首页> 中文期刊> 《现代电子技术》 >神经网络曲线拟合在温补晶振上的应用

神经网络曲线拟合在温补晶振上的应用

         

摘要

石英晶振作为重要的频率源器件,其频率稳定度至关重要。但温度对石英晶振的影响很大。传统的微处理器温度补偿晶振中在拟合曲线时算法简单,导致软件引起的误差较大。利用了神经网络算法在曲线拟合上的应用,拟合出补偿电压与温度之间的函数关系,微处理器根据温度传感器采集的温度控制AD芯片产生补偿电压,从而使压控振荡电路输出稳定的频率的目的。实验结果表明:温度在-10~80℃时,频率稳定度达到±0.35 ppm,比未补偿时提高了近20倍,比其他曲线拟合方法得出的效果要好。%The quartz crystal oscillator is an important frequency source device. Its stability is crucial. However there is a great influence of temperature on the quartz crystal oscillator. The traditional MCU quartz crystal oscillator of temperature com-pensation may result in large error caused by software because its algorithm for curves fitting is too simple. In this paper,a func-tion relation between compensation voltage and temperature was fitted by using a neural network algorithm for curves fitting. The MCU controls AD chip to generate a compensation voltage according to temperature collected by temperature sensors,so as to make the VOC circuit output a stable frequency. The experimental results show that,when the temperature is from -10oC to 80oC,frequency stability gets 0.35 ppm,which increases nearly 20 times greater than the frequency stability without compensa-tion,and the results got from the proposed method is better than those of other curve fitting methods.

著录项

相似文献

  • 中文文献
  • 外文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号