薄膜帽的有限元模拟得到不同结果

在第一个示例中很难看到端口详细信息。
显然，对于您的小型设备和更大的端口，您需要非常小心地进行端口去嵌入，以及接地返回电流的路径长度。
所以第二个例子对我来说很糟糕，返回电流必须在模拟盒周围从端口1流到端口2.最好在两个板之间放置一个集总端口。



     以下为原文

  It is difficult to see the port details in the first example.

Obviously, with your small device under test and the much larger ports, you need to be VERY careful with port de-embedding, and the path length for ground return current. So the second example looks bad to me, with return current having to flow all around the simulation box from port 1 to port 2. You would better place one lumped port between the two plates.

>对于端口校准，我使用了None。
对于您的情况，这会给FEM带来可怕的结果。
所有端口长度都包括在寄生物中。
这里没用！
>第二个结果似乎也不正确但我需要两个端口结果用于我的cosim仿真。
不，除非模型中有明确定义的（物理上有意义的）基础，否则无法创建有用的2端口数据。
您正在寻找的是具有纯系列元素的模型。
这可以通过1端口模拟创建，您可以将（+）引脚连接到一个板，将（ - ）引脚连接到另一个板。
将两个引脚分组到端口编辑器中的一个差分端口。
然后在*符号*中，您需要访问（+）终端和参考终端。
ADS允许您在符号中创建参考引脚，在这里您需要它。



     以下为原文

  > For port calibration I have used None.

This will give terrible results with FEM for your case. All the port length included as parasitics. Not useful here!

> The second one results also does not seem correct but I need a two ports results for my cosim simulation. 

No, you can't create useful 2-port data unless you have a well defined (physical meaningful) ground in the model. What you are looking for is a model with pure series elements. That can be created with a 1-port simulation where you connect the (+) pin to one plate and the (-) pin to the other plate. 
Group the two pins into one differential port in the port editor.


Then in the *symbol* you then need access to both (+) terminal and also the reference terminal. ADS allows you to create the reference pin in the symbol, and here you need that.