Design And Implementation Of An FPGA-based Data Acquisition System For Linear CCD Sensor

Ar F laser lithography machine is the core component of the product. It provides the exposure for the lithography light source. Wavelength and pulse width are two key indicators of Ar F laser. This article is for 193 nm Ar F laser domestic demand, research on laser linewidth laser line detection technology, and complete the data acquisition board development. The main contents are as follows:Analysis of the Ar F laser for laser linewidth measurement functionality, performance requirements and interface constraints, the data acquisition system must be able to determine the inner and outer trigger switch function in both modes, and can freely set the frequency of data collection and integration time according to the frequency of laser. Based on the spectral algorithm realizing in different methods, this article uses a board integrated Ethernet-based implementations. Not only centralized function but also convenient debugging.Select the appropriate spectral measurements linear CCD for laser wavelength, and analysis of the CCD working principle and timing. Sensor is designed for the differential amplifier circuit, and can effectively remove the CCD dark current noise. Hardware use the Ethernet interface and FPGA in RAM as a buffer storage structure to achieve real-time transmission spectral data.A detailed analysis of software workflow and data flow. The embedded software is divided into drive module, Ethernet module, FIFO buffer module. The system uses Verilog HDL hardware description language to implement the various modules and global control. Designed and optimized drive timing and cache timing relationships.Analysis of results integral characteristic of each sensor pixel point sequence begins and ends. Developed the corresponding data acquisition boards, and carray out the laser pulse acquisition experimental tests. The collected data taken recursive average and software compensation data processing method. And compare the results of the line width before and after processing. Achieve a target statistical error of about ±10%.