Asymmetric Dead-Time Optimization

In a standard half-bridge inverter, a brief delay called "dead-time" is inserted between switching transitions to prevent shoot-through. However, dead-time causes a discrepancy between commanded and actual output voltages, introducing control inaccuracies (Jie Shi in Analysis and Compensation Control of Dead-Time Effect on Space Vector PWM).

Traditional designs minimize low-to-high and high-to-low dead-times equally. However, the authors of A 9.2-ns to 1-s Digitally Controlled Multituned Deadtime Optimization for Efficient GaN HEMT Power Converters reveal that the two transitions have completely different physical requirements:

  1. Low-to-High Transition: Must be minimized strictly to prevent short-circuits.
  2. High-to-Low Transition: Requires a specific, load-dependent delay that allows the parasitic node capacitance to perfectly discharge to zero volts (achieving Zero Voltage Switching, or ZVS).
Symmetric Tuning (Inefficient):
Low-to-High Delay:  [─── Fixed Minimum ───]
High-to-Low Delay:  [─── Fixed Minimum ───] -> Misses ZVS discharge window

Asymmetric Tuning (Optimized):
Low-to-High Delay:  [─── Fixed Minimum ───] -> Prevents shoot-through
High-to-Low Delay:  [─────── Load-Dependent Tuned Delay ───────] -> Achieves ZVS

By treating these delays asymmetrically, converters can eliminate unnecessary power losses. To implement this natively, the authors of An Active Dead-Time Control Circuit With Timing Elements for a 45-V Input 1-MHz Half-Bridge Converter split the driver's control circuit into independent left and right architectural branches. This structural separation allows the circuit to generate distinct, nanosecond-level delays tailored precisely to the unique propagation profile of each switch.

Hardware-Level Balancing

Another strategy focuses on redesigning the underlying circuit components to minimize the propagation gap before it compounds. In A Novel Floating High-Voltage Level Shifter with Pre-Storage Technique, Qiang Li et al. introduce a "pseudosymmetrical" input pair combined with a "pre-storage" technique. This architecture pre-charges internal nodes to accelerate the slower transition path, successfully compressing the propagation delay down to the sub-nanosecond range and restoring duty-cycle symmetry.

In the digital domain, rather than redesigning the silicon, researchers use clever encoding schemes to bypass physical limitations. For instance, Wujun Xie et al. counter pulse-vanishing effects in TDCs by implementing a bidirectional encoder alongside manual pulse-width configurations. By tracking transitions from both directions and actively adjusting the initial pulse widths, they ensure that signal transitions survive long delay lines, preserving sub-picosecond timing accuracy across the entire measurement range.

Key Foundational Work

A novel auxiliary-free zero inductor current detection scheme for step down non-isolated LED driver (2014) — contributed 2 concept elements to this direction.
Testable Critical Path Selection Considering Process Variation (2010) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_asymmetric_edge_propagation_d35ad537,
  title = {asymmetric edge propagation},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/asymmetric-edge-propagation/rdr_d35ad53737dd42919a82202a4071643a},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_asymmetric_edge_propagation_d35ad537_v1,
  title = {asymmetric edge propagation},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/asymmetric-edge-propagation/rdr_d35ad53737dd42919a82202a4071643a/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake