标签: regulator

       在线性稳压器环路中，输出电容和后续的负载形成了一个极点，此极点造成了稳压器环路的不稳定。为了补偿这个极点，一般做法是用输出电容的ESR来产生一个零点。所以，在选在稳压器的输出电容时，ESR是一个需要考虑的方面。       处于成本的考虑，在实际设计应用中常常将 多层陶瓷电容用作稳压器的输出端电容，而MLCC的特点就是超低的ESR值，这将引起输出电压的不稳定。     以安森美NCV4949来讲，在其输出端加一个10uF的陶瓷电容，在电容后端测量到的瞬态响应如图(AC mode)：     同时测算的其phase margin只有5度（测试方式见此 http://bbs.ednchina.com/BLOG_ARTICLE_3026685.HTM ）。远远达不到稳定的标准。 查看其规格书，发现该稳压器对ESR的要求如下： 在负载30mA的情况下，其ESR要求在0.1～15ohms之间。而MURATA的10uF 1206 X7R MLCC的ESR查的为远小于0.01ohm：   在上图电路中10uF电容上串联一个1ohm电阻，重新测量输出的瞬态响应   同是算得phase margin已71度。满足稳定标准来。  

It's such a shame that low-dropout regulator (LDO) just doesn't get a lot of respect. Designers who should know better automatically assume that the switching DC/DC regulator is the regulator of choice, because "everyone" knows that the LDO is inefficient. Despite this somewhat-undeserved negative reputation, there are at least 30 credible LDO vendors, offering well over 1000 truly distinct LDOs (not just temperature/performance/package variations of a given model number), and billions are shipped every year in both fairly old and very new designs. Why the contradiction between LDO image and LDO reality? Depending on the line, load, and usage patterns, the LDO may have overall efficiency comparable to the switcher, or at least "good enough" for the applications. After all, not every application's must-have list is totally dominated with concerns about saving every milliwatt. In exchange, the LDO gives you a lot: simplicity in use, good behaviour with various loads, little need for external passives, low noise, good transient response, and more. Perhaps you are wondering, "What do you specifically mean by that word 'more,' anyway?" It turns out that LDOs can be designed to handle harsh, high-temperature environments and still meet some pretty good specs. Two recent introductions demonstrate this aspect of LDOs clearly. The TPS7H1101-SP and TPS7H1201-HT LDOs from Texas Instruments (see figure below) deliver up to 3 A and 0.5 A, respectively—no big deal, so do plenty of other available LDOs. But the 3-A unit is QML Class V qualified for operation up to 125°C, targeting military, satellite, and undersea cable applications; the 0.5-A device is qualified to 210°C, aiming at downhill-drilling needs.   In addition, due to their load-sharing-compatible outputs, each of these LDOs can be paralleled to yield double their individual current outputs, thus reaching 6 A and 1 A (depending on which device you are using). Input range is 1.5 to 7V and dropout is just 75 mV, with output noise of 17µVrms. It's not just TI announcing high-temperature LDOs, either. Linear Technology just introduced the LT1965, an adjustable 1-A unit for operation to 150°C, and just 40µVrms noise (see figure below). This device has an input range of 1.8 to 20 Vdc, and a resistor-settable Vout of 1.9 to 19.5 V, along with 300 mV typical dropout. To counter the bad reputation for low efficiency of LDOs, the regulator also can be put into a shutdown mode for less than 1µA dissipation.   Before you assume that the LDO is yesterday's solution to today's design problem for tomorrow's products, take a look at the huge assortment of LDOs offered by the many reputable vendors. You may be surprised to find that the virtues of the LDO are what you need, while their apparent primary shortcoming of low efficiency is not as bad you think—and you may also get high-temperature performance that is harder to achieve, or requires a more complex, and thus costly, switching architecture.

Pity the poor low-dropout regulator (LDO) as it just doesn't get a lot of respect. Designers who should know better automatically assume that the switching DC/DC regulator is the regulator of choice, because "everyone" knows that the LDO is inefficient. Despite this somewhat-undeserved negative reputation, there are at least 30 credible LDO vendors, offering well over 1000 truly distinct LDOs (not just temperature/performance/package variations of a given model number), and billions are shipped every year in both fairly old and very new designs. Why the contradiction between LDO image and LDO reality? Depending on the line, load, and usage patterns, the LDO may have overall efficiency comparable to the switcher, or at least "good enough" for the applications. After all, not every application's must-have list is totally dominated with concerns about saving every milliwatt. In exchange, the LDO gives you a lot: simplicity in use, good behaviour with various loads, little need for external passives, low noise, good transient response, and more. Perhaps you are wondering, "What do you specifically mean by that word 'more,' anyway?" It turns out that LDOs can be designed to handle harsh, high-temperature environments and still meet some pretty good specs. Two recent introductions demonstrate this aspect of LDOs clearly. The TPS7H1101-SP and TPS7H1201-HT LDOs from Texas Instruments (see figure below) deliver up to 3 A and 0.5 A, respectively—no big deal, so do plenty of other available LDOs. But the 3-A unit is QML Class V qualified for operation up to 125°C, targeting military, satellite, and undersea cable applications; the 0.5-A device is qualified to 210°C, aiming at downhill-drilling needs.   In addition, due to their load-sharing-compatible outputs, each of these LDOs can be paralleled to yield double their individual current outputs, thus reaching 6 A and 1 A (depending on which device you are using). Input range is 1.5 to 7V and dropout is just 75 mV, with output noise of 17µVrms. It's not just TI announcing high-temperature LDOs, either. Linear Technology just introduced the LT1965, an adjustable 1-A unit for operation to 150°C, and just 40µVrms noise (see figure below). This device has an input range of 1.8 to 20 Vdc, and a resistor-settable Vout of 1.9 to 19.5 V, along with 300 mV typical dropout. To counter the bad reputation for low efficiency of LDOs, the regulator also can be put into a shutdown mode for less than 1µA dissipation.   Before you assume that the LDO is yesterday's solution to today's design problem for tomorrow's products, take a look at the huge assortment of LDOs offered by the many reputable vendors. You may be surprised to find that the virtues of the LDO are what you need, while their apparent primary shortcoming of low efficiency is not as bad you think—and you may also get high-temperature performance that is harder to achieve, or requires a more complex, and thus costly, switching architecture.  

Sometimes, the deluge of new ICs seems mostly like an outpouring of pretty much the same parts, albeit with "more and faster" attributes in the form of more memory, more I/O, faster clocks, and so on, or perhaps slight enhancements in specifications. However, that's not the case in the world of power-management and regulator ICs, that's for sure. Unlike components which strive to be solutions to a somewhat broad class of problems, or parts which have achieved amazingly widespread use and longevity (think of basic LDOs or the ubiquitous 2N2222 transistor), most of the power-related ICs that I see are carefully targeted at well-defined application niches. They often have small but essential features and functions added, which make them a really good fit while solving a circuit/system designer's nasty problem. (Marketers often call these parts "ideal," but no part is ideal; I prefer to think of them as "well suited.") As a result, we have literally thousands of power-control ICs to choose from—and sometimes, having so many to look at can be overwhelming. But if you work your way through the various selection guides and parameter/specification tables, you'll see there is a reasonably justified rationale for almost each one. Some recent introductions make the case. For example, look at the LT3795 110V LED DC/DC converter and dimming controller from Linear Technology Corp. Sure, it's efficient, but that itself is not newsworthy these days. What makes this IC especially application friendly is that it adds the features you'd want in such situations, such as robust short-circuit protection and user-selectable spread-spectrum modulation to spread radiated energy and reduce EMI compliance issues and complaints (I have some LED and CFL lamps at home that are high-performance wideband EMI sources). Then there's an IC from Texas Instruments that is somewhat similar to the LTC part at first glance, yet is actually quite different, the TPS92690 . This DC/DC LED driver also has adjustable switching frequency to minimise EMI, but it is designed for automotive headlamps and illumination, so it has nothing to do with the AC line, and instead operates from 4.5 to 75.0-Vdc inputs. Further, it employs low-side current sensing, which is more compatible with some specialised LED power-driver topologies used in this application. Targeting a very different situation, Microchip Technologies has new members in its UCS100X family of programmable USB port power controllers. Along with 12 W charging (2.5 A), the UCS100X supports active cables, such as the Apple Lightning connector. The UCS1002 device will automatically charge a wide variety of portable devices, including USB-IF BC1.2, YD/T-1591 (2009), most Apple and RIM units, and many others. It comes with nine preloaded charger emulation profiles to maximise compatibility coverage of these peripheral devices—that's impressively thoughtful. These are just a few recent examples; there are many more you could find and highlight. In each case, they are associated with the "power control" function, yet they are quite different in what they do. This diversity makes an engineering and selection challenge, but it is one that is worth the effort to solve. Ironically, it also makes the job of the various research organisations that estimate the market size and changes for these types of ICs very foggy and fuzzy, so to speak. Have you seen any power control and management ICs lately that really seemed innovative with clever features and functions, rather than just having better specifications or improved efficiency?  

Sometimes, the new ICs seem to be composed of pretty much the same parts, albeit with "more and faster" attributes in the form of more memory, more I/O, faster clocks, and so on, or perhaps slight enhancements in specifications. However, that's not the case in the world of power-management and regulator ICs, that's for sure. Unlike components which strive to be solutions to a somewhat broad class of problems, or parts which have achieved amazingly widespread use and longevity (think of basic LDOs or the ubiquitous 2N2222 transistor), most of the power-related ICs that I see are carefully targeted at well-defined application niches. They often have small but essential features and functions added, which make them a really good fit while solving a circuit/system designer's nasty problem. (Marketers often call these parts "ideal," but no part is ideal; I prefer to think of them as "well suited.") As a result, we have literally thousands of power-control ICs to choose from—and sometimes, having so many to look at can be overwhelming. But if you work your way through the various selection guides and parameter/specification tables, you'll see there is a reasonably justified rationale for almost each one. Some recent introductions make the case. For example, look at the LT3795 110V LED DC/DC converter and dimming controller from Linear Technology Corp. Sure, it's efficient, but that itself is not newsworthy these days. What makes this IC especially application friendly is that it adds the features you'd want in such situations, such as robust short-circuit protection and user-selectable spread-spectrum modulation to spread radiated energy and reduce EMI compliance issues and complaints (I have some LED and CFL lamps at home that are high-performance wideband EMI sources). Then there's an IC from Texas Instruments that is somewhat similar to the LTC part at first glance, yet is actually quite different, the TPS92690 . This DC/DC LED driver also has adjustable switching frequency to minimise EMI, but it is designed for automotive headlamps and illumination, so it has nothing to do with the AC line, and instead operates from 4.5 to 75.0-Vdc inputs. Further, it employs low-side current sensing, which is more compatible with some specialised LED power-driver topologies used in this application. Targeting a very different situation, Microchip Technologies has new members in its UCS100X family of programmable USB port power controllers. Along with 12 W charging (2.5 A), the UCS100X supports active cables, such as the Apple Lightning connector. The UCS1002 device will automatically charge a wide variety of portable devices, including USB-IF BC1.2, YD/T-1591 (2009), most Apple and RIM units, and many others. It comes with nine preloaded charger emulation profiles to maximise compatibility coverage of these peripheral devices—that's impressively thoughtful. These are just a few recent examples; there are many more you could find and highlight. In each case, they are associated with the "power control" function, yet they are quite different in what they do. This diversity makes an engineering and selection challenge, but it is one that is worth the effort to solve. Ironically, it also makes the job of the various research organisations that estimate the market size and changes for these types of ICs very foggy and fuzzy, so to speak. Have you seen any power control and management ICs lately that really seemed innovative with clever features and functions, rather than just having better specifications or improved efficiency?