--- title: "What does spread spectrum do to a buck’s EMI test?" topic: "Spread spectrum" answer: "Almost nothing at its 400 kHz clock. It takes 10 dB off the 4 MHz reading and 12 dB off 30 MHz." url: "https://nearfieldnotes.com/notes/spread-spectrum/" site: "Near Field Notes" author: "Michael Walker" note: 6 family: "EMC" published: "2026-11-01" revised: "2026-11-01" revision: 1 video: "https://nearfieldnotes.com/videos/spread-spectrum/" --- Note 6 · [EMC](https://nearfieldnotes.com/families/emc/) · Rev 1 · [Watch the video · 1:19](https://nearfieldnotes.com/videos/spread-spectrum/) Spread spectrum # What does spread spectrum do to a buck’s EMI test? **Short answer:** Almost nothing at its 400 kHz clock. It takes 10 dB off the 4 MHz reading and 12 dB off 30 MHz. **Note 6’s opening figure: The spread, harmonic by harmonic.** 220 kHz either side of each ![Note 6’s opening figure: The spread, harmonic by harmonic](https://nearfieldnotes.com/notes/spread-spectrum/fig-opening.png) *What the figure shows:* The buck's conducted noise, 220 kHz either side of three harmonics, with spread spectrum off and on. At 400 kHz, the clock, the spread reaches ±16.9 kHz and the quasi-peak reading drops 1 dB. At 4 MHz, the 10th harmonic, the spread reaches ±169 kHz and the quasi-peak reading drops 10 dB. At 30 MHz, the 75th harmonic, the spread reaches ±1.3 MHz and the quasi-peak reading drops 12 dB. **Key** - Spread off - Spread on The clock moves up to 16.9 kHz either way. Each harmonic moves that much times its number. ## The buck and its test *MNT Reform’s 5 V rail* The MNT Reform laptop makes its 5 V with an LM62460 [buck](https://nearfieldnotes.com/glossary/#g-buck). It pulls its input current from the bus in pulses, 400,000 a second, and that current is the noise a conducted EMI test reads. With spread spectrum on, the chip moves its clock a little every cycle. **Fig. 6: The circuit.** ![Fig. 6: The circuit](https://nearfieldnotes.com/notes/spread-spectrum/fig-6.png) *What the figure shows:* Figure 6: the test circuit. The LM62460 buck switches at 400 kHz from a 25.6 V bus and makes 5 V at 2 A. Its input current comes from the bus capacitors, four 10 µF, four 470 nF and one 2.2 µF, and from a LISN in place of the battery: 50 µH with 5 Ω, beside the receiver's 50 Ω. The LISN stands in for the battery and the rest of the laptop. The receiver reads its 50 Ω. **Fig. 6b: The clock.** ![Fig. 6b: The clock](https://nearfieldnotes.com/notes/spread-spectrum/fig-6b.png) *What the figure shows:* Figure 6b: the switching frequency, cycle by cycle, for 302 µs. A triangle of ±11.4 kHz every 74 µs, plus a random jump each cycle of up to ±5.5 kHz, about 400 kHz. TI’s pattern for the LM62460, measured from their scope shot: a 74 µs triangle of ±11.4 kHz plus a jump every cycle, up to ±5.5 kHz. **Fig. 6c: Three harmonics.** ![Fig. 6c: Three harmonics](https://nearfieldnotes.com/notes/spread-spectrum/fig-6c.png) *What the figure shows:* Figure 6c: three windows, each 220 kHz either side of a harmonic, spread off and on. 400 kHz, the clock: the quasi-peak reading drops 1 dB. 4 MHz, the 10th harmonic: the quasi-peak reading drops 10 dB. 30 MHz, the 75th harmonic: the quasi-peak reading drops 12 dB. Quasi-peak reading, 220 kHz either side of each harmonic. Spread spectrum ### The 4 MHz harmonic drops 10 dB, the clock only 1 **Field marks.** Specimen 2 is the 4 MHz harmonic. Its reading drops 10 dB. **Look-alike.** Specimen 1, the 400 kHz clock, moves only 16.9 kHz either way. Its reading drops 1 dB. **Range.** `Δf_n = n · Δf = 10 × 16.9 kHz = 169 kHz`. Past about 5 MHz each harmonic’s spread reaches its neighbor’s, and the drop stops growing. Specimen 3, at 30 MHz, drops 12 dB. **Habitat.** The MNT Reform laptop’s 5 V buck: an LM62460 at 400 kHz, 25.6 V in, 2 A out. Read through a 50 µH LISN and a CISPR 16 [quasi-peak](https://nearfieldnotes.com/glossary/#g-quasi-peak) receiver. ## Read the whole band *Hover for values Tap or drag for values* **Note 6’s explorer: Spread spectrum.** What a conducted EMI test reads, 150 kHz to 30 MHz ![Note 6’s explorer: Spread spectrum](https://nearfieldnotes.com/notes/spread-spectrum/fig-explore.png) *What the figure shows:* What the receiver reads from 0.15 to 30 MHz with spread spectrum off and on, for the chosen pattern and detector. 9 kHz window, 50 µH LISN at the bus. No limit line: the LISN stands in for the rest of the laptop, so these levels are this buck’s alone. **Figure: Peak, quasi-peak and average.** How much each detector’s reading drops with spread on Fig. 6d | Harmonic | Peak | Quasi-peak | Average | |---|---|---|---| | 400 kHz | −0.7 dB | −1.1 dB | −1.2 dB | | 4 MHz | −7.4 dB | −10.2 dB | −14.5 dB | | 30 MHz | −8.0 dB | −12.4 dB | −18.4 dB | Detectors ### Average readings drop most, peak readings least **Field marks.** The jumps fill the gaps between harmonics with a low floor. At 4 MHz they leave the peak reading 2 dB higher than the triangle alone, and take another 5 dB off the average. **Range.** Conducted limits are set against quasi-peak and average, so those are the columns that pass or fail a test. ## Every harmonic *Quasi-peak, 400 kHz to 30 MHz* **Fig. 6e: The drop at each harmonic.** ![Fig. 6e: The drop at each harmonic](https://nearfieldnotes.com/notes/spread-spectrum/fig-6e.png) *What the figure shows:* Figure 6e: how much spread spectrum lowers the quasi-peak reading at each harmonic, 400 kHz to 30 MHz, for TI's pattern and the triangle alone. TI's pattern takes 1 dB off the clock, 10 dB off 4 MHz, and between 10 and 13 dB off every harmonic from 6 MHz up. How much lower each harmonic reads with spread on, quasi-peak. Key to spread spectrum ### How much does spread spectrum take off each reading? 1. 1a The 400 kHz clock itself 1 dB lower 2. 1b A harmonic Go to 2 3. 2a Below 6 MHz 10 dB lower at 4 MHz 4. 2b 6 MHz and up 10 to 13 dB lower At 30 MHz it reads 12 dB lower. #### Conditions - **Converter:** LM62460, MNT Reform motherboard 3.0, U7 - **Input:** 25.6 V, 8S LiFePO4 pack - **Output:** 5.0 V at 2 A - **Switching:** 400 kHz, spread spectrum on - **Spread:** A 74 µs triangle of ±11.4 kHz plus a jump each cycle within ±5.5 kHz, up to ±16.9 kHz in all - **Inductor:** 3.3 µH, PM2205.332NLT - **Network:** CISPR 16 50 µH / 50 Ω LISN at the bus - **Receiver:** 9 kHz window, quasi-peak, 1 ms charge, 160 ms discharge - **Reading:** The highest within 200 kHz of each harmonic TI publishes the spread's size but not its shape. The shape was measured from TI’s scope shot, datasheet SNVSBZ4A, Figure 8-11. #### Limits - These levels are this buck’s alone, not a test of the Reform. - Below about 0.24 A of load the chip slows its clock, and the spread stops with it. - Other random seeds and another capacitor model keep the drops within 0.6 dB. #### Files - ssim/ti_trace.py The pattern, measured from TI’s figure - ssim/emi.py The input current and the receiver #### Sources - [TI LM62460 datasheet, SNVSBZ4A](https://www.ti.com/lit/ds/symlink/lm62460.pdf) - [MNT Reform](https://mntre.com) What should the next note be? [Ask for one](https://nearfieldnotes.com/#ask-for-a-note) --- Source: https://nearfieldnotes.com/notes/spread-spectrum/ Cite as: Michael Walker, “Spread spectrum”, Near Field Notes, Note 6, Rev 1, 1 Nov 2026. https://nearfieldnotes.com/notes/spread-spectrum/ Figures: free to reuse under CC BY 4.0 (https://nearfieldnotes.com/license/). Each one's image is linked under its name. Credit Michael Walker, Near Field Notes, and link the figure: this address plus #fig- and its number, like https://nearfieldnotes.com/notes/spread-spectrum/#fig-6.