标签: ir

Low-voltage design gets most of the attention these days, but there are many applications which require very-high voltages even though they do not deliver significant amounts of current to the load.   It's easy to think that almost "everyone" is doing low-voltage designs with power-stingy, battery-operated circuits -- but that's a simplistic and myopic perspective. There are well-known exceptions to the low-power world in applications which must deliver significant power to a load, such as a heater or motor. In those situations, using higher voltages allows use of lower currents for a given power rating (P = V × I), while minimizing IR loss and I2R heating. In these cases, current ratings in tens or hundreds of amps are common.   But it’s not all about higher voltages when it comes to reducing current, even though the current may still be in the tens or hundreds of amps . There are many unavoidably high-voltage situations which are also fairly low current, often under 100 mA. This became fairly evident when I walked the exhibit floor at the recent fall meeting of the Materials Research Society , the world's leading scientific organization for researching, developing, and applying new and existing materials. Approximately 6,800 attendees explored nanomaterials, ultrapure materials, cryogenics and ultravacuum chambers, high-temperature furnaces, sputtering, vapor deposition, specialized instrumentation, and much more.   We routinely and somewhat casually rely on many materials-related technologies which enable our hi-tech advances and innovations; it's a synergistic relationship, of course, as these advances in turn drive new materials. Beyond the development of these materials, there are major issues in determining their electrical, physical, and chemical properties. After all, once you have created that amazing ultrapure nanomaterial, how do you measure its hardness?   Why the need for the high voltages and low currents, as seen at many exhibits at the MRS event? These systems and their instrumentation are not "power devices" in the conventional sense, and minimizing IR loss and I 2 R dissipation is not the primary concern. The need is simple: it’s the law of physics. These systems require high voltages to steer electron beams, attract and accelerate particles, and change the energy state of atoms. I saw many specialty vendors whose supply product lines began at 10 kV, as well as many high-voltage supplies embedded within highly specialized analysis, fabrication, and measurement systems. Consumers also have a need for voltages in the 10 kV range, to power the magnetron in their microwave oven, and in past times, for the venerable CRT of now-obsolete television displays.   The standard home microwave oven has a high-voltage supply, usually between 10 and 20 kV, to energize the 2.4 GHz magnetron inside; would knowing that detail scare the consumer? SOURCE: Wikipedia   For designers who have little or no exposure to high-voltage/low-current design, it's a very different world. There's little need to minimize IR loss by using heftier connectors and conductors, since voltage drop is not a primary concern at these low currents. Instead, it's a world of thin conductors, thick insulation, safety interlocks, and mandated minimum physical spacing between conductors and anything nearby. It's also an unforgiving world where marginal design, inadequate attention to tiny details, and microscopic cracks in insulation can have dangerous consequences for equipment and users.   Nothing is done quickly, easily, or casually: voltage/current monitoring, probing with test equipment during debug or repair, and designing for user access must all take into account high voltages and its tendency to go through any breach in system physical integrity. Seeing a 25-kV rail cause spark-over in an poorly placed capacitor is an experience you won't forget (who, me?). Every component must have appropriate ratings and certifications, and even something as normally mundane as insulation material, thickness, and breakdown rating is a concern. Further, any tiny nicks in the insulation which may occur during prototyping and debug stages, or in production manufacturing are also a concern. Routine electrical insulation and standard electronic signal-isolation techniques and components are neither routine nor standard.   Designers who work in low-voltage, low-current domain have the luxury of being able to route power as they want to and make changes as needed to satisfy PC board-layout and cable-harness priorities; you can rout a 3.3V rail (on the PC board or in a cable) pretty much wherever you wish. In high-voltage designs however, that degree of freedom doesn’t exist and any change in routing must be carefully assessed to make sure it doesn't violate appropriate design guidelines or numerous regulatory standards for placement, creepage, clearance, and safety. Low-voltage designers don't have to worry that the basic design or any change in it will trigger safety-related regulatory testing and approval cycles.   Have you ever been directly involved with high-voltage/low-current design? Have you ever worked on a project which had that aspect? If so, what intrigued, impressed, worried, or scared you the most?

作者：飞兆半导体Joyce Patrick 和Robert Krause 光电晶体管光耦合器自从 40 年前面世以来就不断演进，目前仍然广泛使用。 目前应用最广的技术是使用砷化镓（GaAs）材料的红外发光二极管（IR LED）。 它们在许多应用中都能发挥很好的性能，但在低电流时效果不佳，因此，对于那些使用较低电流以提高发光效率的系统，这是一个问题。 飞兆半导体已采用一种不同类型的 IR LED，材料为铝镓砷化物（AlGaAs），有助于解决这个问题。 AlGaAs IR LED 具有更快的速度，改进的温度性能和一致性，并且与常规 GaAs 材料相比，在较低电流时更加高效。 FODM8801 OptoHiT™ 系列采用了新型的 IR LED，而且该系列可保证 LED 电流低至 1mA 时各个温度下的规格不变。 LED 发光效率的长期和短期性能 所有 IR LED 都要考虑其长期性能，无论材料是 GaA 还是 AlGaAs。 在应用过程中，LED 性能开始退化，导致 LED 发光效率长期永久性降低，这通过电流传输比（CTR）来衡量。 幸运的是，当 IR LED 在低电流下工作时，长期性能退化将降至最低。 这并不罕见，例如，在低电流下长期 CTR 变化只有 (IF 5mA) 每 1,000 小时 0.1%，这种情况下，性能退化不会造成什么影响。 虽然如此，问题是，工作温度的上升会对短期 LED 发光效率造成影响。 当结温上升时，CRT 会大幅下降，问题会很快出现。 在短短 5 分钟内，(LED)周围温度可能上升 50°C。 这种效应常常被忽视，因为当周围温度恢复到原来的起始温度时，CTR 也恢复到其初始值。 （大多数数据表仅仅指定在室温下工作。） 飞兆半导体的新型 AlGaAs IR LED 不易受到温度变化影响。 图 1 对比了 AlGaAs 880 nm IR LED 与标准的 GaAs 940 nm IR LED，并显示了在结温上升时，光输出如何变化。 (该图显示了典型的数据，但不能保证。) 图 1. 光输出减少（%/°C）与 LED 电流 在 1mA 下工作时，GaAs IR LED 光输出减少 1%/°C。 这意味着结温升高 50 °C 时，LED 的输出会降低 50%。 使用 AlGaAs IR LED，光输出仅降低 0.225%/°C，因此，结温升高 50°C 而光输出仅降低 11%。 光电晶体管速度与 LED 光电流 长期性能中另一个要考虑的因素是光电晶体管的速度。 光电晶体管的速度是由 LED 产生的光电流、晶体管电流增益（hFE）和负载情况所决定的。 随着 LED 光通量下降，时间也会变化。 如果一个设计在低电流时的温度范围内工作良好，则长期的变化将是最小的。 一般情况下， AlGaAs LED 的发光效率是同类 GaAs 产品两到三倍，温度稳定性是 4.5 倍。 使用 AlGaAs 材料制成的 IR LED 在更低的电流和更低的温度下支持的时间更长。 飞兆半导体的 FODM8801 OptoHiT 系列采用专有 OPTOPLANAR® 共面封装技术进行封装，这有助于提高低电流下的性能。 图 2 展示了 OPTOPLANAR 技术的截面图。 图 2.飞兆半导体 OPTOPLANAR® 共面封装技术 该封装可提供低输入-输出电容（CIO），相比采用面对面封装结构，其电容要低 30%，即使诸如半节距微型扁平封装（MFP）之类的小型封装也是如此。 绝缘内部厚度仅为 0.4mm，提供了高度的绝缘鲁棒性。 从而设计师无需牺牲隔离/绝缘性来减小封装尺寸。 结论 飞兆半导体的 FODM8801 OptoHiT 系列产品是新一代的光电晶体管光耦合器，配有 AlGaAs 工艺技术制造的 IR LED，能在更低电流和更低温度下更持久、更可靠地工作。 这让设计师能够更轻松地设计即使在极端严酷的条件下仍能正常工作的系统。

  转：http://hi.baidu.com/rfic_seu/blog/item/c2d8ec8e7a805208b21bba95.html 芯片电源网络上的IR drop按照形成原因可以分成 静态 和 动态 两种.IR drop现象 影响了标准单元的电源电压,使得标准单元不能很好的工作 . 静态IR drop VDD电压的静态IR drop现象产生的原因 主要是电源网络的金属连线的分压 ,是由于金属连线的自身电阻分压造成的.电流经过内部电源连线的时候,根据欧姆定律产生电源压降.所以静态IR drop主要跟电源网络的结构和连线细节有关,比如:金属连线的宽度,金属连线所用层,该路径流过的电流大小,尤其需要注意的是通孔的个数和打孔的位置. 动态IR drop VDD电压的动态IR drop是电源在电路开关切换的时候电流波动引起的电压压降.这种现象产生在时钟的触发沿,时钟沿跳变不仅带来自身的大量晶体管开关,同时带来组合逻辑电路的跳变,往往在短时间内在整个芯片上产生很大的电流,这个瞬间的大电流引起了IR drop现象.同时开关的晶体管数量越多,越容易触发动态IR drop现象.比如高扇出的一些结构,同一时间需求大量电流, 再比如扫描测试电路,扫描链非常大也容易引起这个现象.这个在设计的时候,电路结构中应该尽可能避免掉.后端设计的时候能做的就是给出足够强壮的电源网络,给出足够大的裕度. IR drop现象在芯片设计中的影响 如果通过电源网络到达某flip flop的电源压降IR drop太大, 这个flip flop可能无法正常工作. 最直接的影响是由于标准单元看到的电源假如是VDD’ 低于VDD,那么根据MOS模型这个单元的transition将变大,响应速度变慢.通常,如果是在clock时钟路径上产生IR drop会带来hold-time错误,如果在数据路径上的IR drop压降会带来setup错误. 一般情况下,5%的电源压降会增大10%-15%的线延迟.另外,还有可能引起信号完整性方面的问题,因为一个弱的电源导致弱的信号,弱的信号更容易被噪声影响甚至淹没. 但是,我们知道IR drop这种现象是每个芯片都会有的(我们不可能所有连线都用超导材料对吧? ^_^)所以,面对IR drop其实大家只能是尽可能去减小这种影响, 尽可能让IR drop所产生的影响不至于对信号完整性和时序产生不良影响. IR drop的兄弟,EM现象 EM是电子迁移现象. 作为特定的工艺来说,某个金属层会有一个最大电流密度,在这层金属上允许的最大电流密度是有一定限度的,过大的电流会造成金属断裂. 电流经过金属层时,电子和原子可能会碰撞,电流过大这种碰撞会产生热效应,同时热效应会加剧原子和电子的碰撞. 某种意义上来说,EM效应也有热效应的影子.这种大电流的情况下,不仅仅是电子的移动,有可能会在金属层造成阳极原子的迁移,这种迁移造成了金属连线的破损断裂,同时也可能会顺着迁移方向短路到其他线路. IR drop 的分析 对芯片电源进行有效的设计必须理解IR drop的产生原理. 静态IR drop和动态IR drop之间没有什么必然的联系. 现在的芯片设计中,电源设计和IR drop分析被不断的前提,应为深亚微米之后的电源压降现象更严重,早日分析早日发现问题并在设计中考虑到是很有必要的.在后端设计中,布局之时就应该考虑电源的规划,快速布局之后就应该对电源设计做一次分析. 到Sign-off阶段再做更详细的分析:功耗\压降\电子迁移\热. 做静态分析需要用到的是TCF(toggle count format)文件；做动态功耗分析需要知道两个重要的信息:电源点的位置(一般是PAD,也有的是用片内Regulator等)和动态仿真所用的文件VCD(Value change dump)或者TWF(timing window format)文件. 电源分析部分所用的工具一般是以Apache公司的Totem和其旗舰产品Redhawk来做. 其中Redhawk被很多设计公司作为Sign-off工具. 转：http://blog.csdn.net/wen8201/article/details/4284329 因为U=IR,所以IR-drop顾名思义就是压降。其危害有： 1。性能（performance） 由管子的Tdelay＝c/u可知，电压降低，门的开关速度越慢，性能越差。 2。功能(function） 实际上在极端的情况下甚至功能也会受影响的。在深亚微米下，如果Power/Ground network做的也很差，然后碰上了很不好的case，IR drop会很大，如果用的是high Vt的process，则DC noise margin就比较小了。这样就有可能功能错误。 3。功耗（power） 如果没有做详细的IR drop分析，又想功能正确，那就只有留很大的margin了，本来1.2v可以跑的，也只能用1.5v了。但是这样功耗也就上去了。 4。面积(area) 如果要在一定程度上限制IR drop，就要在chip里面加上很多的decoupling capacitance.占用了很多面积。 5。成本(cost) 功耗上去了，响应的散热，封装都成了问题，需要额外花费啦。而且面积变大，也是钱啊～～ 所以，IR drop还是一个比较讨厌的问题，需要小心对待。