LM算法在OCD低频噪声参数拟合中的应用

Application of low frequency noise parameters fitting in OCD with LM algorithm

用测量低频噪声的方法对航天用光电耦合器件(OptoelectronicCoupledDevices,OCD)的可靠性进行了筛选,并在测量大量OCD器件的基础上,采用Levenberg-Marquardt(LM)算法对实际噪声功率谱进行了低频噪声谱参数拟合,进而确定了导致噪声增加的器件内部缺陷。该方法快速有效,为改进光电耦合器件的生产工艺,提高产品的质量和可靠性提供了理论依据。

A reliability selectionmethod for astronautic optoelectronic coupled devices(OCD) by using low frequency noise measuring was presented.Based on large amount of measurement for OCD,with Levenberg-Marquardt algoritm,the fit for real curve of low frequency noise spectrum parameters was conducted in order to detect defects leading to the noise increase in the devices.Practical application for OCD selection proved that this method is valid and fast,and can provide a theoretical foundation to improve OCD production techniques,qualities,and reliabilities.