拆分测量的S参数出现一些错误

大家好，回到以前的公司，我得到了一个模拟文件，允许我将测量的S参数分成两个S参数文件。
假设是两个文件背靠背或S11A = S22B S22A = S11B（需要）S21A = S12A（需要）S21B = S12B（需要）并且如果需要S21A = S21B我认为它是迭代解决方案。
这样做的原因是去嵌入。
我有一个评估板，测量范围从100 kHz到9 GHz，我需要数据到设备。
我有一个删除设备的直通线。
是的，会出现一些错误，但由于低频和端口扩展不能用于高频，因此不能使用TRL校准。
由于我不再为前公司工作，我无法访问模拟。
有没有人记得或有这个解决方案。
我当时确实是从安捷伦那里得到的。
提前致谢。
Mike Virostko集成设备技术



     以下为原文

  Hello All,

   Back at a previous company I had gotten a simulation file that allowed me to split a measured S-parameter into two S-parameter files.  The assumption was the two files were back to back or

       S11A  =  S22B
       S22A  =  S11B (required)

       S21A  = S12A  (required)
       S21B  = S12B (required)

and if needed S21A=S21B

I believe it was an iterative solution.

The reason for doing this is is for de-embedding.  I have an evaluation board which is being measured from 100 kHz to 9 GHz and I need the data to the device.  I have a through line that has the the device removed.  Yes there will be some error to this, but TRL calibration can not be used due to the low frequency  and port extension can not be used to to the high frequency.

Since I no longer work for the former company I have no access to the simulations. Does anyone remember or have this solution.  I did get it from Agilent at the time.

Thanks in advance.

Mike Virostko
Integrated Device Technology