简介:文章对0.5μm/40V高压工艺中形成的N阱电阻的SPICE模型进行研究。因为高压电路的实际应用,N阱电阻的寄生效应不可忽略,所以精确反应其电学特性的SPICE模型也显得尤为重要。从N阱电阻的测量结果反映其IV曲线的非线性特性和结型栅场效应管JFET输出特性曲线具有相似性,并通过对高压阱电阻与JFET的结构分析认为可以采用JFET的电压电流关系来建立合理的数学模型反映高压阱电阻的这种非线性特性。因此在文中借用JFET的SPICE模型作为基础,用宏模型的方法为高压N阱电阻建立了一套精确的SPICE模型、此模型适用于各类仿真器,具有一定的通用性.
简介:Photoluminescenceevaluationofpandntype6H-SiCsampleshasbeendone.Resultsshowthatatlowtemperaturethephotoluminescenceof6H-SiCisclearlydominatedbydonor-acceptorpairtransitions,insomecase,free-to-donortransitioncouldbeobservedathighertemperature.Thethermalquenchingprocessesofthephotoluminescencehavebeeninvestigatedtodeterminethepossibleionizationnenergiesoftheimpurities.
简介:8310进水机经过超声波处理后正常使用了一周左右,又返修,其故障现象:用维修电源加电按开机键有50mA左右的电流反应,先查电源IC逻辑供电是否正常,在电容C205的非接地端测VFLASH—1其电压只有1.8V,查阅图纸,该点的电压正常值为2.8V,主供字库,再测VCORE只有0.8V(在电容C208非接地端测)正常值为1.5V,查阅此电压供CPU。再测VIO为0V(在电容C207的非接地端测),VIO的正常值为1.8V,查图可知此电压供CPU、字库,造成VIO电压为0V的原因:一是电源IC坏、二是VIO滤波电容击穿、三是CPU内部短路。我先采用最简单的方法:用万用表蜂鸣档测VIO滤波电容C207时发出蜂鸣,断定是C207击穿所至,把C207取掉,再在C207焊盘处测VIO电压为1.1V,(如图1):
简介:Intheopennetworkenvironment,maliciousattackstothetrustmodelhavebecomeincreasinglyserious.Comparedwithsinglenodeattacks,collusionattacksdomoreharmtothetrustmodel.Tosolvethisproblem,acollusiondetectorbasedontheGNalgorithmforthetrustevaluationmodelisproposedintheopenInternetenvironment.Byanalyzingthebehavioralcharacteristicsofcollusiongroups,theconceptofflattingisdefinedandtheG-Ncommunityminingalgorithmisusedtodividesuspiciouscommunities.Onthisbasis,acollusioncommunitydetectormethodisproposedbasedonthebreakingstrengthofsuspiciouscommunities.Simulationresultsshowthatthemodelhashighrecognitionaccuracyinidentifyingcollusionnodes,soastoeffectivelydefendagainstmaliciousattacksofcollusionnodes.
简介:Animprovedgeneticalgorithmforsearchingoptimalparametersinn-dimensionalspaceispresented,whichencodesmovementdirectionanddistanceandsearchesfromcoarsetoprecise.Thealgorithmcanrealizeglobaloptimizationandimprovethesearchefficiency,andcanbeappliedeffectivelyinindustrialoptimization,dataminingandpatternrecognition.
简介:ThisletterpresentsanewchunkingmethodbasedonMaximumEntropy(ME)modelwithN-foldtemplatecorrectionmodel.Firsttwotypesofmachinelearningmodelsaredescribed.Basedontheanalysisofthetwomodels,thenthechunkingmodelwhichcombinestheprofitsofconditionalprobabilitymodelandrulebasedmodelisproposed.Theselectionoffeaturesandruletemplatesinthechunkingmodelisdiscussed.ExperimentalresultsfortheCoNLL-2000corpusshowthatthisapproachachievesimpressiveaccuracyintermsoftheF-score:92.93%.ComparedwiththeMEmodelandMEMarkovmodel,thenewchunkingmodelachievesbetterperformance.