Before any electronic product can be launched to market, it needs to achieve compliance certification from appropriate regulatory agencies such as the Federal Communications Commission (FCC) and The International Special Committee for Radio Interference (CISPR) - French acronym for ‘Comité International Spécial des Perturbations Radioélectriques’ ,to ensure it does not exceed peak radiated and conducted EMI emission limits. This is an ongoing challenge that electronic system designers must address adequately at an early stage in the design cycle, or it could adversely impact the product launch schedule. The emphasis on faster clock speeds and smaller form factors in both consumer and industrial electronic products only serves to heighten the challenge.
Fortunately, there are now a number of tools, processes and devices that system designers have at their disposal to help address EMI emissions. These include approaches such as shielding, sound PCB design practices, spread spectrum clocking and passive components including filters, ferrite beads and common mode chokes.
This article focuses on how Spread Spectrum Clock Generators can be employed to reduce peak EMI and help achieve regulatory compliance.
What is EMI?
Electromagnetic Interference (EMI) is unwanted emissions caused by varying electric and magnetic fields and transmitted by conduction or radiation. The amount of EMI generated is directly proportional to the rate at which the electric and magnetic fields vary. The worst offenders of EMI are periodic signals such as clocks, since most of the energy peaks are concentrated at the clock frequencies.
Why is it important to contain EMI?
EMI generated by one electronic product can interfere with the normal frequency of operation of another electronic product. Typical examples are phone signals interfering with TV operation resulting in distorted images, phone signals interfering with conference call equipment resulting in poor call quality, interference in Bluetooth devices, baby monitors and cordless phones. To address these problems, regulatory agencies specify emission limits that electronic products must comply with before they are launched to market.
Spread Spectrum Clock Generation
Spread Spectrum Clock Generation (SSCG) is a technique that modulates the frequency of a clock source in a controlled fashion. The result is reduced peak spectral energy at the fundamental and harmonic frequencies achieved by spreading the energy contained in the narrow band of the clock source over a wider band. Since SSCG reduces EMI at the clock source, all downstream clock and data signals derived from this clock source will also be spread if the bandwidth of the downstream system components allows it. This provides system wide EMI suppression thereby reducing the cost and number of components needed when compared to alternative approaches. As shown in Fig. 1, EMI suppression is seen both at the fundamental and harmonic frequencies.
(*See Figure 1)
The amount of peak EMI suppression achieved depends on three variables:
modulation rate, modulation depth (or spread) and the type of modulation profile used.
Modulation rate is the rate at which the nominal clock is varied over the modulation depth. In Fig.2, this is represented by ‘MR’.
Modulation depth (or spread) is the total amount of variation of the nominal clock frequency. For example, and as shown in Fig.2, a 100MHz nominal clock with +/-1% deviation indicates that the frequency of this 100MHz clock varies from 99MHz to 101MHz. This type of spread is called center-spread because the average frequency remains centered at 100MHz. Another commonly used spread scheme is down-spread. A 100MHz clock that is -1% down-spread indicates that the clock frequency varies between 100MHz and 99MHz. The average frequency in this case would be 99.5MHz.
Modulation profile determines the shape of the spectral energy re-distribution. There are a number of modulation profiles: Sine, triangle (linear), Lexmark (non-linear) etc.
(*See Figure 2)
It can be seen in Fig. 3 that the Lexmark profile has a flat spectral energy re-distribution where the side-lobes are flat compared to the center. The sine profile has side-lobes that are a few dB higher than the center because more spectral energy is concentrated at the side-lobes.
(*See Figure 3)
More EMI suppression can usually be achieved by increasing the amount of spread, provided the modulation rate and modulation profile are the same. ON Semiconductor offers a wide range of EMI suppression devices in various packages, supply voltages, spread profiles, spread percentages, frequency ranges and modulation rates to help achieve EMI compliance for a wide array of applications. Many of these devices have digital or analog pins to adjust the amount of spread, thereby providing flexibility to system designers in choosing the right amount of spread needed for EMI compliance.
For portable applications, package size and low power are important requirements for a system designer to consider in order that battery life is maximized. As an example, ON Semiconductor’s P3MS650100H & P3MS650103H peak EMI suppression ICs are offered in a 4-pin 1mm x 1.2mm WDFN package and have an auto-power down mode and are well suited for such applications.
Conventional spread spectrum techniques discussed so far involve usage of spread spectrum at the clock source. This approach suppresses EMI system-wide if the downstream interfaces can tolerate spread spectrum. However, if some of the downstream interfaces cannot tolerate spread spectrum, then devices such as ON Semiconductor’s patented Timing-SafeTM technology offer EMI suppression in interfaces where clock and data paths need to be synchronized. Timing-Safe technology provides system designers with additional flexibility to apply spread spectrum and reduce EMI locally without disturbing the clock and data synchronization, unlike conventional spread spectrum techniques. The oscilloscope waveforms in fig. 4 show an input clock to two devices and their outputs: one device generates a Timing-Safe TM clock output, and the other generates a conventional spread spectrum clock output. The persistence feature of the oscilloscope is turned ON in this waveform. It can be clearly seen that a Timing-Safe clock output can be used in place of the input clock to achieve peak EMI suppression, since this clock is synchronized with data, but a conventional spread spectrum clock can no longer be used because it is not synchronized with data.
(*See Figure 4)
When using spread spectrum, it is advisable for system designers to check if the bandwidth of their system can tolerate spread spectrum and ensure that the amount of spread used is within the jitter budget of the system. By including a provision early in the design cycle to use a spread spectrum device, system designers can mitigate the risk of failing EMI compliance and avoid losing valuable time to market in debug and re-design of the system boards. If the system fails EMI compliance, system designers can quickly resort to using the spread spectrum provision already built into the design.