"Electromagnetic radiation (EMR), electromagnetic interference (EMI) and electromagnetic compatibility (EMC) are terms involving energy from charged particles and may interfere with circuit performance and signal transmission. With the surge of wireless communication, communication devices are not Number, plus more and more spectrum used in communication methods (including honeycombs, Wi-Fi, satellites, GPS, etc.) (some frequency bands overlap each other), electromagnetic interference into objective facts. In order to alleviate In this influence, many government agencies and regulatory organizations set restrictions on communication devices, equipment and instruments. One of such specifications is CISPR 16-1-3, which involves radio interference and immunity measurement. Equipment and measurement methods.
According to the characteristics, electromagnetic interference can be divided into conduction interference (via power transmission) or radiation interference (via air). Switching power supplies generate two types of interference. A technique for reducing conduction interference and radiation interference is spreading frequency modulation (SSFM). This technique is used for our switching power supplies, silicon oscillators, and LED drivers based on inductance and capacitors, extends noise to a wider frequency band, thereby reducing peak noise and flat average noise at a particular frequency.
SSFM does not allow emission energy to stay in the frequency band of any receiver, thereby improving EMI. The key determinants of effective SSFM are frequency expansion and modulation rates. For switching regulator applications, typical expansion is ± 10%, and the optimal modulation rate depends on the modulation method. SSFM can use a variety of frequency extensions, such as using a sine wave or triangular modulation clock frequency.
Be
Modulation method
Most switching regulators will present the frequency-related ripple: the lower the switching frequency, the more the ripple is, the higher the switching frequency, the less ripple. Therefore, if the switching clock is frequency modulated, the ripple wave of the switching regulator will present amplitude modulation. If the modulation signal of the clock is periodic (eg, a sine wave or triangular wave), periodic ripple modulation will be present, and there is a significant spectral component in the modulation frequency (Fig. 1).
Be
Figure 1. Switch regulator ripple pattern caused by the sinusoidal frequency modulation of the clock.
Be
Since the modulation frequency is much lower than the clock frequency of the switching regulator, it may be difficult to filter out. This may result in a problem such as audio or obvious artifact due to the coupling or limited power supply in the downstream circuit. Pseudo-random frequency modulation can eliminate this cycle of cycle. When using pseudo-random frequency modulation, the clock is converted from one frequency to another from a pseudo-random manner. Since the output ripple of the switching regulator is amplified by the type noise signal, the output seems to have not been modulated, and the effect of the downstream system can be ignored.
Be
Modulation amount
As the SSFM frequency range increases, the percentage of internal time is reduced. As can be seen from Figure 2 below, the modulation frequency is presented as a broadband signal compared to a single unmodulated narrowband signal, and the peak is reduced by 20 dB. If the transmitted signal does not usually enter the frequency band of the receiver and the time of the stay is short (relative to its response time), the EMI can be significantly reduced. For example, in low EMI, ± 10% frequency modulation is much more effective than ± 2% frequency modulation. 1 However, the frequency range that the switching regulator can allow is limited. In general, most switching regulators can easily tolerate the frequency change of ± 10%.
Be
Figure 2. Spread spectrum modulation produces lower peak energy in a wider clock frequency band.
Be
Modulation rate
Similar to the amount of modulation, for a given receiver, as the frequency modulation rate increases (frequency hopping rate), the time of the given receiver in the belt will be reduced, so the EMI will decrease. However, the frequency change rate (DF / DT) that the switching regulator can track has a limit. Its solution is to identify the highest modulation rate that does not affect the output adjustment of the switching regulator.
Be
Measuring EMI
The typical method of measuring EMI is peak detection, quasi-peak detection or average detection. For these tests, the bandwidth of the test device is appropriately set to reflect the actual target bandwidth and determine the effectiveness of the SSFM. When frequency modulation is performed, the detector responds with the frequency band of the entire detector as transmitted. When the bandwidth of the detector is less than the modulation rate, the limited response time of the detector causes the EMI measurement value attenuation. Instead, the response time of the detector does not affect the fixed frequency emission, so that EMI attenuation is not observed. The peak detection test shows that the improvement obtained by SSFM corresponds directly to the amount of attenuation. The peak detection test can also display further EMI improvements because it includes the effect of duty cycle. Specifically, the fixed frequency emission generates 100% duty ratio, and the duty ratio from SSFM is reduced with the amount of time occupied by the transmitted frequency band. Finally, the average value detection test can display the most obvious EMI improvement because it uses low by filter peak detection signals to generate an average band energy. When the fixed frequency is transmitted, the average value and the peak energy are equal, and the SSFM is different, and it is attenuated by the peak detection energy and the internal amount of time, resulting in a lower average value detection result. Many regulatory tests require two detection tests by quasi-peak and average value.
Be
SSFM and receiver bandwidth
Regardless of whether SSFM is enabled, the peak emission of the switching regulator may appear the same at any time. How can this be? The effectiveness part of the SSFM depends on the bandwidth of the receiver. To receive an instantaneous emission snapshot, you need an unlimited bandwidth. The bandwidth of each actual system is limited. If the clock frequency changes faster than the bandwidth of the receiver, it will significantly reduce the reception interference.
Be
Figure 3. Output spectrum (9 kHz resolution bandwidth) using the LTC6908 switching regulator (9 kHz resolution bandwidth) that enables SSFM and does not enable SSFM.
Be
SSFM in the silicon oscillator
The LTC6909, LTC6902 and LTC6908 are multiphase silicon oscillators having an eight-phase, four-phase and double phase output of spread spectrum modulation. These devices are typically used to provide clocks for switching power supplies. The multiphase operation effectively increases the switching frequency of the system (since the phase is expressed as an increase in the switching frequency), and the spreading modulation enables each device to switches within a certain frequency range, thereby expanding the conduction EMI on a wider frequency band. The LTC6908 has a frequency range of 5 kHz to 10 MHz, providing two outputs: LTC6908-1 provides two outputs with 180 ° phase shift, and LTC6908-2 provides two of the 90 ° phase shift. Output. The former is ideal for synchronization of two single-switching regulators, which are ideal for synchronous two-phase dual switching regulators. The four-channel LTC6902 has a frequency range of 5 kHz to 20 MHz, which can be programmed for two-phase, three-phase or four-phase outputs such as equal spacing. The LTC6909 has a frequency range of 12 kHz to 6.67 MHz and can be programmed to provide eight-phase output.
In order to solve the above periodic ripple problem, these silicon oscillators use pseudo-random frequency modulation. With this technique, the switching regulator clock is converted from one frequency from one frequency to another in a pseudo-random manner. The higher the frequency offset or frequency hopping rate, the shorter the working time of the switch regulator at a given frequency, and the EMI will be shorter in the tape for a given receiver interval.
Be
Figure 4. Pseudo-random modulation illustrates the impact of the LTC6908 / LTC6909 internal tracking filter.
Be
However, there is a limit on the frequency hopping rate. If the frequency is hopped at a rate of exceeding the switch stretcher bandwidth, the peak may occur at the clock frequency conversion edge. Smaller switching regulator bandwidth can result in a more obvious spike. Thus, the LTC6908 and LTC 6909 include a proprietary tracking filter that can achieve smooth conversion from one frequency to the next frequency (LTC6902 uses a 25 kHz internal low-pass filter). The internal filter tracks the hopping rate, providing the best smooth performance for all frequencies and modulation rates.
For many logical systems, this filter modulation signal may be acceptable, but must carefully consider the progressive jitter problem.
Even if the tracking filter is used, the bandwidth of the given regulator may still be insufficient to meet the requirements of high speed frequency modulation. To address bandwidth limitation, the frequency hopping rate of the LTC6908 / LTC6909 can be reduced from 1/32 or 1/64 of the nominal frequency from the default rate (i.e., 1/16 of the nominal frequency).
Click here to view the spread frequency modulation silicon oscillator.
Be
SSFM in the power supply
The switching regulator is based on cycle-by-cycle to transfer power to the output. In most cases, the operating frequency is either fixed or is based on the constant output load. This conversion method produces a larger noise component at the operational frequency (basis) and the multiplier (harmonics) of the operating frequency. Click here to view the expansion frequency modulation buckstick list.
Be
LTM4608A: 8 A, 2.7 V to 5.5 Vin DC / DC μModule® buck regulator for SSFM
In order to reduce the switching noise, the CLKIN pin of the LTM4608A can be connected to the SVIN (low power circuit power supply voltage pin) to enable spread spectrum function. In spread spectrum mode, the internal oscillator of the LTM4608A is designed to generate clock pulses, and the cycle is randomly based on weekly, but fixed to between 70% and 130% of the nominal frequency. This facilitates the expansion of switching noise in a certain frequency range, thereby significantly reduces peak noise. If the CLKIN is grounded or driven by the external frequency synchronous signal, the spread spectrum operation is disabled. Figure 5 shows an operating circuit that enables spread spectrum operation. A 0.01 μF grounding capacitor must be placed on the PLL LPF pin to control the dispel rate of the expansion frequency change. The component value is determined by the following formula:
Be
Be
LT8609: 42 V input with SSFM, 2 A synchronization converter
The LT8609 is a micro-power buck converter that maintains high efficiency at high switching frequencies (93% of 2 MHz), allowing for smaller external components. The operation of the SSFM mode is similar to the jumper pulse working mode, which is mainly different that the switching frequency is modulated by 3 kHz triangular waves. The low end of the modulation range is set by the switching frequency (which is set by the resistance on the RT pin), and the high end is set to about 20% higher than the frequency set by the RT. To enable the spread spectrum mode, you must connect the SYNC pin to the INTVCC or drive it to a voltage between 3.2 V and 5 V.
Be
Figure 5. LTM4608A that enables spread spectrum.
Be
LTC3251 / LTC3252: charge pump buck regulator with SSFM
The LTC3251 / LTC3252 is 2.7 V to 5.5 V, single output 500 mA / dual output 250 mA charge pump bucking regulator, can generate clock pulses, and the cycle is randomly based on cycle, but fixed Between 1 MHz to 1.6 MHz. Figures 6 and 7 show that the spread spectrum characteristics of the LTC 3251 significantly reduces the peak harmonic noise compared to the conventional buck converter. The LTC3251 provides an optional spreading operation, while the LTC3252 is always enabled.
Be
Figure 6. LTC3251 of SSFM is disabled.
Be
Figure 7. LTC3251 of the SSFM is enabled.
Be
SSFM in the LED drive
LT3795: 110 V multi-topological LED controller with SSFM
The switching regulator LED driver is also a trouble for the EMI problem applied by the car and display illustrated. In order to improve EMI performance, the LT3795 110 V multi-top topology LED drive controller integrates SSFM. If there is a capacitor on the RAMP pin, a triangular wave between 1 V and 2 V is generated. The signal is then fed into the internal oscillator, and the switching frequency is modulated between the 70% and the base frequency of the fundamental frequency, and the base frequency is set by the clock frequency to set the resistor RT setting. The modulation frequency calculation formula is as follows:
Be
Figures 8 and 9 show the noise spectrum comparison between the conventional boost switch converter circuit (connecting the RAMP pin to GND) and enabling spread spectrum modulation. . Figure 8 shows an average conducting EMI, Figure 9 shows a peak conduction EMI. The results of EMI measurement are susceptible to the effect of RAMP frequencies selected by the capacitance. 1 kHz is a good starting point for optimizing peak measurements, but in order to obtain the best results of overall EMI in a particular system, it may be necessary to fine-tuning this value.
Be
Figure 8. LT3795 average conduction EMI.
Be
Figure 9. LT3795 Peak conduction EMI.
Be
LT3952: Multi Topology with SSFM 42 VIN, 60 V / 4 A LED Driver
The LT3952 is a 60 V / 4 A power switch type, constant current, constant voltage, multi-topological LED driver, providing an optional SSFM. The oscillator frequency is changed from the nominal frequency (FSW) to 31% higher than the nominal value in a pseudo-random manner, and the step size is 1%. This one-way adjustment enables the LT3952 to program a nominal frequency to one point it can avoid the sensitive band (e.g., AM radio spectrum) in the system. The proportional step allows the user to easily determine the clock frequency value (RT pin) applied to the specified EMI test warehouse, and the pseudo-random method can provide tone suppression from the frequency change itself.
The update of pseudo-random values uses the rate of FSW / 32, and is proportional to the oscillator frequency. This rate allows the entire set frequency to pass multiple times in standard EMI test residence time.
Be
Figure 10. LT3952 Average Transmission EMI.
Be
ADI also provides many other products that can effectively use design techniques to reduce EMI. As mentioned above, using SSFM is one of the techniques. Other methods include slowing down the fast internal clock edge and internal filtering. Use our SilentSwitcher® technology has realized another innovative method that effectively reduces EMI by layout. The LT8640 is a unique 42 V input, a micro-power synchronous buffet switching regulator, which combines Silent Switcher technology and SSFM to reduce EMI. Therefore, when you encounter EMI questions again in your design, be sure to view our low EMI products to help you easily meet EMI standards.
Note:
For microprocessors and data clocks, ± 2% SSFM is common because they cannot tolerate larger frequency changes.
The repeated rate of complete pseudo-random sequence is guaranteed to be less than 20 Hz.
Be
About the Author
Greg Zimmer is the marketing manager of the ADI's battery management system, has a wealth of experience in various high-performance signal conditioning IC product marketing. GREG has a background of marketing, technology marketing, application engineering and analog circuit design.
Greg has a bachelor's degree in electrical engineering and computer science in the University of California University and a bachelor's degree in economics in Santa Cruz, California.
Kevin Scott is the product marketing manager of the Department of Power Products, ADI, is responsible for managing boost, buck-boosting and isolation converters and LED drivers and linear regulators. He has served as a senior strategic marketing engineer, which is responsible for developing technical training content, training sales engineers, and writes a large number of website articles on the company's numerous product technology advantages. He has been in the semiconductor industry for 26 years of experience, serving a number of applications, business management and marketing. KEVIN graduated from Stanford University in 1987 to obtain a bachelor's degree in electrical engineering, and started his engineering and technical career after a short NFL (American Rugby League) career.
Be
Article from the network
Our other product:
