【作者】 黄俊；

【导师】 戴志坚；

【作者基本信息】 电子科技大学 ， 工程硕士（专业学位）， 2020， 硕士

【摘要】 集成电路测试贯穿了集成电路设计、制造、封装测试整个产业链及应用环节,一般采用专用的集成电路自动测试设备(Automatic Test Equipment,ATE)进行测试,而被测器件电源模块(DPS:Device Power Supply)是ATE的基础性功能模块,用于对被测器件(DUT:Device Under Test)供电。随着集成电路的发展,对DPS模块在输出精度,纹波,动态响应特性,电源转换效率等方面提出了更高的要求。本课题研制的DPS模块要求实现±20V输出电压范围,±6A电流范围,解决大负载条件下电源快速响应、电压/电流精确控制、电源的高效转换等问题。主要研究内容包括:1、电源输出的精确控制:以ARM控制器为核心,通过DAC调节功率放大器的输入,实现可调电源输出。为了实现输出电源的快速响应及精确控制,采用电压电流双闭环负反馈控制。对输出电压采集时,采用四线开尔文测量消除DPS模块输出端与被测器件DUT间由于附加电阻和引线电阻引起的误差。对电流采集时,采用分档采集提高测量精度。2、高效率的电源转换:根据线性稳压和开关稳压电源的特点,提出了四开关管预稳压加功率放大稳压的两级稳压方案,并利用四开关管升降压电路降低大负载下功率放大器的功率耗散。同时,为减小四开关管升降压电路中功率管的导通关断损耗,提出了一种ZVS软开关技术。3、大负载下电源的快速响应调节:根据本课题DPS模块的特点经过推导得到其等效的传递函数,利用simulink建立其仿真模型。使用matlab工具分析DPS模块的开环频率响应特性,结合SISO TOOL工具分析补偿控制器的设计需求。为提高DPS模块的相位裕度和穿越频率,改善电源输出响应特性,先后设计了增量式PID控制算法、BP神经网络自适应PID控制算法、RBF神经网络自适应PID控制算法。仿真结果说明基于RBF神经网络辨识的自适应PID控制算法性能最优,符合本课题的指标需求,最后详细阐述了控制算法的工程实现。测试结果表明,本课题研制的DPS模块实现了输出电压电流的精确控制,纹波电压≤20mVpp,满足大多数集成电路测试的要求。更多还原

【Abstract】 Integrated circuit testing runs through the entire industrial chain and application links of integrated circuit design,manufacturing,packaging and testing.Generally,special integrated circuit automatic test equipment(Automatic Test Equipment,ATE)is used for testing,while the power supply module of the device under test(DPS: Device Power)Supply)is a basic functional module of ATE,which is the the power of device under test(DUT: Device Under Test).With the development of integrated circuits,higher requirements have been proposed on DPS modules in terms of output accuracy,ripple,dynamic response characteristics,and power conversion efficiency.The DPS module developed in this subject can achieve an output voltage range of ± 20 V and a current range of ± 6A,which solves the problems of rapid response under high power load conditions,accurate control of voltage and current,and efficient power conversion.The main research contents include:1.Studying the precise control method : the ARM controller is the core,and the input of the power amplifier is adjusted through a DAC to realize the adjustable power supply output.In order to achieve fast response and precise control of output power.Using voltage and current double closed-loop control.When collecting the output voltage,the four-wire Kelvin collection is used to eliminate the error caused by the additional electronics and lead resistance between the output of the DPS module and the DUT of the device under test.For current collection,the use of sub-collection to improve the measurement accuracy.At the same time,the Butterworth active filter and LC passive filter are used to decrease the output power ripple.2.Researching on the method of high-efficiency power conversion: According to the characteristics of linear voltage stabilization and switching power supply,a two-stage voltage stabilization method of four-switch pre-regulation plus power amplification is proposed.A four-switch buck-boost circuit is used to reduce the power dissipation of the power amplifier under large loads.At the same time,in order to reduce the turn-on and turn-off losses of the power tube in the four-switch tube buck-boost circuit,a ZVS soft switching technology is proposed.3.Fast response of power supply under heavy load: obtain the equivalent transfer function according to the hardware circuit of the DPS module and use Simulink to establish its simulation model.Using matlab tools to analyze the open-loop frequency response characteristics of the DPS module,and design a compensation controller to improve the system’s phase margin and crossover frequency.Firstly,the compensation controller based on the conventional PID control is designed in combination with the characteristics of the incremental PID control algorithm and the defects of the conventional PID control algorithm are explained.On this basis,the design method of adaptive PID controller based on neural network is discussed.The characteristics of BP back propagation neural network and RBF neural network are studied,and the adaptive PID control algorithm of RBF neural network with identification is been designed.The test results show that the DPS module developed in this project achieves precise control of the output voltage and current,and the ripple voltage is ≤20mVpp,which meets the requirements of most integrated circuit tests.更多还原