超大规模集成电路第四次作业2016秋-段成华

1.Shownbelowarebuffer-chaindesigns.(1)Calculatetheminimumdelayofachainofinvertersfortheoveralleffectivefan-outof64/1.Solution：由题可知：64F根据经验6.3optf为最合适的值，所以6.364NNFf，所以24.3N，但是级数必须为整数所以取3N，又因为1，所以：15)641(3,464303ppttf，所以时最合适4f。(2)UsingHSPICEandTSMC0.18umCMOStechnologymodelwith1.8Vpowersupply,designacircuitsimulationschemetoverifythemwiththeircorrespondentparametersofN,f,andtp.Solution:根据（1）中计算知道三级最合适，所以验证如下：A）、一级无负载测本征延时代码如下：.titlebuffer-chain1.lib'C:\synopsys\Hspice_D-2010.03-SP1\tsmc018\mm018.l'TT*set0.18umlibrary.optscale=0.1u*setlambda.optionspost=2list.temp27.globalvddVddvddgnd1.8vinvin00.9pulse01.825n5p5p49.99n100n$频率为10MhzClvoutgnd0f$Cg1=2.46fF，负载为CL=157.44fF.subcktinvinoutwn=3.5wp=10t=7.5mnoutingndgndNCHl=2w=wnad='wn*t'pd='wn+2*t'as='wn*t'ps='wn+2*t'mpoutinvddvddPCHl=2w=wpad='wp*t'pd='wp+2*t'as='wp*t'ps='wp+2*t'.endsX1vinvoutinvwn=3.5wp=10t=7.5.op.tran5p5n.meastranvoutmaxmaxv(vout)from=5pto=5n.meastranvoutminminv(vout)from=5pto=5n$一级.meastrantphl1+trigv(vin)+val=0.9+rise=1+targv(vout)+val='0.5*(voutmax-voutmin)+voutmin'+fall=1.meastrantplh1+trigv(vin)+val=0.9+fall=1+targv(vout)+val='0.5*(voutmax-voutmin)+voutmin'+rise=1.end1）一级无负载测得本征延时约为17ps；2）带上64倍Cg1大小的负载测得延时为750.35ps，是本征延时的44倍B）、三级带负载测延时代码如下：.titlebuffer-chain3.lib'C:\synopsys\Hspice_D-2010.03-SP1\tsmc018\mm018.l'TT*set0.18umlibrary.optscale=0.1u*setlambda.optionspost=2list.temp27.globalvdd.paramfan=4Vddvddgnd1.8vinvin00.9pulse01.825n5p5p49.99n100nClvoutgnd0f$Cg1=2.46fF，负载为CL=157.44fF.subcktinvinoutwn=3.5wp=10t=7.5mnoutingndgndNCHl=2w=wnad='wn*t'pd='wn+2*t'as='wn*t'ps='wn+2*t'mpoutinvddvddPCHl=2w=wpad='wp*t'pd='wp+2*t'as='wp*t'ps='wp+2*t'.endsX1vin2invwn=3.5wp=10t=7.5X223invwn='fan*3.5'wp='fan*10't=5X33voutinvwn='fan*fan*3.5'wp='fan*fan*10't=5.op.tran50p500n.meastranvoutmaxmaxv(vout)from=50pto=500n.meastranvoutminminv(vout)from=50pto=500n$三级.meastrantphl3+trigv(vin)+val=0.9+rise=1+targv(vout)+val='0.5*(voutmax-voutmin)+voutmin'+fall=1.meastrantplh3+trigv(vin)+val=0.9+fall=1+targv(vout)+val='0.5*(voutmax-voutmin)+voutmin'+rise=11）带上64倍Cg1大小的负载测得延时为174.6ps，是本征延时的10.27倍总结如下：经过调整参数近似时每一级的1，所以经过手工计算得到一级带负载和三级带负载的延时比值为：2344.065151300pptttptp，而仿真得到的结果为2327.035.7506.174，所以符合手工计算的比值，同理其他级的延时代码也是如上的写法，经过仿真得到三级延时最小。.end2.Considerthelogicnetworkbelow,whichmayrepresentthecriticalpathofamorecomplexlogicblock.Theoutputofthenetworkisloadedwithacapacitancewhichis5timeslargerthantheinputcapacitanceofthefirstgate,whichisaminimum-sizedinverter.TheeffectivefanoutofthepathhenceequalsF=CL/Cg1=5.UsingHSPICEandTSMC0.18umCMOStechnologymodelwith1.8Vpowersupply,designacircuitsimulationschemetoverifytheOPTIMAZATIONparametersofg,f,andsforeachoftheinverterandgates.Solution:由题得到路径逻辑努力925135351G，由于没有分支B=1，所以9125GFBH，所以使延时最小的逻辑努力为93.191254NHh，得到如下的扇出系数：93.1,16.1,16.1,93.14321ffff，利用书上公式6.18计算得到尺寸系数6.2/,34.1/,16.1/,14132143121321121ggfffSggffSggfSS。电路仿真代码如下：.titleINV2NAND2NOR.lib'C:\synopsys\Hspice_D-2010.03-SP1\tsmc018\mm018.l'TT*set0.18umlibrary.optionspost=2list.temp27.globalvddVddvddgnd1.8vinvin00.9pulse0.01.8150p5p5p290p600pC1voutgnd12.3f$Cg1=2.46fF，所以负载为12.3fF.subcktinv1inoutwn=0.35uwp=1ut=0.75umnoutingndgndNCHl=0.2uw=wnad='wn*t'pd='wn+2*t'as='wn*t'ps='wn+2*t'mpoutinvddvddPCHl=0.2uw=wpad='wp*t'pd='wp+2*t'as='wp*t'ps='wp+2*t'.ends.subcktnand3NAND-A1NAND-D1NAND-B1NAND-C1wn='0.35u*1.16'wp='1u*1.16't=0.5u$优化尺寸系数S2*.subcktnand3NAND-A1NAND-D1NAND-B1NAND-C1wn=0.35uwp=1ut=0.5u$未优化尺寸系数S2mn3NAND-S2NAND-C1gndgndNCHl=0.2uw=wnad='wn*t'pd='wn+2*t'as='wn*t'ps='wn+2*t'mn2NAND-S1NAND-B1NAND-S2gndNCHl=0.2uw=wnad='wn*t'pd='wn+2*t'as='wn*t'ps='wn+2*t'mn1NAND-D1NAND-A1NAND-S1gndNCHl=0.2uw=wnad='wn*t'pd='wn+2*t'as='wn*t'ps='wn+2*t'mp1NAND-D1NAND-A1vddvddPCHl=0.2uw=wpad='wp*t'pd='wp+2*t'as='wp*t'ps='wp+2*t'mp2NAND-D1NAND-B1vddvddPCHl=0.2uw=wpad='wp*t'pd='wp+2*t'as='wp*t'ps='wp+2*t'mp3NAND-D1NAND-C1vddvddPCHl=0.2uw=wpad='wp*t'pd='wp+2*t'as='wp*t'ps='wp+2*t'.ends.subcktnor2NOR-A1NOR-D1NOR-B1wn='0.35u*1.34'wp='1u*1.34't=0.5u$优化尺寸系数S3*.subcktnor2NOR-A1NOR-D1NOR-B1wn=0.35uwp=1ut=0.5u$未优化尺寸系数S3mn2NOR-D1NOR-B1gndgndNCHl=0.2uw=wnad='wn*t'pd='wn+2*t'as='wn*t'ps='wn+2*t'mn1NOR-D1NOR-A1gndgndNCHl=0.2uw=wnad='wn*t'pd='wn+2*t'as='wn*t'ps='wn+2*t'mp1NOR-S1NOR-A1vddvddPCHl=0.2uw=wpad='wp*t'pd='wp+2*t'as='wp*t'ps='wp+2*t'mp2NOR-D1NOR-B1NOR-S1vddPCHl=0.2uw=wpad='wp*t'pd='wp+2*t'as='wp*t'ps='wp+2*t'.ends.subcktinv2inoutwn='0.35u*2.6'wp='1u*2.6't=0.5u$优化尺寸系数S4*.subcktinv2inoutwn=0.35uwp=1ut=0.5u$未优化尺寸系数S4mnoutingndgndNCHl=0.2uw=wnad='wn*t'pd='wn+2*t'as='wn*t'ps='wn+2*t'mpoutinvddvddPCHl=0.2uw=wpad='wp*t'pd='wp+2*t'as='wp*t'ps='wp+2*t'.endsX1vin2inv1X223vddvddnand3X334gndnor2X44voutinv2.op.tran5p3000p.meastranvoutmaxmaxv(vout)from=5pto=3000p.meastranvoutminminv(vout)from=5pto=3000p.meastrantphl+trigv(vin)+val=0.9+rise=2+targv(vout)+val='0.5*(voutmax-voutmin)+voutmin'+rise=2.meastrantplh+trigv(vin)+val=0.9+fall=2+targv(vout)+val='0.5*(voutmax-voutmin)+voutmin'+fall=2.end$Cg1=2.46fF，所以负载为12.3fF仿真结果如下：尺寸系数全部优化得到的tphl和tplh尺寸系数全部未优化得到的tphl和tplhVout对比图：粉线是尺寸系数全部未优化、浅蓝线是尺寸系数未优化的输出电压、