Center Spreading
Understanding Center Spreading
EMI Reduction through Spectral Energy Dispersion
High-speed digital circuits rely on stable clock signals to coordinate data transfers. These clocks are typically square waves with fast rise and fall times, containing high energy at the fundamental frequency and its odd harmonics. These discrete energy peaks can radiate from PCB traces, acting as unintentional antennas and causing the system to fail electromagnetic compatibility (EMC) tests. Center spreading resolves this challenge by applying frequency modulation (FM) to the clock source. Instead of operating at a fixed frequency, the clock frequency is continuously swept back and forth around the nominal value, dispersing the spectral energy over a wider band.
Because EMI receivers measure power within a specific resolution bandwidth (RBW), spreading the clock signal reduces the peak power measured at any single frequency coordinate. A typical spread-spectrum clock generator (SSCG) uses a triangular modulation profile (often called the Lexmark profile) at a modulation frequency of 30 kHz to 33 kHz (above the audio range but slow enough for digital phase-locked loops to track). This technique can lower peak radiated emissions by 10 dB to 18 dB at higher harmonics, avoiding the need for expensive physical shielding or filtering.
Comparison: Center Spreading vs. Down Spreading
Spread spectrum clocking is classified by how the frequency modulation is applied relative to the nominal clock frequency. In center spreading, the frequency is modulated symmetrically above and below the nominal center (e.g., a +/-1.0% spread on a 100 MHz clock sweeps the frequency between 99 MHz and 101 MHz). In down spreading, the frequency is modulated only below the nominal value (e.g., a -2.0% spread sweeps between 98 MHz and 100 MHz). Down spreading is preferred for microprocessors and memory controllers to prevent timing violations caused by running components above their maximum rated clock frequency, while center spreading is used in systems where the nominal frequency represents the target average speed.
Key Mathematical Relations
Technical Specifications Comparison
| Modulation Type | Frequency Range Profile | Peak EMI Reduction | Jitter & Timing Impact | Typical Applications |
|---|---|---|---|---|
| No Spreading (Fixed Clock) | Fixed at nominal (e.g., 100.00 MHz) | 0 dB (Reference baseline) | Ultra-low jitter (Clean phase noise) | High-speed RF synthesizers, ADCs, precision timing |
| Center Spreading | Symmetrical (e.g., 99.0 to 101.0 MHz) | 10 - 15 dB (High reduction) | Moderate clock jitter, potential overclocking stress | Display interfaces (HDMI, DisplayPort), consumer electronics |
| Down Spreading | Asymmetrical (e.g., 98.0 to 100.0 MHz) | 8 - 12 dB | Moderate clock jitter, protects maximum timing margins | High-speed digital buses (PCIe, SATA, DDR memory) |
Frequently Asked Questions
What is the difference between center spreading and down spreading in SSCG?
Center spreading modulates the clock frequency symmetrically above and below the nominal target frequency. Down spreading modulates the clock frequency exclusively below the nominal target, preventing timing errors in components that cannot exceed their maximum rated clock frequency.
How does center spreading affect clock jitter in high-speed digital designs?
By continuously varying the clock period, center spreading introduces cycle-to-cycle and period jitter. In high-speed serial links like PCIe, this jitter can degrade the eye diagram and increase bit error rates, requiring careful clock configuration.
Why does frequency modulation reduce peak radiated emissions in compliance testing?
Compliance receivers measure radiated emissions using a fixed resolution bandwidth filter. Spreading the clock frequency distributes the signal's energy across a wider bandwidth, meaning less energy is present within the receiver's measurement window at any single moment.