Understand_Capacitor_Soakage_to_Optimize_Analog_Systems(理解电容的浸润效应--翻译版)

1、Understand Capacitor Soakage to Optimize Analog Systems解电容器浸润优化模拟系统 ）Dielectric absorpti on can cause subtle errors in an alog applicati ons（在模拟应用中电解质吸收可能导致微妙的错误）such as those employing S/H circuits（例如使用采样保持电路）,integrating ADCs （集成ADC ）and active filters（ 和有源滤波器）.But knowing how to measure this soak

2、age（但是知道如何测量这些浸润）and compensate for it helps you minimize its effects（并且补偿它能帮助你减小它的影响 ）.Vetera n circuit desig ners ofte n got a shock ing in troducti on to dielectric absorpti on （老的电路设计者对电容器浸润经常有令人震惊的介绍）when supposedly discharged high-voltage oil-filled papercapacitors reached out and bit them（当很可

3、能是高压油浸纸电容器放电并且击穿电容）.1 ndeed, the oldoil-filled paper capacitors were no torious for what was once called soakage（实际上老的油浸纸电容器是因为" 浸润”而声名狼藉 ）-a capacitor's propensity to regain some charge after removal of amome ntary short（ 浸润电容在短路放电后，开路时倾向于恢复原有电荷（也就是电路中的"长尾"现象）.Today, you won'

4、t find very ma ny of these capacitors in use（今天你会发现这种电容应用的不是太多）,butyou will still en cou nter soakage（但是仍然会遭遇浸润 ）.Do you know how to deal with it?（你知道怎样解决吗？）Nowadays, you're more likely to notice the effects of dielectric absorption in some more subtle way（当今，你更可能在一些更微妙的方式下注意到电介质吸收的影响），perhaps i

5、n the performanee of anin tegrator that can't be reset to zero（积分器表现岀拒绝复位至0） or a sample/hold that refuses to workcorrectly（或采样保持器表现岀变化不定的误差）.But whether you literally feel its effects or merelyobserve them in a circuit's behavior（但是不管你真正感受到它的影响或是仅仅是观察它们在电路中的响应）,dielectric absorpti on is an

6、un desirable characteristic that every capacitor possesses to some degree（电介质吸收这种不良特征每种电容或多或少都会有）.This characteristic is in here nt in the dielectricmaterial itself（这种特征天生存在于电介质材料当中）,although a poor man ufacturi ng procedure （即使不良的制造工序）or in ferior foil electrodes can con tribute to the problem（或者劣质

7、箔电极都能导致这种问题）.Fig 1 - A simple test fixturelets you evaluate dielectric absorpti on at low speeds（一个简单的让你在低速下评估电介质吸收的测试固件）.To use the one shown here（举个例子）,start with all switches offand throw S 1 and S 2 on for 1 min（ 首先将所有开关打到 OFF并且S1和S2打到 ON 分钟）;throw S 1 and S2 off and wait 6 sec（ S1 和 S2 打到 OFF 并

8、且等待 6S ）, throwing S 3 on during the wait period（在等待周期内将S3打到ON）. Next, turn S 2 on and watch V out for 1 min（接下来将S2打到ON并且观察 Vout 分钟）.To compensate for leakage（为了弥补泄露 ），leave all switches off for 1 min and then throw S2 andS3 on（将所有开关打到 OFF 分钟并且将 S2 和 S3 打到 ON）. Monitor V out for 1 min and subtract t

9、his valuefrom the V out value obtained earlier（ 监测 Vout 分钟并且减去早先获得的Vout 值）.（View a largerversi on of the image. ） In deed, soakage seems an apt term for dielectric absorpti on （实际上，浸润似乎是描述介电吸收一个恰当的词）when you note what the capacitor seems to be doi ng（当你注意到电容表现岀的特性时）.Consider a typical example（考虑到一个特

10、殊的例子）:A capacitor charges to 10V for along time T （一个电容在一个长的时间T 内充电至U 10V ）and then discharges through a small-valueresistor for a short time t（然后通过一个小的电阻放电一小会时间t）. If you remove the short circuit andmon itor the capacitor termi nals with a high-impeda nee voltmeter（女口果你在断开放电回路之后用一个高阻抗的电压表测量电容的一端 ）,

11、you see the capacitor charge back to 0.1%, 1% or as much as 10% of the original voltage（ 你会观察到电容被重新充电到原电压的 0.1% , 1%甚至是10% ）. For example, a 1-F卩Mylar capacitor charged to 10V for 60 sec 仃charge ）（例如一个 1uF 的聚酯薄膜电容器充到10V 需要60s（Tcharge） ） and discharged for 6 sec （Tdischarge ） charges to 20 or 30 mV a

12、fter 1 min （Thold ）（并且放电 6s（Tdischarge），断开一分钟后会重新充电到20-30mv ）. Fig 1 shows a simple evaluation circuit formeasur ing this characteristic（图1显示了一个简单的测量这一特点的评估电路）.Fig 2 - To model the soakagecharacteristic of a 1-F Mylarcapacitor（ 用一个 1-uF 的聚脂薄膜电容模拟浸润 ）,con sider a circuit that in corporates a 0.006-卩

13、F capacitor to represe nt the dielectric'scharge-storage characteristics（用0.006-uF的电容代表电介质的电荷存储特性）.A capacitor exhibit ing dielectric absorpti on（种具有介电吸收的电容器）acts as if duri ng its longprecharge time the dielectric material has soaked up some charge that rema ins in the dielectricduring the brie

14、f discharge period（好像在一个长的预充电时间内介电材料已经吸收了一些电荷，这些电荷在短暂的放电期间内仍然保留在电介质当中）.This charge then bleeds back out of the dielectric during therelaxati on period and causes a voltage to appear at the capacitor term in als（这些电荷在张弛周期内由电介质中释放并且引起电容两端的电压的改变）.Fig 2 depicts a simple model of this capacitor（图2描述了这种电容

15、的简单模型）:When 10V is applied for 1 min, the 0.006-卩 F capacitor gets almost completelycharged（ 10V 电压在 1 分钟内几乎可以将 0.006-uF 电容充饱）,but during a 6-sec discharge period it only partially discharges（ 但是在6秒放电周期内只有一部分电荷放掉）.Then, over the next minute（接下来在下一分钟内 ）,the charge flows back out of the 0.006-卩 F and

16、charges the 1- 卩 F capacitor to a couple ofdozen millivolts（ 电荷从0.006uF电容回流并且重新给1-uF 电容充电至几十mV）. This example indicatesthat a longer dischargi ng time reduces soakage error （这个例子表明一个长的放电时间可以减小浸润效应）but that discharg ing for only a small fracti on of that time results in a larger error（但是放电一小段时间会导致更大的

17、问题）.Illustrating this point（图解这一要点）,Fig 3 shows the results of conductingFig1's basic test sequence（图3显示了按照图1的流程基本测试结果1/6/ 12s的放电时间）for 1-, 6- and12-sec discharge times. Note that the capacitor tries to remember its old voltage（注意到电容在尝试记住以前的电压 ）,but the longer you hold it at its new voltage, the

18、 better it forgets（但在新的电压上持续的时间越长，它的记忆性越差）-in the Fig 3 case, soakage errors equal 31 mV at tdischarge =l sec, 20mV at t discharge =6 see and 14 mV at t discharge =12 sec.（ 在图 3，在 T 放电=1s 时，浸润电压 =31mv ;在 T放电=6s 时，浸润电压 =20mv ;在T放电=12s 时，浸润电压 =14mv）0.250.5075TIME AFTER SHORT REMOVAL （MINIFig 3 - Obta

19、ined using Fig 1's test circuit,these dielectric-absorpti on-measureme nt results fora 1- F capacitor shown that longer tdischarge times reduce soakage-caused errors（用图 1 电路得到，对于1uF电容的电介质吸收测试结果表明：越长的放电时间越能减缓浸润效应）.High-speed tests predict S/H performanee（高速测试预测采样/保持电路的性能）You might now ask whether

20、 these low-speed tests have any bearing on a capacitor's suitability infast millisecond or microsecond sample/hold applications（你会问是否这种低速测试结果在ms 或 us 级的采样保持电路中有同样的影响）.If you repeat the Fig 1 experiment for T charge = T hold = 1000 sec andt discharge = 100s（ec（如果你重做电路 1的实验，T充电=T保持=1000us 并且T放电=10

21、0us ）, you see verysimilar capacitor-voltage waveforms but with about 10-times-smaller amplitudes（你可以看到同样的电容-电压波形，但是幅度只有原来的1/10 ）. In fact, for a constant T（事实上，对于恒定不变的T（这里指充电时寸间）：t ratio, the resulting soakage error decreases only slightly in tests ranging in length fromminutes to microseconds（对于T的

22、，浸润结果误差仅仅在测试时常从分钟到毫秒细微的减少）.?Fig 4 - More precise than Fig 2's equivale nt circuit（比电路 2 更准确的等效电路）,a capacitormodel employi ng several time con sta nts（使用几个时间常数的电容模型）proves valid for a wide range ofcharge and discharge times（对于大范围的充放电时间是有效的）.This model approximates a Mylarcapacitor（这个模型近似聚酯薄膜电容器）

23、.Fig 4's circuit approximates this capacitor characteristic（图 4 电路近似这种电容特征）,which you canobserve on actual capacitors by usingFig 5's test setup（你可以用图 5 的测试设置观察实际电容器）.Here,a sample/hold IC exercises the capacitor un der test at various speeds and duty cycles（这里，一个采样保持电路练习对电容在多种速率和占空比下进行测试）,a

24、nd a limiter amplifier facilitates close study ofthe small residual waveforms（并且一个限幅放大器帮助你更深入学习小的残余波形）,without over-driv ingthe oscilloscope whe n the capacitor is charged to full voltage（当电容被充满时不会过驱动示波器）.Fig 5 - Capable of automatically sequencing the dielectric-absorption tests（可以自动排序电介质吸收测试 ）,a c

25、ircuit employing timers（you to make measureme nts for a wide range of T 阶段允许对充电/保持/放电有广泛时间的测试obta ined using the circuit show n here（电路采用定时器 ）,a sample/hold and limiting stages allowsCHARGE , T HOLD , and t DISCHARGE values（采样/保持和限制（也就是可以在很多时常数下测试）.Fig 7 shows the results图7示岀此电路的测试结果）.（View a larger

26、 version of theimage.）Notes:1. ALL DIODES = 1N9142. IC5, IC6 = LM301A3. IC7 = MM74C044. USE R4 OR -10 GAIN TO KEEP SCOPE WAVEFORM BELOW 200mV SO AS TO AVOIDDISTORTION OR FALSE ATTENUATIONS（用R4或-10 倍的增益来保持波形的范围低于200mv 以至于避免失真或错误的衰减）Such experiments illustrate that if you put a certain amount of charg

27、e into a less-than-idealcapacitor（这个实验例子说明如果你给一个不太理想的电容充电）,you will get out a differe nt amou nt ofcharge（电路中将会岀现额外电荷）,depe nding on how long you wait（电荷的多少取决于你等多久（也就是放电时间的长短 ）.Thus, using low-soakage capacitors proves important in applications such as thoseinvolving high-resolution dual-slope integ

28、rating ADCs（因此在高分辨率的双积分ADC中应用低的浸润效应的电容是十分重要的）.And sure en ough, many top-of-the-l ine digital voltmeters do use polypropyle ne（a low-soakage dielectric） devices for their ma in in tegrat ing capacitors（果不其然，许多数字电压表确实使用的聚丙烯电容）.But dielectric-absorption characteristics are most obviously detrimental i

29、n applications involving sample/holds（但是电介质吸收特点很明显在采样/保持电路中是有害的 ）.Man ufacturers guara ntee howfast these devices can charge a capacitor in their Sample mode（厂商们会保证他们的器件（电容）在采样模式下充电是有多快 ）and how much their circuits' leakage causes capacitor-voltage droop duri ng theHold mode（ 和电路泄露泄露导致电容上电压在保持模式

30、时掉了多少），but they don't give any warni ng abouthow much the capacitor voltage cha nges because of soakage（但是他们不会给岀任何的关于电容电压因为浸润效应会改变多少的警告）.This factor is especially importa nt in a data-acquisiti on system（这个因素在采集系统中尤为重要）,where some cha nn els might han dle small voltages while others operatenear

31、full scale（在一些通道工作于满量程状态时，另外一些通道可能有感应电压）.Eve n with a good dielectric,a sample/hold can hurt your accuracy（即使是好的电介质，采样保持电路也会影响精度），especially if thesample time is a small fraction of Thol）.（特别是采样时间只占保持时间的一小部分 ）For example, althougha good polypropylene device can have only 1-mV hysteresis per 10V step

32、 if T/t=1OO msec/10 msec, this figure in creases to 6 mV if the T/t ratio equals 100 msec/0.5 msec（例如，虽然好的聚丙烯器件能达到只有没10V有1-mv的滞后，当 T采样/T保持=100msec/10msec时.这个指标可以达到6mv当T采样/T保持=100msec/0.5msec 时）（滞后，指的是输入信号实际值与采样值的误差， 如图2-1所示）.Because most sample/hold data sheets don't warn you of such factors（因为大

33、多数采样 /保持数据手册对这一因素没有警告你）,you should evaluate capacitors in a circuit such asFig 5's （你应该用电路 5 测量电容 ）,using timescal ing suited to your applicatio n（时间常数依你的应用定 ）.图21非线性系统的输入输岀特性Fi各21 Input and output characteristic of nonlinear systemCltOJ mF tantalum or mylarFig 6 - Soakage can prese nt problemsw

34、he n you're desig ning a fast-settli ng amplifier or filter（当你设计一个快速建立的放大器或者滤波器时浸润效应可能会引起问题）.In the circuit show n here, forexample, C 1 can be a Mylar or tan talum un it, but making C2 a polypropyle ne device improvesperformance（在这个电路中，C1可以用聚脂薄膜电容或坦电容，但是C2必须用聚丙烯电容以改善性能）.Other applicati ons in w

35、hich soakage can degrade performa nee are those in volv ing fast-settl ing acactive filters or ac-coupled amplifiers（其他的因为浸润效应降低性能的应用还包括：快速建立的有源滤波器或交流耦合放大器）.In Fig 6's circuit, C 1 can be a Mylar or tan talum unit because it always has 0V dc on it（在电路6种C1之所以可以用聚脂薄膜电容或者坦电容是因为它的直流电压一直是OV）, but mak

36、ing C 2polypropyle ne in stead of Mylar noticeably improves settli ng（但是在 C2 中用聚丙烯电容替代聚苯乙烯电容会显著地提高建立时间）.For example, settling to within -0.2 mV for a 10V step improves from 10 to1.6 sec with the elimination of Mylar's dielectric absorption（例如，以 2mv 每步建立到 10V 如果消除电解质吸收的话，可以从 10s 提高到 1.6s） . Simil

37、arly, voltage-to-freque ncy con verters ben efit fromlow-soakage timing capacitors, which improve V/F linearity（通常，电压 / 频率转换器受益于低侵润的定时电容能够提高V/F的线性度）.Some dielectrics are excelle nt at all speeds（有些电介质在任何速率下都是很好的）Fortu nately, good capacitors such as those emplo ying polystyre ne, polypropyle ne, NPO

38、 ceramic andTeflon dielectrics perform well at all speeds（幸运地是，用聚苯乙烯/聚丙烯/NOP的陶瓷和特氟龙制成的电容在任何速率下都表现的很好）.Fig 7 shows the characteristics of capacitors using these dielectrics andothers such as silver mica and Mylar（图7画岀了用这些优质电介质和其他像是镀银云母和聚脂薄膜制成的电容的特点）.1 n general, polystyrene, polypropylene or NPO-cera

39、mic capacitors furnish goodperformanee, although polystyrene can't be used at temperatures greater than 80C（尽管聚苯乙烯不能够在超过80 C以上工作，但是聚苯乙烯，聚丙烯或者NOP的陶瓷电容器能表现岀很好的性能）.And althoughF（尽管NOP他们有着NPO ceramic capacitors are expensive and hard to find in values much larger than 0.01的陶瓷电容器昂贵并且大于 0.01uF 的很难找到）,

40、they do achieve a low temperature coefficient （ 低的温度系数 ）（a spec not usually significant for a S/H but one that might prove advantageous forprecisi on in tegrators or voltage-to-freque ncy con verters（这种参数对采样保持电路来说意义不大，但是对精密积分器或者电压频率转换器来说很有用）.Teflon is rather expe nsive but defi nitely the best mater

41、ialto use whe n high performa nee is importa nt（特氟龙材料虽然贵但是在高性能应用中是最好的材料）.Furthermore, only Teflon and NPO ceramic capacitors suit use at 125C（此外，只有特氟龙和NOP 陶瓷电容适用于125 C ）.50 mV20 mV10 mV (0.1%)S mV2jt)V1 mV 0.01 %O.SinV乂 、HIGHXMYLAR弋、CERAMIC /sample BpTANTALUM silver mica"eVNOTES:帕LO - 1 口暑 101

42、严 n “Gt TCitAflOE - 11 xMYLARISAM PLEA、/polystyrene*< (SAMPLE BlJ_-ZNPOCERAM -一 一(SAMPLE B) POLYPROPYLENE *SAMPLE A)POLYSTYRENE(SAMPLE Al、*NPO CERAMIC忙 ”、(SAMPLE A、l*OLYPROP¥LENE、 -(&AMPLEB)J fbFLDN100 mV 厂fl 4 m W III丄<OqqV；.;1 6SEG 1 GC<- 1W rnrc 10 flight； 1 rriKtG lUU&ECIni

43、RCHAHatFig 7 - Soakage-measureme nt resultsfor a variety of capacitors illustrate the effects of tdischargevalues on dielectric-absorptio n-caused errors. Note the curves for two differe nt samples of NP0ceramic capacitors in tersect(各种各样的电容随Tdischarge的值不同电解质吸收引起的误差浸润测试结果).If you look at Fig 7's

44、 dielectric-absorption values, you can see wide differences in performanee for a given dielectric material(如果你看图7电解质吸收值，你可以不同电解质材料的巨大差异).For example,polypropylene sample A is about as good as B at t=6 see,(polypropylene sample A禾口 B 差不多)butB is four times better at high speeds(但是 B 在高速下比 A 好 4 倍).Si

45、milarly, NP0-ceramic sample A isslightly worse tha n NPO-ceramic sample B at low speeds(同样地， NPO-ceramic sample A在低速下比NP0-ceramic sample B 更差),but A is definitely better at high speeds(但很明显 A 在高速下性能更好 ).And some Mylar capacitors (sample A) get better as speed in creases from 1000 to 100sec( 一些聚 卩酯薄膜

46、电容(sample A)当速度从1000us 上升到100us 时性能会变好),but others (sample B) get worse(但是另一些 (sample B) 会变差).So if you want consistently good performanee from your capacitors(所以如果你想让你的电容在性能上一直保持稳定),evaluate and specify them for the speed at which they'll beused in your applicatio n（在他们被应用的特定速度的电路中，评估并指定这些电容）.K

47、eep in mind thatbecause most sample/holds are used at much faster speeds tha n those corresp onding to the 1- or5-min ratings usually given in data sheets（大多数采样/保持被用在相当于它们数据手册中1-5倍的速率下）,a published specification for dielectric absorptio n has limited value（一个已发布的对电解质吸收值的限制的说明书）.In additi on, other d

48、ielectrics fur nish various levels of performa nce（此外，其他的电解质 ）:« Because any long word that starts with poly seems to have good dielectric properties, howabout polycarb on ate or polysulfo ne? No - they are about as bad as Mylar（似乎聚酯的电容有好的介电性能，那聚碳酸酯或聚丙烯呢？不是这样的一-他们像聚酯薄膜一样烂）.« Does an air or

49、 vacuum capacitor have low soakage?（是不是空气或真空介质电容器有低的浸润）Well, it might, but many sta ndard capacitors of this type are old desig ns with ceramicspacers, and they might give poor results because of the ceramic's hysteresis（也许是，但因为陶瓷的迟滞使得有的结果很差）.* If a ceramic capacitor is not an NPO device（如果陶瓷电容器

50、不是NPO 材料的）,is it any good?（ 它性能更好吗？）Most of the conventional high-K ceramics are just terrible（大多数传统的高K 陶瓷是很差的）-20 to 1000 times worse than NP0 and even worse than tantalum（比 NPO 差20 1000倍，甚至比钽电容还差 ）.* Is silic on dioxide suitable for small capacita nces（二氧化硅小电容合适吗 ）? Although Fig 5's testsetup,

51、 used in preparingFig 7's chart（用图 5 的测试电路绘制岀图7 样的图）,only measuresmoderate capacita nces （500 to 200,000 pF）, silic on dioxide appears suitable for the smallcapacitors needed for fast S/Hs or deglitchers（它只能制造岀中等数值容量的电容（500-200000pF,）二氧化硅表现岀适合于低电容需求快速采样/保持电路或变阻器）.Cancellation circuit improves ac

52、curacy（浸润消除电路和提升精度 ）A practical method of gett ing good performa nee with less-tha n-perfect capacitors is to use a soakage-cancellation circuit such as one of the form shown inFig 8 （常用的使不怎么好的电容变得性能更好的办法，就是像电路8 一样使用浸润消除电路）,in which a capacitor of the type modeled inFig4 serves as an integrator. （On

53、ly the first two soakage elements are shown.）（这个电路中电容像图4一样，是它的实际等效模型（其中只有只有其中两路）The in tegrator's output is in verted with a scale factorof -0.1, and this voltage is then fed through one or more experimentally chosen RC networks tocan cel the equivale nt n etwork in here nt in the capacitor'

54、s dielectric material（这个积分器的输岀用被用-0.1倍的比率引回，这个用一个RC网络的引回来消除电容内部电解质材料所产生的那个网络）.COMPENSATION FEEDBACK PATHOrOO3 #F1( I200M M0.Q06MODEL OF 1 j*F MYLAR CAPACHOR1000M20MI0 06FFig 8 - You can compe nsate an in tegrator for dielectric absorpti onby feedi ng its in vertedoutput back to the in put through on

55、e or more experime ntally chose n RC n etworks, which can cel theequivale nt network in here nt in the capacitor's dielectric material(你可以用一个 RC 网络将输岀反相引回到输入来补偿这个积分器当中的电解质吸收).Fig 9 shows a practical sample/hold circuit with an easily trimmed compensator. This networkprovides about a 10-fold impr

56、ovement for sample times in the 50- to 2000-sec range ( Fig 10 ).Although this compensation is subject to limitations at very fast or slow speeds, the number of RCsecti ons and trim ming pots employed can be exte nded.SAMPLEHOLDVIN O_U15VPOLYPROPYLENEO 15VILF398AS/H&100kSLOW SOAKAGE TRIM_ Pl100k

57、 FAST SOAKAGETRIM200k<T 0.0033MYLAR200kT 0.022MYLARa good Teflon capacitor furnishes a dynamic range of 15 to 17 bits（ 能够达到从15bits 17bits 精度的大动态范围ADC ）.AFTER COMPENSATIONBEFORECOMPENSATIONCIRCUITWITH FIG 9mSEC/cm10mSEC = T HOLD 5 mSEC/cm1 mSEC = T HOLD 0.5mSEC/cm1 mSEC = T HOLD 0.510 mSEC = T HOLD 5 mSEC/cmFig 10 - Addi ng Fig 9's compe nsati on n etworkperforma nee improveme nt for sample times of 50 to 2000络，对于50-2000us的采样时间能够提升10