DC_DC 变换的PCB layout 注意事项

合集下载

pcb layout流程和注意事项

pcb layout流程和注意事项

pcb layout流程和注意事项下载温馨提示:该文档是我店铺精心编制而成,希望大家下载以后,能够帮助大家解决实际的问题。

文档下载后可定制随意修改,请根据实际需要进行相应的调整和使用,谢谢!并且,本店铺为大家提供各种各样类型的实用资料,如教育随笔、日记赏析、句子摘抄、古诗大全、经典美文、话题作文、工作总结、词语解析、文案摘录、其他资料等等,如想了解不同资料格式和写法,敬请关注!Download tips: This document is carefully compiled by theeditor.I hope that after you download them,they can help yousolve practical problems. The document can be customized andmodified after downloading,please adjust and use it according toactual needs, thank you!In addition, our shop provides you with various types ofpractical materials,such as educational essays, diaryappreciation,sentence excerpts,ancient poems,classic articles,topic composition,work summary,word parsing,copy excerpts,other materials and so on,want to know different data formats andwriting methods,please pay attention!PCB Layout流程详解及设计注意事项PCB(Printed Circuit Board)布局是电子设计中至关重要的一步,它直接影响到电路的性能、可靠性和生产成本。

自己总结的电源板Layout的一些注意点

自己总结的电源板Layout的一些注意点

电源板Layout注意点做了几年的电源板的layout,自己总结了一些主要注意的地方,我认为主要是从下面这几个地方考虑:一、 功率回路部分功率板中比较重要首当其冲的就是功率回路部分,在layout的时候应该首先要知道所布的功率部分的电路性质,在电源中功率电路主要分di/dt电路和dv/dt电路,这两种电路在布局走线的时候走法是不一样的。

di/dt电路因为它的单位时间内电流的变化比较大,所以这部分电路在走线的时候重点这样电路的环路面积最小,本身产生的干扰可以自身就耦合掉dv/dt电路它的侧重点就完全不一样,因为这种电路在单位时间内电压变化会比较大,所以它容易对外界产生干扰,所以这种电路在走线的时候铜皮不能太宽,在满足承载电流的情况下铜皮宽度尽可能的小,不同层的重叠区域尽可能小,敏感信号尽可能远离这些走线二、 驱动部分驱动部分的线首先要考虑整个驱动回路的面积,要尽可能的小,要远离干扰源,离被驱动的部分尽可能的近。

像MOS管之类工功率元件的驱动,在走线的时候要特别注意G极和D极的走线不要平行走,因为在大多数情况下MOS管的D极部分的电路是dv/dt的电路,G极是驱动电路,如果平行走的话,驱动信号很容易被干扰,从而导致MOS的误动作三、 采样信号在功率板中像一些电压采样和电流采样之类的采样信号也是至关重要的,因为这些信号准确与否直接关系到控制端,所有这些采样信号也要尽量避开其他信号,如果有条件的话这些采样信号可以用差分采样,并且在相对应的走线地方能够给他们一个完整的地平面四、 地的处理地的重要性就更不用说了,无论在哪种板子上,对于地的处理都是非常重要的。

在功率板中地相对来说会比较复杂,因为很多时候功率部分走大电流的地、控制部分一些小电流的地都是共地的,所以这时候这些地的处理显的非常重要,在我的经验中处理好这些地,关键是选择一个正确的单点连节点,因为每个电源的设计不一样,所以这个单点连接点的选择也是不一样的,我在小功率光伏逆变器中一般都是选择BUS电容的一个地管脚,变频器中我一般是将大电流中的一个电容的地管脚引一根比较粗的走线到开关电源输入端的那个电容的地管脚上,然后再从这个地管脚引到开关电源后面出来的那些小电流的地平面上,当然还有一些别的地,如晶振的地、采样的地等,每个公司的设计规则不一样,走法也不一样,网上对于地的处理的资料也比较多五、 安规安规在电源产品的设计中是不可或缺的,对于不同国家不同地区相应的安规法规要求也是有区别的,还有应用环境的污染等级和海拔高度都会对安规要抓的距离有比较大的影响,所有我们在设计之初一定要搞清楚上面这些因素,如果有安规工程师的话可以请他们给出比较专业的爬电和电气间隙的距离,我们实际PCB Layout的时候要特别注意那些金属元件在PCB上所在区域,比如保险丝,它两头是金属的中间是非金属,如果没有座子的话,保险丝的两头金属的会和PCB接触,所有保险丝周围的表层走线要注意避开这些金属区域六、 散热对于那些功率比较大的系统来说,散热也是至关重要的,这个一般情况下要和结构配合好,在设计之前要了解整体结构的散热方式,是自然冷却、风冷还是水冷,其中风冷又分吸风和吹风,这些都会对布局产生比较大的影响七、 E MC主要是一些功率部分的走线宽度尽量不要发生突变,如果需要拐弯,拐弯的地方也尽量做的平和一点,不要突变,还有就是有时候会有大电流、小电流、采样信号中有些虽然有些是共用一个网络,但是自在走线的时候不要 共用一个回路,要分开走,各走各的回路比较好。

PCBLAYOUT安规设计注意事项

PCBLAYOUT安规设计注意事项

PCBLAYOUT安规设计注意事项PCB(Printed Circuit Board)Layout的设计是电子工程师在电路设计中不可或缺的一部分。

PCB Layout的设计必须遵循一定的安规设计准则和注意事项,以确保最终产品的质量符合相关法规和标准,同时还要保证电路板能够正常工作。

下面将介绍一些PCB Layout的安规设计注意事项。

1. 防静电破坏静电对于电子元器件的损坏是十分严重的。

在PCB Layout 中,我们必须考虑如何减少静电破坏的风险,并确保PCB板及其上元器件不遭受静电损坏。

对于一些静电敏感的元器件,如场效应晶体管等,我们需要注意以下几点:(1)在装配元器件之前,要确保工作区域的接地系统得到有效的连接;(2)元器件需要使用袋式包装或者静电包装,确保元器件表面的防静电材料不受损坏;(3)在PCB Layout上,为防止静电累积,要合理安排元器件的布局,对那些静电敏感的部分,需要进行特殊处理。

2. 灵敏度和抗干扰能力在PCB Layout设计中,元器件的灵敏度和通信接口的干扰容忍度十分重要。

在光、磁、电场和射频辐射等电磁干扰的环境下,必要时需要采取一些措施来保证电路板的抗干扰能力。

例如,为了减少介质损失,一种方法是使用高频线路的微带线(microstrip lines)。

3. 温度和湿度电子元器件的温度和湿度对它们的性能和寿命都有很大的影响。

在PCB Layout设计中,我们需要考虑环境条件,并采取必要的措施来确保元器件长期稳定工作。

例如在元器件周围设置散热装置或者风扇,以保持元器件周围的温度。

这样可以有效降低元器件电阻和电容的漂移,同时还可以提高元器件的稳定性。

4. 接地和电源接地和电源设计是PCB Layout安规设计中很重要的一部分。

在接地设计中,应该遵循单点接地和保持最小全流接地的原则。

这种方法可以减少环路电流和降低噪声。

在电源设计中,需要考虑到电源稳定性和供电电流等因素。

5. 安全性和可靠性在PCB Layout安规设计中,需要考虑到电路板的安全性和可靠性。

DC-DC的PCB Layout之降压非同步型续流二极管

DC-DC的PCB Layout之降压非同步型续流二极管

DC-DC 的PCB Layout 之降压非同步续流二极管本文主要介绍降压型开关DC-DC 中非同步续流二极管的PCB Layout 问题 首先了解概念:开关型DC-DC 有同步型(Synchronous )和非同步型(Non-synchronous )两种,下图以Buck 型电路为例图1.1 非同步型Buck 图1.2 同步型Buck非同步型:1,效率偏低 同步型:1,更高效 2,成本较低,IC 外部使用续流二极管 2,成本更高,内部集成MOS 管 3,比较适宜较高的输出电压 3,MOS 管具有更低的导通压降 对于Non-Synchronous 型的降压DC-DC ,我们的续流二极管Layout 时怎么布局? 1,负端靠近SW 脚放置 √ 2,正端靠近输出电容放置 ? 续流二极管是在开关管关断后,对电感形成回路,为什么对续流二极管靠近输出电容放置存在疑问?那么正确的放置是怎样? 后面做详细分析,首先给出答案:续流二极管正端接靠近输入电容的地 理由:从交流路径分析,续流二极管正端接靠近输入电容的地,为最小辐射路径面积,有助于EMI 指标和性能。

分析如下:图2a图2b图2c图2a粗线标出了开关管打开时的高频电流路径,当开关管关闭后,形成图2b的交流回路,如此循环的在这两个回路切换。

那么就有了图2c,它是开关动作中电流会出现有到无,无到有的突变段。

电流突变会在线路的寄生电感上形成电压尖峰,大小也与寄生电感的值相关(Vl=L*di/dt)。

突变段受到电流基波的丰富谐波的上升或下降边影响,通常可以描述为突变段有“交流电流”流过,基本的开关PWM频率构成“交流”突变段中电流波形总谐波的一部分。

一般1盎司铜箔,2英寸长50mil宽走线的寄生电感值约50nH,如果走线有打过孔跨层,寄生电感值会再明显增加;开关管的开通关闭时间在十几到几十ns,比如EUP3490的开关管开通时间典型值为25ns,这样条件瞬态电流为2A来计算,产生的感应电压V=L*di/dt=50*2/25=4V。

LAYOUT应注意事项

LAYOUT应注意事项

LAYOUT應注意事項:1.如果兩個銲點之間,只走一條線,應儘量走在中間,以減少短路的機會。

2.繞線時,除非不得已的情況下,不要走90度角,容易造成斷裂。

3.繞完線後,儘可能使用淚滴,以增加線與銲點的接觸面,接觸面積愈大則線愈不容易斷裂。

4.繞線距離板邊,最少不要低於0.5MM,以免成型時將線截斷。

5.文字面避免放在銲點上面,將參考位置放在實體物面積之外。

6.注意FPC要折彎或擺動之處,必須儘量設計不要太硬,不要舖太多的銅,使其具有良好的耐折性。

7.導線的寬度:銅導線的寬度關係到耐電流和溫昇,所以盡量使用寬一點的導體較佳)。

通常信號用0.8mm寬,電源用1.5mm以上。

必要時可以加大或減小。

太細的線製作容易導致失敗。

8.焊點不要太小以免脫落,孔徑可以設成0.5mm以利鑽孔時的定位。

如果你技術好,可以直接設成要鑽孔的孔徑,這樣子銅箔比較不容易突起,但是相對鑽孔定位會差一點,要是鑽歪了,焊點內會有留白。

9:零件排列时各部份电路盡可能排列在一起,走线盡可能短。

10:IC地去耦电容应尽可能的靠近IC脚以增加效果。

11:如果两条线路之间的电压差较大时需注意安全间距。

12:要考量每条回路的电流大小,即发热状况来决定铜箔粗细。

13:线路拐角时尽量部要有锐角,直角最好用钝角和圆弧。

14:对高频电路而言,两条线路最好不要平行走太长,以减少分布电容的影响,一般采取顶层底层众项的方式。

15:高频电路须考量地线的高频阻抗,一般采用大面積接地的方式,各点就近接地,减小地线的电感份量,讓各接地点的电位相近。

16:高频电路的走线要粗而短,减小因走线太长而产生的电感及高频阻抗对电路的影响。

17:零件排列时,一般要把同类零件排在一起,盡量整齐,对有极性的元件盡可能的方向一致,降低淺在的生产成本。

18:对RF机种而言,电源部份的零件盡量遠离接收板,以减少干擾。

19:对TF机种而言,发射器应盡可能离PIR远一些,以减少发射时对PIR造成的干扰。

Layout(集成电路版图)注意事项及技巧总结材料

Layout(集成电路版图)注意事项及技巧总结材料

Layout(集成电路版图)注意事项及技巧总结材料Layout主要工作注意事项●画之前的准备工作●与电路设计者的沟通●Layout 的金属线尤其是电源线、地线●保护环●衬底噪声●管子的匹配精度一、l ayout 之前的准备工作1、先估算芯片面积先分别计算各个电路模块的面积,然后再加上模块之间走线以及端口引出等的面积,即得到芯片总的面积。

2、Top-Down 设计流程先根据电路规模对版图进行整体布局,整体布局包括:主要单元的大小形状以及位置安排;电源和地线的布局;输入输出引脚的放置等;统计整个芯片的引脚个数,包括测试点也要确定好,严格确定每个模块的引脚属性,位置。

3、模块的方向应该与信号的流向一致每个模块一定按照确定好的引脚位置引出之间的连线4、保证主信号通道简单流畅,连线尽量短,少拐弯等。

5、不同模块的电源,地线分开,以防干扰,电源线的寄生电阻尽可能较小,避免各模块的电源电压不一致。

6、尽可能把电容电阻和大管子放在侧旁,利于提高电路的抗干扰能力。

二、与电路设计者的沟通搞清楚电路的结构和工作原理明确电路设计中对版图有特殊要求的地方包含内容:(1)确保金属线的宽度和引线孔的数目能够满足要求(各通路在典型情况和最坏情况的大小)尤其是电源线盒地线。

(2)差分对管,有源负载,电流镜,电容阵列等要求匹配良好的子模块。

(3)电路中MOS管,电阻电容对精度的要求。

(4)易受干扰的电压传输线,高频信号传输线。

三、layout 的金属线尤其是电源线,地线1、根据电路在最坏情况下的电流值来确定金属线的宽度以及接触孔的排列方式和数目,以避免电迁移。

电迁移效应:是指当传输电流过大时,电子碰撞金属原子,导致原子移位而使金属断线。

在接触孔周围,电流比较集中,电迁移更容易产生。

2、避免天线效应长金属(面积较大的金属)在刻蚀的时候,会吸引大量的电荷,这时如果该金属与管子栅相连,可能会在栅极形成高压,影响栅养化层质量,降低电路的可靠性和寿命。

电源、DCDC电路原理设计及PCB布线注意事项大全.doc

电源、DCDC电路原理设计及PCB布线注意事项大全.doc

电源、DC-DC-电路原理设计及PCB布线注意事项大全注:本文内容摘抄整理自网络、论坛,仅供大家参考学习,谢谢!!!电源、DC-DC 电路原理设计及PCB布线注意事项大全一般的降压型的DC-DC变换的典型原理电路,如下图:一.DC-DC电路设计至少要考虑以下条件:1.外部输入电源电压的范围,输出电流的大小。

2.DC-DC输出的电压,电流,系统的功率最大值。

二.基于以上两点选择PWM IC要考虑:1.PWM IC的最大输入电压。

2.PWM开关的频率,这一点的选择关系到系统的效率。

对储能电感,电容的大小的选择也有一定影响。

3.MOS管的所能够承受的最大额定电流及其额定功率,如果DC-DC IC内部自带MOS,只需要考虑IC输出的额定电流。

4.MOS的开关电压Vgs大小及最大承受电压。

三.电感(L1),二极管(CR1),电容(C2)的选择1. 电感量:大小选择主要由开关频率决定,大小会影响电源纹波;额定电流,电感的内阻选择由系统功耗决定。

2. 二极管:通常都用肖特基二极管。

选择时要考滤反向电压,前向电流,一般情况反向电压为输入电源电压的二倍,前向电流为输出电流的两倍。

3. 电容:电容的选择基于开关的频率,系统纹波的要求及输出电压的要求。

容量和电容内部的等效电阻决定纹波大小(当然和电感也有关)。

如何得到一个电源纹波相对较小、对系统其他电路干扰相对较小,而且相对稳定可靠的DC-DC电路,需要对以上电路的原理做如下修改:1.输入部分:电源输入端需要加电感电容滤波。

目的:由于MOS管的开关及电感在瞬间的变化会造成输入电源的波动,尤其是在系统耗电波动较大时,影响更为明显。

2.输出部分:(1)假定C2的选择的100uF是正确的,我们想得到更小的纹波,可以将100u F的电容改成两颗47uF的电容(基于相同类型的电容);如果100uF电容采用的是铝电解,可以在原来的基础上加一颗10uF的磁片电容或钽电容。

(2)在输出端再加一颗电容和一颗电容对原来的电源做一个LC滤波,会得到一个纹波更小的电源。

pcb设计注意事项及设计原则

pcb设计注意事项及设计原则

pcb设计注意事项及设计原则
1. 注意电路的布局:将关键的电路元件和元件之间的连接线尽量短,并且按照电路信号流的路径进行布局,以降低电路的干扰和噪声。

2. 确保供电和地线的良好连接:供电和地线必须足够宽,以确保电流的充分通畅,同时尽量减少导线的长度和阻抗。

3. 保持信号的完整性:重要的高频信号和低噪声信号应该有独立的接线层进行隔离,并且保持信号线之间的最小交叉和最小输入/输出延迟。

4. 尽量减少板层数量:增加板层会增加制造成本和装配难度,因此应该尽量减少板层数量,并合理布局各种信号。

5. 为高功率模块提供散热解决方案:对于功率较大的模块,应该考虑合适的散热解决方案,如散热片、散热孔等。

6. 注意阻抗匹配:对于高速信号线,应该根据需求确定合适的阻抗,并尽量避免阻抗不匹配。

7. 考虑EMC问题:应该尽量减少电磁干扰并提高抗干扰能力,如采用合适的屏蔽、阻尼材料和接地。

8. 保证良好的可维护性:电路的布局应该考虑到维修和更换元件的方便性,如保留合适的测试点和备用元件位置。

9. 注意元器件的热分布:对于容易发热的元件,应该注意合适的散热和降温措施。

10. 使用规范的命名和标记:为了方便阅读和维护,应该使用规范的元件命名和标记方法,并为电路板添加清晰的标签和说明。

  1. 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
  2. 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
  3. 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。

Layout Considerations for Non-Isolated DC-DC ConvertersDC-DC converters are an excellent source of electric fields and magnetic fields. Their EMI spectrum begins at the switching frequency and often extends over 100MHz. To minimize capacitive couplings and magnetic couplings care must be exercised in printed circuit board (PCB) layout. Parasitic capacitance and parasitic inductance of the circuit must be evaluated so that the proper trade-off can be made early in the design phase.For many years, repeated introductions of integrated DC-DCpower-supply controllers have given us ever-higher levels of performance. These ICs unburden the systems engineer by removing the task of power-supply design, but this simplification has led to a loss of knowledge. Switching converters should therefore serve as a reminder to be careful. The following discussion presents rules for avoiding surprises when designing board layouts for non-isolated DC-DC converters.The first rule in optimizing such a layout is to isolate the converter. DC-DC converters are an excellent source of electric and magnetic fields. Their EMI spectrum begins at the switching frequency and often extends over 100MHz. To minimize capacitive couplings and "magnetic-field-to-loop" couplings, you should locate the converter away from other circuitry, especially from low-level analog circuitry. Isolating the converter is not always easy. Some boards accept input voltage on one side of the converter and distribute output voltages on the other side. VME cards or telecom cards, for example, include very complex routings with currents as high as 20A. A single connector brings in the input voltage and distributes several output voltages to the backplane. Therefore, there's a strong temptation to place the converter near this connector to reduce resistive drop. The area, however, is dense with interface drivers, backplane buses, and so forth, with the associated risk of noise coupling. A power connector can be added in some cases, but that solution entails extra board area and cost.Resistance in the copper traces is the most constraining factor. For a trace of a given length and thickness, this resistance iswhere l is the trace length in meters, S is the trace area in square meters, and (the resistivity of copper) is 1.7x10-8/m at 20°C, or 2.1x10-8/m at 70°C. As an example, the resistance of a 20°C trace 0.5cm wide and 35µm thick is 1m/cm. That value may appear negligible, but it commands attention if you are distributing 2.5V at 10A through two connectors and a backplane board.On some boards, the trace thickness includes a tin-lead layer. This layer can nearly double the equivalent resistance:lead resistivity = 2.07x10-7/m at 20°Ctin resistivity = 1.14x10-7/m at 20°CA trade-off between accuracy and trace loss lets you move the converter away from the connector. You can limit the effect of a resistive drop by performing a remote V OUT measurement near the connector, but beware of capactive couplings! To confine large currents to a defined area, route all the supply lines through pins at one end of the connector.MOSFET DriversAs switching frequencies increase, the switching time becomes shorter and shorter: typically 10ns for a 500kHz converter. At that frequency, even the shortest trace has a significant impedance. It's also important to note that a peak gate current can rise to several amperes in an extremely short time. Therefore, the proper routing of MOSFET-drive signals begins with an analysis of the converter's block diagram.For example, consider a synchronous step-down controller for notebook computers (Figure 1). The MOSFETs are driven by the transference of energy from the tank capacitors (C6 and C7) to the gates through the few ohms of the driver outputs. Note that the gate drive for the high-side n-channel MOSFET (Q1) is floating. Then-channel driver works like a charge pump!For larger image, click on figures.Figure 1. Operation of the MAX1710 synchronous step-down controller is depicted by an application circuit (a) and an internal block diagram (b). Figure 2 highlights the current paths during turn-on. Any series inductance can lead to disaster. In the best case, the spikes are higher, but simply increase the switching losses. In the worst case, the two MOSFETs can blow up due to cross conduction (simultaneous turn-on). Consequently, an optimal routing implies very short and wide traces between:C6 and V ddC6 and Q2(S)C7 and BST, and C7 and LXQ1(G) and DHQ2(G) and DLQ1(S) and LXQ2(S) and PGNDKeep in mind that the parasitic inductance for a 1cm trace is about 10nH.Looking closely at C6, you can see that it supplies Q1 and Q2, but not in the same way. It acts as a filter capacitor for Q1 and as a tank capacitor for Q2. Because we cannot place C6 near the high-side and low-side drivers at the same time, we place it as close as possible to V dd and PGND (where the peak currents flow), and also near C7 (almost average current). Notice that the PGND, DL, and V dd pins are side by side, and not by chance! Q2 and C6 are placed to minimize theground-trace lengths between PGND, C6(-), and Q2(S). Connect this ground trace to the ground plane at a single point, near the PGND pin. To avoid common-impedance coupling, LX should be connected to Q1, and PGND/C6(-) to the source of Q2. Figure 3 highlights the current paths during turn-off.Figure 2. The dotted lines indicate heavy current flow in the gate-drive circuits for Q1 (a) and Q2 (b).Figure 3. The dotted lines indicate heavy current flow in the gate-discharge (turn-off) paths for Q1 (a) and Q2 (b).The number of "via" should be limited as much as possible. Indeed,the few tens of nanohenries added by a via is embarrassing whendi/dt is high. For that reason, you should place all power components on the component layer, even the SMD ones. If you have no other choice, put several via in parallel.We must remember that controllers are often oversized for the application at hand. Common practice, for instance, employs a 10A controller to produce a 3A output. Because we generally choose minimum-sized MOSFETs for cost reasons, the on-chip drivers remain oversized and therefore capable of more gate drive than is necessary. Because the earlier discussion sought to avoid slowing the MOS gate drive, it seems paradoxical to place a small (10 to 100) resistor in series with the gate drive. Oversized and/or fast drive waveforms, however, produce more switching noise and RF interference. At the other extreme, slow waveforms produce more switching loss in the MOSFET(s) and diode (if any). A good compromise is to reduce EMI by slowing the waveform slopes as much as possible, while maintaining an acceptable level of efficiency. (Components in the gate drive of Figure 4b let you trim the rise and fall times separately.) Fortunately, large drivers allow a final bit of optimization.Routing the Power StageFigure 4 illustrates the two high-current loops common to many power converters. In responding to the perturbations caused by switching, these loops support high di/dt, and (at some nodes) high dv/dt as well. Identifying these loops helps to reduce their effects. Note that di/dt is great within the converter (at the switching node) but small outside the converter. Indeed, currents before the input capacitor and after the output capacitor reach a high value, but they are nearly continuous. Their AC components are low when the converter is well designed.Figure 4. These simple schematics illustrate the basic operation of the step-up (a) and step-down (b) switching converters.First, minimize parasitic inductance. We will consider a step-up converter, but the reasoning can be transposed for a step-down type. Figure 5 illustrates the kinds of parasitic inductance that cause the most problems.Figure 5. When the main switching transistor (T) turns off, the diode capacitance discharges as shown.Without describing the 10 phases of a switching cycle, we can consider the MOSFET turn-off, when inductor current has been short-circuited by the MOSFET. The diode's inverse capacitor chargesextremely fast via loop two, and the node-voltage Va at the diode's anode (normally at Vout - Vd) goes to near 0V. The serial inductors (LfT + LfD + LfC) increase this discharge time, thereby increasing switching loss in the MOSFET. These types of inductance also generate noise.Peak current is limited by the transistor, which operates as a current source (Vds still equals a few volts). For a 2A MOSFET, this current source could be 10A! The current level is large for a very short time (a few 10ns). Remember that varying the current through an inductor produces a voltage proportional to the current variation:This transition phase is a good spike generator! Once more, we minimize parasitic inductance by minimizing trace lengths and using short, wide traces around the MOSFET, diode, and Cout. You can now see how noise can be reduced by the control of slopes in thegate-drive waveforms.To limit resistive voltage drops and the number of via, power-stage SMD components should be placed on the component side of the board and power traces routed on its component layer. If possible, the power ground should also be routed on the same layer. This arrangement has another benefit: less perturbation of the ground plane. To cancel radiated fields, take care to minimize the area of the power-current loop.When it becomes necessary to route a power trace on a layer other than the component layer, choose a trace from the inductor or filter capacitors (i.e., Cout for a step-down converter, or Cin for a step-up converter). Because current through such traces is nearly continuous, it produces no noise, just a resistive drop. Parasitic inductance will be smaller if you route this trace on the layer just under the component layer. To avoid coupling by common impedance, you should separate PGND, the power-circuit ground, and the general ground plane (Figure 6).Figure 6. These details illustrate the routing of PGND versus gate-control traces in the controller circuit.Capacitors and Other ComponentsIt's important to pay close attention to the routing of traces from capacitor terminals in a DC-DC converter circuit. Large-valuedlow-ESR capacitors are expensive, and bad routings can cancel their performance. A good routing, on the other hand, can lower the output noise from 150mV to 50mV!Ripple is directly related to the inductor value, the capacitor ESR, the switching frequency, and so forth, but HF noise (spikes) depend on parasitic elements and the switching action. We can anticipate spiking frequencies from 1MHz to 10MHz, depending on the switching frequency.In a bad routing (Figure 7a), parasitic inductance associated with trace lengths causes trouble: L1 brings about an increase in noise, and L2 limits the attenuation of HF capacitor CoHF. The fix (Figure 7b) is to bring the input trace in on one side of the capacitor pad, and theoutput trace out on the other side of the pad.Figure 7. Improper routing of capacitor traces introduces unwanted parasitic inductance.Now, having placed and routed the bulkiest parts, we turn our attention to the inductor. A deliberate but unwanted coupling (Figure 8a) allows current in the power loop to pollute the controller supply (Vcc). High di/dt produced by switching through Lp1 causes Vcc overshoots that easily reach several hundred millivolts. Usingwith L = 10nH, I = 1A, and t = 50ns, V = 0.2V!As mentioned earlier, the first precaution is to separate the powertraces and be careful with PGND. The second precaution is to avoid connecting traces to the power loop (see Figure 8b). Traces that distribute the input voltage should be connected before the input capacitor and the controller's Vcc connection. Output voltage is distributed after the connections to the HF output capacitor.Figure 8. Unwanted common-impedance coupling (a) is prevented by the connections shown in (b).Finally, some miscellaneous advice: Bad routing of the PGND trace can cause common-impedance coupling (one reason for the PGND terminal is to avoid polluting the controller's internal ground node). Move high-impedance traces (especially the output resistor divider that adjusts the output voltage) away from nodes that support high dV/dt. Currents induced by such coupling can make the controller oscillate.z。

相关文档
最新文档