EMI / EMC suppression of audio / video interface
Abstract: In the past, large external filters and shielded cables could be used to solve EMI / EMC (electromagnetic interference or electromagnetic compatibility) problems in audio / video products. These methods not only increase the cost, but also affect the performance of the end product and increase the size. As the size of these products continues to shrink and evolve into current audio / video players, EMI / EMC solutions must also reduce the size of the solution while maintaining or even improving system performance. To achieve this, small devices such as the MAX9511 graphic video interface and the MAX9705 Class D audio amplifier have been developed. These devices can provide excellent EMI performance in a smaller size. To illustrate how this performance improvement is achieved, let's examine the audio and display interface of a common computer to understand the EMI performance provided by these small devices. This article discusses various EMI problems that must be solved in the audio / video interface design, and then gives methods to solve these problems.
Introduction All electronic products sold globally must be tested for EMI / EMC before going to market to prove that they will not cause interference or be interfered by other equipment. For testing purposes, these products can be divided into two categories: active radiation products and non-active radiation products, for example, cell phones and walkie-talkies that actively radiate energy, while TVs, computers, and laptops should not radiate energy.
Different types of products and test institutions, EMI / EMC test requirements are also different. However, EMI / EMC tests can be roughly divided into two categories: Radiation: This test limits the amplitude and frequency of the signal radiated or conducted by a product so that it will not interfere with other products. Sensitivity (also called immunity): This test shows the product's ability to suppress radiation by limiting the amplitude and frequency of radiated and conducted signals that interfere with the normal operation of the device. As mentioned above, EMI can be divided into conducted interference and radiated interference. Since all EMI radiation is generated by current, these two types of interference are related to each other. But not all currents will produce radiation. Therefore, we must first analyze and suppress the problem of radiated interference, and then deal with the problem of conducted interference. For these two types of interference, radiated interference is more difficult to predict and suppress. Therefore, it is the main reason for failure of EMI test of most non-active radiation products. Here, we will focus on how to solve the radiated interference problem of the audio / video interface that is common in many products.
Various methods can be used to meet the conditions defined in the EMI / EMC regulations. But most of these methods can be classified into two categories: shielding and filtering. In actual products, these methods must be combined with specific applications to achieve a comprehensive EMI solution. For example, in most products, a metal shell is used to shield the radiation, and LC or RC filters are used to reduce the conducted interference of the input / output lines. In addition, a dithered clock can also be used to extend the spectral range to reduce the filtering or shielding requirements of specific applications.
When the EMI performance of the product basically meets the requirements, it will be taken to a certified laboratory for formal testing. If the product passes the test, it can be put on the market; if it fails, it means there is a problem. When solving problems, even a small change can take a long time. This may delay the time to market of products, because both the international and domestic markets require products to pass EMI / EMC compatibility tests¹. As a result, EMI design often sacrifices the video performance of the product to ensure that it passes the test. In modern design, the physical size and cost of the components required to pass the EMI test need to be considered, so the video performance will be sacrificed.
The size of modern audio / video analog interfaces continues to decrease, but the performance expectations are very high, which poses a very serious challenge to the design. To solve this problem, first find out the source of most EMI / EMC test failures; and then explore feasible solutions. The source of the test failure EMI / EMC test failure usually occurs in the weakest link in the product design-the signal (and interference) enters or leaves the shielded and filtered mechanism from this link. Among the audio / video interfaces, the weakest point is the cable connecting the devices. They are equivalent to antennas. For computers, the cable connecting the display and speakers to the PC is the weakest link, and it often causes EMI / EMC problems. We may think that only high-bandwidth video interfaces will cause this problem, while low-frequency audio interfaces will not have this problem. This is true when all amplifiers use Class A audio amplifiers. However, the currently used high-efficiency class D amplifiers² have high-frequency switching signals, and if there is no proper filtering and shielding, there will also be EMI problems.
In the past, large external filters and / or shielded cables could be used to solve these problems. But these methods not only increase the cost, but also affect product performance and increase product size. As these products continue to shrink in size and evolve into current audio / video players, EMI / EMC solutions must reduce product size while maintaining or even improving system performance. To achieve this goal, small devices such as the MAX9511 graphic video interface and the MAX9705 Class D audio amplifier have been developed, which can provide excellent EMI performance. In order to show how this improvement is achieved, one can examine the audio and display interface 7³ of a common PC and the EMI performance provided by these slim devices. First, we should understand the various EMI problems that must be solved in the audio / video interface design, and then give a solution to these problems. The video format commonly used by video and EMI computers, which is what we call "graphics", is not the same as the video format of TV4. Computer video has red, green, and blue (R, G, B) analog video signals, as well as logic signals consisting of line, field sync, and DDC5, all of which have fast rise / fall times. The video connector usually uses a high-density ultra-micro D-type connector to connect the monitor and the PC (Figure 1). Although this solution combines measures such as video signal shielding (coaxial) and common-mode choke (CMC) to reduce radiated and conducted EMI, it still needs to increase filtering to ensure that EMI requirements are met. In broadcast video applications, similar filtering measures are used to eliminate aliasing defects in TV images. However, this cannot be done in graphic video because the purpose of graphic video is to reproduce a checkerboard pattern of "on" and "off" pixels at the highest possible resolution. Therefore, in order to achieve the best display performance, we hope that the larger the bandwidth, the better. However, in practical applications, EMI and video performance must be weighed, so video bandwidth has to be sacrificed. For multi-signal video interfaces, multiple factors need to be weighed.
Figure 1. Typical VGA connection and video signal generating radiated EMI
For example, when filtering a video signal, a time delay will occur, and if the delay times of the video channels (R, G, and B) cannot be precisely matched, a color edge effect will occur. In order to avoid this phenomenon, the group delay and group delay matching of the video channel must be precisely controlled6. RGB video is extremely susceptible to these parameters7. For best performance, the group delay must be consistent with the frequency, and the minimum group delay match between channels must be kept within ± 0.5 pixel time. If the matching can be so precise, then the synchronization signal must also track the channel delay to correctly display the image frame. After this is done, the multi-video resolution problem supported by the PC needs to be resolved.
In this application, it is very difficult to achieve the best performance with a fixed frequency filter. If we design a filter to suppress EMI at the lowest resolution, the filter's stopband will intervene in the signal bandwidth of the higher resolution format, thereby affecting the higher resolution video performance. If the filter is designed for the highest resolution format, it may not meet EMI requirements. Obviously, the best solution is to use a "tunable" filter whose frequency response can track the display resolution, but this method will increase the cost and may also increase the product size. In addition, the impact of synchronous and fast rise / fall times of DDC drivers on EMI performance is also important. Therefore, in any complete EMI solution, methods that can delay these rise / fall times must be included. There are still some historical issues, such as the video DAC load detection function to meet the plug and play requirements.
MAX95118 can realize all these functions. As shown in Figure 2, the high-resolution graphics card output using MAX9511, using LC filter scheme, and the EMI characteristics of the original output without filtering.
Figure 2. EMI radiated under three conditions: a) no filtering, b) passive LC filter, c) MAX9511 complete EMI solution (MAX9511) The MAX9511 graphic video interface shown in Figure 3 provides RGB video A matched, three-channel adjustable EMI filter with a resolution range from VGA to UXGA and a skew error between channels of less than 0.5ns. The slew rate adjustment function is realized by changing the resistance of a single resistor (Rx). Corresponding to different VESA resolutions and sampling clock ranges, Table 1 lists the relationship between resistance and slew rate. In the circuit of Figure 4, the MAX54329 potentiometer controlled by I²C provides 32-stage filter control. However, it can be seen from Table 1 that only Level 3 or Level 4 control is required in most applications. In the final EMI / EMC test, without any mechanical or electrical changes, you can improve the EMI performance of a product.
Figure 3. MAX9511 VGA interface with EMI suppression
Figure 4. The MAX9511 drives multiple outputs. Adjustable filtering is controlled by the MAX5432 I²C digital potentiometer.
Table 1. MAX9511 slew rate, bandwidth, and Rx resistance values
The RGB video output is low impedance (ZOUT <1Ω), plus a 75Ω reverse termination resistor to provide 45dB to 50dB of isolation between the remote monitor and the dock. Previously, when this method was used to drive two different outputs, a switch was needed to avoid connecting the LC filter output to a long unterminated branch. As shown in Figure 4, it can be seen how the MAX9511 detects the output load. Whether the output load is connected or not will cause a significant change in the impedance of the DAC termination at the input. The video controller driving the RGB input can detect this impedance change, and if the load is not connected, the video output and the synchronous output are turned off by turning off the pin. DDC has been in the normally open state to support plug and play, the driver also has a level conversion function, which can convert the low voltage controller level to a standard 5V interface level. The synchronous driver has a 50Ω (typical) output impedance, and an external capacitor can be used to limit the edge slew rate (Figure 4). Synchronous jitter (without adding capacitance) is generally less than 0.5ns. Video performance also includes: + 6dB gain, 50dB SNR, 0.036% linearity error and less than 1% overshoot / undershoot (with excellent damping response characteristics). Audio and EMI audio interfaces need to solve a series of different problems to obtain efficiency and performance without generating EMI. In portable applications, we want to maximize battery life without expecting inefficient designs to generate heat, so Class D amplifiers are widely used. The problem is that Class D amplifiers use PWM to achieve high efficiency, which is very similar to switching power supplies. When an unshielded speaker cable is connected to the output, the cable will radiate EMI like an antenna. Although the clock frequency (typically 300kHz to 1MHz) is higher than the audio spectrum, it is a square wave with a lot of harmonic components. The size of the filter used to filter out the harmonic component is relatively large, and the cost is high. In portable applications such as laptops, this is not a viable solution due to size reasons10.
The general design topology cannot solve these two problems at the same time. To maximize the output audio power, portable applications use a bridge-to-load (BTL) connection, in which case the two wires of the speaker are effectively driven (Figure 5). In a class D amplifier, a comparator is used to monitor the analog input voltage and compare the input voltage with a triangle wave. When the amplitude of the triangle wave is higher than the audio input voltage, the comparator flips, and the inverter generates a complementary PWM waveform to drive the other side of the BTL output stage. Because of this BTL topology, the output filter actually requires twice the number of components for single-ended audio output: two inductors (L1 and L2) and two capacitors (C1 and C2). These two inductors need to handle the peak output current, so the size is relatively large and takes up most of the space.
Figure 5. The active radiation limiting technique used in a typical Maxim Class D audio amplifier.
Class D amplifiers can use the coil inductance and discrete capacitance of the speaker to form a filter, thereby eliminating the need for additional filters. Since the speaker cable will still radiate a considerable amount of energy, this method is limited to internal speakers. One approach is to change the switching process so that the amplifier maintains high efficiency while reducing EMI, thereby requiring only a small filter. To achieve this, the clock frequency can be modulated to reduce the energy per Hertz bandwidth11. This method is called clock spread spectrum modulation 12, or clock frequency jitter. However, the effectiveness of spectrum spreading has a certain range. The typical radiation spectrum shown in Figure 6 illustrates the effectiveness of this technique.
Figure 6. The MAX9705 radiation data obtained using the MAX9705EVKIT (12-inch unshielded twisted pair) shows the effect of spread spectrum modulation.
For devices that only provide spread-spectrum modulation, when the output power is higher than a few hundred milliwatts, speaker wires longer than a few inches will radiate too much energy. At this time, increasing the clock frequency does not help, as the frequency increases, the output spectrum of the class D amplifier will decrease. However, the wiring of the speaker will become as efficient as the antenna, which cancels out any improvement in performance. To further improve EMI performance, it is required to change the PWM waveform used by the class D amplifier itself. A specific method called active radiation limitation can be used to achieve this.
The active radiation limiting circuit sets the minimum pulse width of the amplifier, while the design in Figure 5 does not limit the maximum boundary. Combined with the control of crossover switching, rise / fall time and clock frequency, the power spectrum generated during operation can be limited to a given output power level. The purpose of this is to reduce the frequency spectrum to a certain level, so that without any external filtering and the connection of up to 24in external speaker connection, its EMI characteristics can still meet the radiation limit requirements.
We also hope to get good audio performance, which requires a peak power output greater than 2W. At the same time, it is also expected to minimize heat generation and maximize battery life. Therefore, the device needs to achieve high efficiency under a low voltage single power supply, and at the same time have a low power consumption shutdown mode suitable for headphone applications. THD + N must be low and SNR must be high. To have click suppression, the input must be compatible with single-ended or differential inputs. The MAX9705 can not only complete the above tasks, but also has more functions, as you will see from the following. Active Radiation Limitation (MAX9705) The active radiation limitation technology used by Maxim's Class D amplifiers is shown in Figure 7. It is not clear from this figure how the switch operates. Through careful design of the drive circuit and zero dead time control, the efficiency of the MAX9705 Class D amplifier can exceed 85%. The unique and proprietary spread spectrum modulation mode flattens the spectrum components and reduces the EMI radiation generated by connecting cables and speakers. In stereo or multi-channel applications, the synchronous input locks the clock frequency of all amplifiers in the 800kHz to 2MHz common clock range, thereby minimizing intermodulation, otherwise multiple independent clock sources will produce intermodulation. Maxim's Class D audio amplifier combines two unique technologies: spread spectrum modulation and active radiation limitation. It is possible to connect up to 24in of unshielded speaker cables without "filter", while the radiated interference remains within the EMI limits specified in FCC Part 15 (Figure 8).
Figure 7. The MAX9705 Class D amplifier generates a sawtooth wave and provides a differential input. If a single-ended input is used, a differential input can be generated internally.
Figure 8. Radiation data when the MAX9705 is connected to 24in unshielded twisted pair in spread spectrum modulation mode
In addition to EMI, audio performance is also very good, THD + D is 0.02% at 1W, increased to 1% at 2.3W, and SNR is 90dB. The input can be a differential input or a single-ended input, providing + 6dB, + 12dB, + 15.6dB, or + 20dB fixed gain, which can meet any application (Figure 7). The power consumption is minimized in shutdown mode. In addition, the sync input allows the MAX9705 to provide mono, stereo, or multi-channel high-performance audio, which can still meet EMI emission requirements when connected to external speakers without filters. Conclusion MAX9511 and MAX9705 represent the advanced technology of EMI / EMC control. Applying these devices to products can effectively reduce EMI. There is no need to rely on large-scale external filters and shielding methods that will increase cost and size as before. These devices use the most advanced technology today, effectively ensuring electromagnetic compatibility and performance.
Different types of products and test institutions, EMI / EMC test requirements are also different. However, EMI / EMC tests can be roughly divided into two categories: Radiation: This test limits the amplitude and frequency of the signal radiated or conducted by a product so that it will not interfere with other products. Sensitivity (also called immunity): This test shows the product's ability to suppress radiation by limiting the amplitude and frequency of radiated and conducted signals that interfere with the normal operation of the device. As mentioned above, EMI can be divided into conducted interference and radiated interference. Since all EMI radiation is generated by current, these two types of interference are related to each other. But not all currents will produce radiation. Therefore, we must first analyze and suppress the problem of radiated interference, and then deal with the problem of conducted interference. For these two types of interference, radiated interference is more difficult to predict and suppress. Therefore, it is the main reason for failure of EMI test of most non-active radiation products. Here, we will focus on how to solve the radiated interference problem of the audio / video interface that is common in many products.
Various methods can be used to meet the conditions defined in the EMI / EMC regulations. But most of these methods can be classified into two categories: shielding and filtering. In actual products, these methods must be combined with specific applications to achieve a comprehensive EMI solution. For example, in most products, a metal shell is used to shield the radiation, and LC or RC filters are used to reduce the conducted interference of the input / output lines. In addition, a dithered clock can also be used to extend the spectral range to reduce the filtering or shielding requirements of specific applications.
When the EMI performance of the product basically meets the requirements, it will be taken to a certified laboratory for formal testing. If the product passes the test, it can be put on the market; if it fails, it means there is a problem. When solving problems, even a small change can take a long time. This may delay the time to market of products, because both the international and domestic markets require products to pass EMI / EMC compatibility tests¹. As a result, EMI design often sacrifices the video performance of the product to ensure that it passes the test. In modern design, the physical size and cost of the components required to pass the EMI test need to be considered, so the video performance will be sacrificed.
The size of modern audio / video analog interfaces continues to decrease, but the performance expectations are very high, which poses a very serious challenge to the design. To solve this problem, first find out the source of most EMI / EMC test failures; and then explore feasible solutions. The source of the test failure EMI / EMC test failure usually occurs in the weakest link in the product design-the signal (and interference) enters or leaves the shielded and filtered mechanism from this link. Among the audio / video interfaces, the weakest point is the cable connecting the devices. They are equivalent to antennas. For computers, the cable connecting the display and speakers to the PC is the weakest link, and it often causes EMI / EMC problems. We may think that only high-bandwidth video interfaces will cause this problem, while low-frequency audio interfaces will not have this problem. This is true when all amplifiers use Class A audio amplifiers. However, the currently used high-efficiency class D amplifiers² have high-frequency switching signals, and if there is no proper filtering and shielding, there will also be EMI problems.
In the past, large external filters and / or shielded cables could be used to solve these problems. But these methods not only increase the cost, but also affect product performance and increase product size. As these products continue to shrink in size and evolve into current audio / video players, EMI / EMC solutions must reduce product size while maintaining or even improving system performance. To achieve this goal, small devices such as the MAX9511 graphic video interface and the MAX9705 Class D audio amplifier have been developed, which can provide excellent EMI performance. In order to show how this improvement is achieved, one can examine the audio and display interface 7³ of a common PC and the EMI performance provided by these slim devices. First, we should understand the various EMI problems that must be solved in the audio / video interface design, and then give a solution to these problems. The video format commonly used by video and EMI computers, which is what we call "graphics", is not the same as the video format of TV4. Computer video has red, green, and blue (R, G, B) analog video signals, as well as logic signals consisting of line, field sync, and DDC5, all of which have fast rise / fall times. The video connector usually uses a high-density ultra-micro D-type connector to connect the monitor and the PC (Figure 1). Although this solution combines measures such as video signal shielding (coaxial) and common-mode choke (CMC) to reduce radiated and conducted EMI, it still needs to increase filtering to ensure that EMI requirements are met. In broadcast video applications, similar filtering measures are used to eliminate aliasing defects in TV images. However, this cannot be done in graphic video because the purpose of graphic video is to reproduce a checkerboard pattern of "on" and "off" pixels at the highest possible resolution. Therefore, in order to achieve the best display performance, we hope that the larger the bandwidth, the better. However, in practical applications, EMI and video performance must be weighed, so video bandwidth has to be sacrificed. For multi-signal video interfaces, multiple factors need to be weighed.
Figure 1. Typical VGA connection and video signal generating radiated EMI
For example, when filtering a video signal, a time delay will occur, and if the delay times of the video channels (R, G, and B) cannot be precisely matched, a color edge effect will occur. In order to avoid this phenomenon, the group delay and group delay matching of the video channel must be precisely controlled6. RGB video is extremely susceptible to these parameters7. For best performance, the group delay must be consistent with the frequency, and the minimum group delay match between channels must be kept within ± 0.5 pixel time. If the matching can be so precise, then the synchronization signal must also track the channel delay to correctly display the image frame. After this is done, the multi-video resolution problem supported by the PC needs to be resolved.
In this application, it is very difficult to achieve the best performance with a fixed frequency filter. If we design a filter to suppress EMI at the lowest resolution, the filter's stopband will intervene in the signal bandwidth of the higher resolution format, thereby affecting the higher resolution video performance. If the filter is designed for the highest resolution format, it may not meet EMI requirements. Obviously, the best solution is to use a "tunable" filter whose frequency response can track the display resolution, but this method will increase the cost and may also increase the product size. In addition, the impact of synchronous and fast rise / fall times of DDC drivers on EMI performance is also important. Therefore, in any complete EMI solution, methods that can delay these rise / fall times must be included. There are still some historical issues, such as the video DAC load detection function to meet the plug and play requirements.
MAX95118 can realize all these functions. As shown in Figure 2, the high-resolution graphics card output using MAX9511, using LC filter scheme, and the EMI characteristics of the original output without filtering.
Figure 2. EMI radiated under three conditions: a) no filtering, b) passive LC filter, c) MAX9511 complete EMI solution (MAX9511) The MAX9511 graphic video interface shown in Figure 3 provides RGB video A matched, three-channel adjustable EMI filter with a resolution range from VGA to UXGA and a skew error between channels of less than 0.5ns. The slew rate adjustment function is realized by changing the resistance of a single resistor (Rx). Corresponding to different VESA resolutions and sampling clock ranges, Table 1 lists the relationship between resistance and slew rate. In the circuit of Figure 4, the MAX54329 potentiometer controlled by I²C provides 32-stage filter control. However, it can be seen from Table 1 that only Level 3 or Level 4 control is required in most applications. In the final EMI / EMC test, without any mechanical or electrical changes, you can improve the EMI performance of a product.
Figure 3. MAX9511 VGA interface with EMI suppression
Figure 4. The MAX9511 drives multiple outputs. Adjustable filtering is controlled by the MAX5432 I²C digital potentiometer.
Table 1. MAX9511 slew rate, bandwidth, and Rx resistance values
| Rx (kΩ) | MAX9511 Slew Rate vs. Rx | ||
| Slew Rate (V / ms) | Pixel Clock Frequency (MHz) | VESA ResoluTIon | |
| 7 | 1408 | 160 to 230 | UXGA (1600 x 1200) |
| 10 | 1255 | 160 to 230 | UXGA (1600 x 1200) |
| 12 | 1050 | 100 to 150 | SXGA (1280 x 1024) |
| 15 | 810 | 100 to 150 | SXGA (1280 x 1024) |
| 20 | 613 | 45 to 95 | XGA (1024 x 768) |
| 25 | 470 | 45 to 95 | XGA (1024 x 768) |
| 30 | 368 | 45 to 95 | XGA (1024 x 768) |
| 35 | 298 | 35 to 50 | XGA (1024 x 768) |
| 40 | 255 | 35 to 50 | SVGA (800 x 600) |
| 45 | 203 | 35 to 50 | SVGA (800 x 600) |
| 50 | 158 | 25 to 30 | VGA (640 x 480) |
| > 50 | <150 | <25 | QCIF |
The RGB video output is low impedance (ZOUT <1Ω), plus a 75Ω reverse termination resistor to provide 45dB to 50dB of isolation between the remote monitor and the dock. Previously, when this method was used to drive two different outputs, a switch was needed to avoid connecting the LC filter output to a long unterminated branch. As shown in Figure 4, it can be seen how the MAX9511 detects the output load. Whether the output load is connected or not will cause a significant change in the impedance of the DAC termination at the input. The video controller driving the RGB input can detect this impedance change, and if the load is not connected, the video output and the synchronous output are turned off by turning off the pin. DDC has been in the normally open state to support plug and play, the driver also has a level conversion function, which can convert the low voltage controller level to a standard 5V interface level. The synchronous driver has a 50Ω (typical) output impedance, and an external capacitor can be used to limit the edge slew rate (Figure 4). Synchronous jitter (without adding capacitance) is generally less than 0.5ns. Video performance also includes: + 6dB gain, 50dB SNR, 0.036% linearity error and less than 1% overshoot / undershoot (with excellent damping response characteristics). Audio and EMI audio interfaces need to solve a series of different problems to obtain efficiency and performance without generating EMI. In portable applications, we want to maximize battery life without expecting inefficient designs to generate heat, so Class D amplifiers are widely used. The problem is that Class D amplifiers use PWM to achieve high efficiency, which is very similar to switching power supplies. When an unshielded speaker cable is connected to the output, the cable will radiate EMI like an antenna. Although the clock frequency (typically 300kHz to 1MHz) is higher than the audio spectrum, it is a square wave with a lot of harmonic components. The size of the filter used to filter out the harmonic component is relatively large, and the cost is high. In portable applications such as laptops, this is not a viable solution due to size reasons10.
The general design topology cannot solve these two problems at the same time. To maximize the output audio power, portable applications use a bridge-to-load (BTL) connection, in which case the two wires of the speaker are effectively driven (Figure 5). In a class D amplifier, a comparator is used to monitor the analog input voltage and compare the input voltage with a triangle wave. When the amplitude of the triangle wave is higher than the audio input voltage, the comparator flips, and the inverter generates a complementary PWM waveform to drive the other side of the BTL output stage. Because of this BTL topology, the output filter actually requires twice the number of components for single-ended audio output: two inductors (L1 and L2) and two capacitors (C1 and C2). These two inductors need to handle the peak output current, so the size is relatively large and takes up most of the space.
Figure 5. The active radiation limiting technique used in a typical Maxim Class D audio amplifier.
Class D amplifiers can use the coil inductance and discrete capacitance of the speaker to form a filter, thereby eliminating the need for additional filters. Since the speaker cable will still radiate a considerable amount of energy, this method is limited to internal speakers. One approach is to change the switching process so that the amplifier maintains high efficiency while reducing EMI, thereby requiring only a small filter. To achieve this, the clock frequency can be modulated to reduce the energy per Hertz bandwidth11. This method is called clock spread spectrum modulation 12, or clock frequency jitter. However, the effectiveness of spectrum spreading has a certain range. The typical radiation spectrum shown in Figure 6 illustrates the effectiveness of this technique.
Figure 6. The MAX9705 radiation data obtained using the MAX9705EVKIT (12-inch unshielded twisted pair) shows the effect of spread spectrum modulation.
For devices that only provide spread-spectrum modulation, when the output power is higher than a few hundred milliwatts, speaker wires longer than a few inches will radiate too much energy. At this time, increasing the clock frequency does not help, as the frequency increases, the output spectrum of the class D amplifier will decrease. However, the wiring of the speaker will become as efficient as the antenna, which cancels out any improvement in performance. To further improve EMI performance, it is required to change the PWM waveform used by the class D amplifier itself. A specific method called active radiation limitation can be used to achieve this.
The active radiation limiting circuit sets the minimum pulse width of the amplifier, while the design in Figure 5 does not limit the maximum boundary. Combined with the control of crossover switching, rise / fall time and clock frequency, the power spectrum generated during operation can be limited to a given output power level. The purpose of this is to reduce the frequency spectrum to a certain level, so that without any external filtering and the connection of up to 24in external speaker connection, its EMI characteristics can still meet the radiation limit requirements.
We also hope to get good audio performance, which requires a peak power output greater than 2W. At the same time, it is also expected to minimize heat generation and maximize battery life. Therefore, the device needs to achieve high efficiency under a low voltage single power supply, and at the same time have a low power consumption shutdown mode suitable for headphone applications. THD + N must be low and SNR must be high. To have click suppression, the input must be compatible with single-ended or differential inputs. The MAX9705 can not only complete the above tasks, but also has more functions, as you will see from the following. Active Radiation Limitation (MAX9705) The active radiation limitation technology used by Maxim's Class D amplifiers is shown in Figure 7. It is not clear from this figure how the switch operates. Through careful design of the drive circuit and zero dead time control, the efficiency of the MAX9705 Class D amplifier can exceed 85%. The unique and proprietary spread spectrum modulation mode flattens the spectrum components and reduces the EMI radiation generated by connecting cables and speakers. In stereo or multi-channel applications, the synchronous input locks the clock frequency of all amplifiers in the 800kHz to 2MHz common clock range, thereby minimizing intermodulation, otherwise multiple independent clock sources will produce intermodulation. Maxim's Class D audio amplifier combines two unique technologies: spread spectrum modulation and active radiation limitation. It is possible to connect up to 24in of unshielded speaker cables without "filter", while the radiated interference remains within the EMI limits specified in FCC Part 15 (Figure 8).
Figure 7. The MAX9705 Class D amplifier generates a sawtooth wave and provides a differential input. If a single-ended input is used, a differential input can be generated internally.
Figure 8. Radiation data when the MAX9705 is connected to 24in unshielded twisted pair in spread spectrum modulation mode
In addition to EMI, audio performance is also very good, THD + D is 0.02% at 1W, increased to 1% at 2.3W, and SNR is 90dB. The input can be a differential input or a single-ended input, providing + 6dB, + 12dB, + 15.6dB, or + 20dB fixed gain, which can meet any application (Figure 7). The power consumption is minimized in shutdown mode. In addition, the sync input allows the MAX9705 to provide mono, stereo, or multi-channel high-performance audio, which can still meet EMI emission requirements when connected to external speakers without filters. Conclusion MAX9511 and MAX9705 represent the advanced technology of EMI / EMC control. Applying these devices to products can effectively reduce EMI. There is no need to rely on large-scale external filters and shielding methods that will increase cost and size as before. These devices use the most advanced technology today, effectively ensuring electromagnetic compatibility and performance.
Antenna For Automotive Product,Magnetic Antenna Coil Element,Antenna Impedance Matching Transformer,Transmitting And Receiving Antenna
IHUA INDUSTRIES CO.,LTD. , https://www.ihua-magnetics.com