Powering High-Speed ​​ADCs with Switching Regulators

Publisher:平章大人Latest update time:2012-03-31 Source: EDNCHINA Reading articles on mobile phones Scan QR code
Read articles on your mobile phone anytime, anywhere

For designers selecting high-speed data converters, power consumption is the most important system design parameter. Whether it is a portable design that requires long battery life or a small product that consumes less heat, power consumption is critical. In the past, system designers have used low-noise linear regulators, such as low-dropout regulators, to power data converters instead of switching regulators because they were concerned that switching noise would enter the converter's output spectrum, greatly degrading AC performance.

However, a newer generation of noise-optimized switching regulators (used in cell phones) has enabled a shift in applications by minimizing interference with adjacent low-noise power amplifiers. They can power high-speed data converters directly from a DC/DC converter without significantly degrading AC performance. This design can immediately improve power efficiency by 20% to 25%.

Modern high-speed converters can consume about 50% less power than previous generations, in part by reducing the supply voltage from 3.3V to 1.8V. In a design using low-dropout regulators, as the supply rails go down, the regulator voltage drop and the available supply rails become more important to power efficiency. The data section of the board typically has many voltage rails, providing various core and I/O voltages for FPGAs and processors. In the analog section, there may be only a few "clean" voltages to choose from, such as 3.3V and 5V.

For a high-speed data converter, a linear regulator can be used to derive 3.3V from a common 5V rail. This results in a 1.7V drop across the low-dropout regulator, which translates to a power loss of about 35%. When a low-dropout regulator (such as the ADS4149) is used to provide 1.8V to an ADC from a 3.3V bus (Reference 1), the power loss in the linear regulator increases to about 45%, meaning that the low-dropout regulator dissipates almost half the power. This example shows how an inefficient power supply design can easily lose 50% of the power. The efficiency of a switching regulator is independent of the size of the input rail, so considerable power savings can be achieved. The impact on AC performance can be minimized through careful design.

Power supply filtering

A key component to isolate the switching noise from the ADC is the power supply filter, which includes a ferrite bead and bypass capacitors. Several key features should be considered when selecting a ferrite bead. First, the ferrite bead must have sufficient current rating for the data converter, and it must have a low DCR (DC resistance) to minimize the voltage drop across the bead itself. For example, when a 200mA power supply passes through a bead with a DCR of 1Ω, a 200mV voltage drop is generated. This voltage drop may push the voltage on the ADC to the edge, and considering the standard deviation of the power supply voltage, the ADC voltage may even be lower than the recommended operating voltage.

Secondly, the ferrite bead must have high impedance to the harmonics of the switching frequency and DC/DC converter to block switching noise and switching glitches. The impedance of most ferrite beads on the market is 100MHz, while the typical switching frequency of modern DC/DC converters is 500kHz~6MHz. In our example, the ADS4149 evaluation module uses a TPS625290 switching regulator with a switching frequency of 2.25MHz (Reference 2). Since the DC/DC regulator has a square wave output, higher-order harmonics must also be considered. Murata's NFM31PC276B0J3EMI filter has high impedance and low DCR in this frequency range.

Figure 1 compares the insertion loss of a traditional ferrite bead with a Murata EMI filter with a resistance of 68Ω at 100MHz. The power circuit has a low impedance, and the insertion loss is measured in a 50Ω environment. Therefore, the insertion loss value of the power filter may be slightly different even though the resonant frequency does not change.

Figure 1. Murata's NFM31PC276B0J3 EMI has high impedance and low DCR compared to a traditional ferrite bead with a resistance of 68Ω at 100 MHz.

The other components in the power filter are the bypass capacitors. These capacitor values ​​should be chosen so that their resonant frequency (creating a low impedance path to ground) is close to the switching frequency. This way, the switching noise through the bead is shorted to ground. The power filter insertion loss comparison in Figure 2 shows that the correct bypass capacitor value can produce a resonance close to the switching frequency, even for a traditional ferrite bead, such as the EXCML32A680. However, at low frequencies, if it is placed together with a 0Ω resistor, there is not much difference. On the other hand, the Murata EMI filter provides about 20dB of additional attenuation around the switching frequency. The power filter in Figure 3 uses a 33μF tantalum capacitor for broadband decoupling, while the 10μF, 2.2μF, and 0.1μF ceramic capacitors have narrow resonant frequencies.

Figure 2. The correct bypass capacitor value can produce a resonance close to FS (switching frequency) even when used in conjunction with a traditional ferrite bead such as the EXC-ML32A680.

Figure 3: This power filter uses a 33μF tantalum capacitor for broadband decoupling, while 10μF, 2.2μF, and 0.1μF ceramic capacitors have narrow resonant frequencies.

AC Performance

Depending on the PSRR characteristics of the data converter, some amount of noise on the power rail can still enter the ADC and degrade the AC performance. Figure 4 shows a comparison of SNR and SFDR (spurious-free dynamic range) for a benchmark power supply (such as a 1.8V clean lab supply) using a low-dropout regulator and a DC/DC converter with different power supply filter options using the ADS4149 evaluation module.

Figure 4 These SNR (a) and SFDR (b) sweep plots compare a baseline power supply (such as a 1.8V clean lab supply) using a low dropout regulator and a DC/DC converter using the ADS4149 evaluation module with different power supply filter options.

The test results show that at an intermediate frequency of 300MHz, when powered by a switching regulator, the SNR performance is about 0.3dB lower than that of a low-noise LDO regulator. The SFDR performance is almost identical for all approaches. A careful look at the normalized FFT plot (starting with the input signal and plotting noise vs. offset frequency) shows that when using a suboptimal EXC ferrite bead, the background noise is slightly elevated throughout the Nyquist zone without any evidence of switching frequency feed-in (Figure 5).


Figure 5, a normalized FFT plot starting with the input signal and plotting noise vs. offset frequency, shows that when a suboptimal EXC ferrite is used, the background noise is slightly elevated across the Nyquist zone without any evidence of switching frequency feed-through.

Power efficiency

The main advantage of replacing linear regulators with DC/DC converters is energy savings. In all experiments with the ADS4149 evaluation module, an external 3.3V supply and a common analog power rail were used to power the LDO and switching regulators, respectively. Table 1 shows the measured power efficiencies, along with their corresponding quiescent currents. This comparison shows that the LDO almost always consumes much more power than the ADC. The switching regulator consumes only 32mW more power than an ideal solution, resulting in an efficient power supply design. When the input voltage is stepped down (from 3.3V to 2.5V or 2.2V), the efficiency of the LDO can be further improved, but at the expense of higher system cost and larger size.

Although a DC/DC converter design with a lower dropout voltage requires more external components, its overall footprint can be smaller because newer DC/DC converters have higher switching frequencies, which greatly reduces the size of the inductor, for example, 33μH is needed at 500kHz, while only about 2.2μH is needed at 2.25MHz.

Table 1: Comparison of converters

Conversely, linear regulators may not require power filtering, but they also have size limitations because they typically dissipate more power. From a cost perspective, switching regulators may cost slightly more because of the higher component count. However, the increased efficiency can offset the cost of cooling techniques and the system power budget (References 3 and 4).

As system designers seek more power efficient components, significant energy savings can be achieved by changing the power architecture in high-speed data converter designs to switching regulators. You can directly power a low-power, high-speed data converter with a switching regulator without significantly degrading its AC performance.

Reference address:Powering High-Speed ​​ADCs with Switching Regulators

Previous article:Design Tips for Low-Power, Low-Cost Wireless ECG Monitors
Next article:Using the boost regulator above the rated voltage

Recommended ReadingLatest update time:2024-11-16 17:33

About the scanning mode of STM32 ADC
The screenshot below is the information about the ADC status register ADC1_SR in the STM32 reference manual that I read.   After referring to the information on the Internet, I feel that the EOC bit   seems to be set when each channel conversion is completed. Instead of being set when the channel group convers
[Microcontroller]
STM32f030c8t6 MCU ADC analog watchdog use
According to the description provided in the official manual of ST, it is not difficult to see that the ADC analog watchdog is actually that the user can set an upper and lower limit of the AD value, and then turn on the analog watchdog interrupt switch, so that when the IO input conversion AD value is detected to be
[Microcontroller]
STM32ADC multi-channel DMA transmission
ADC multi-channel acquisition is written based on ADC single-channel DMA transmission. The code is as follows: volatile u16 adcconverdata ={0,0}; static void ADC_GPIO_Config(void) {   GPIO_InitTypeDef  GPIO_InitStructure; RCC_APB2PeriphClockCmd(RCC_APB2Periph_GPIOA, ENABLE); GPIO_InitStructure.GPIO_Pin=GPIO_Pin_1|GPIO
[Microcontroller]
The MAX11102 family of low-power, successive-approximation analog-to-digital converters (ADCs)
MAX11102/MAX11103/MAX11105/MAX11106/MAX11110/MAX11111/ The MAX11115/MAX11116/MAX11117 are compact, high-speed, low-power, successive-approximation 12-/10-/8-bit analog-to-digital converters (ADCs). These high-performance ADCs feature very high sample/hold dynamic range and a high-speed serial interface
[Analog Electronics]
The MAX11102 family of low-power, successive-approximation analog-to-digital converters (ADCs)
Industrial measurement application of domestic dual-channel high-reliability Sigma-Delta ADC
ADC is the bridge connecting the real world and the digital world. Because it involves high technical barriers such as processes, circuits, and systems, its design is extremely difficult and it is known as the jewel in the crown of analog circuits. In order to precisely focus on the field of industrial measurement,
[Test Measurement]
Industrial measurement application of domestic dual-channel high-reliability Sigma-Delta ADC
[Lianshengde W806 Hands-on Notes] 10. ADC
Windows 10 20H2 HLK-W806-V1.0-KIT WM_SDK_W806_v0.6.0 Excerpted from "W806 Chip Design Guide V1.0", "W806 MCU Chip Specification V2.0" ADC The acquisition module based on Sigma-Delta ADC integrates 4-channel 16-bit ADC, completes the acquisition of up to 4 analog signals, or two differential signals. The sampling r
[Microcontroller]
[Lianshengde W806 Hands-on Notes] 10. ADC
STM32 ADC accuracy description
All the following descriptions are from the "STM32F103xx Data Sheet" (7th edition, May 2008). There is no new information here, just a translation of the corresponding part of the data sheet. When reading the various parameters in this data sheet, you must first understand how these parameters are obtained. Accordin
[Microcontroller]
STM32 ADC accuracy description
Application of 24-bit Σ-Δ analog-to-digital converter CS1240 in electronic scales
  CS1240 is a precision analog-to-digital conversion (ADC) chip developed in China with a resolution of 24 bits and an effective accuracy of up to 21 bits. It can be widely used in industrial process control, electronic scales, gas/liquid detectors, blood meters, etc. kind of application. This article describes the ma
[Analog Electronics]
Latest Power Management Articles
Change More Related Popular Components

EEWorld
subscription
account

EEWorld
service
account

Automotive
development
circle

About Us Customer Service Contact Information Datasheet Sitemap LatestNews


Room 1530, 15th Floor, Building B, No.18 Zhongguancun Street, Haidian District, Beijing, Postal Code: 100190 China Telephone: 008610 8235 0740

Copyright © 2005-2024 EEWORLD.com.cn, Inc. All rights reserved 京ICP证060456号 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号