怎么对不同的端口阻抗执行TRL校准

如果您提供更多信息，您可以自助。
比如* VNA模型。
*固件（有时很有用）*适合的选项（有时有用）*频率*波导，同轴电缆，在PCB上？
（我想我们可以排除波导，因为你说50欧姆）*连接器接口*你所拥有的校准套件都可以提供帮助。
戴夫



     以下为原文

  You might help yourself if you gave more information. Things like

* Model of VNA. 
* Firmware (sometimes useful)
* Options fitted (sometimes useful)
* Frequency
* Waveguide, coax, on a PCB? (I guess we can rule out waveguide, since you said 50 Ohms)
* Connector interface
* Cal kit(s) you have

could all be helpful. 

Dave

- 我可以使用N5230A或E8363B网络分析仪。
- 我可以使用的校准套件是3.5mm 85052D或N4691-6006 ecal。
- DUT是1对3带状线功率分配器，具有50R输入和33R输出阻抗，我更喜欢使用50R SMA连接器作为接口。
- 频率：10-18GHz非常感谢，Amin



     以下为原文

  - I can use either N5230A or E8363B network analyzers.
- The calibration kits I can use are 3.5mm 85052D or N4691-6006 ecal.
- The DUT is a 1-to-3 stripline power divider with 50R input and 33R output impedances and I'd prefer to use 50R SMA connectors as interface.
- Frequency: 10-18GHz

Thanks a lot,
Amin

而且我不需要完整的4端口矩阵。
唯一重要的参数是s21，s31，s41和s11，端口1是输入端口，端口2-4是功率分配器的输出端口。



     以下为原文

  And I don't need the complete 4 port matrix. The only important parameters are s21, s31, s41 and s11 with port 1 being the input port and ports 2-4 being the output ports of the power divider.

事实上，我认为理论上可以做到这一点，我不确定是否有人已经实现了数学运算。
TRL的理论依赖于以下事实：一些元素的S参数是众所周知的，例如Thur的S参数是S11＆amp; S22 = 0，S21＆amp; S12 = 1;
在更复杂的imiplementation中，值可能是非理想的并且是已知的（也就是说，通过的阻抗可以不是50欧姆），但我没有看到任何解决方案在输入和输出上提供不同的阻抗。
您说您想使用SMA 50欧姆连接器进行连接，但是您打算在另一端使用33欧姆的一端制作基于50欧姆的TRL校准套件吗？
您如何建议构建Thru标准或线标准？
你打算如何构建DUT？
你可以张贴你的布局的代表性草图（使用描述性文字比简单的草图更糟糕）。
在这种情况下，通常更容易构建所有50欧姆标准，以50欧姆校准您的灯具，然后在非50欧姆部件上使用阻抗变换来创建基于33欧姆的测量。
在这种情况下，您将在33欧姆DUT的接口处具有不连续性（线宽的离散变化），其中50欧姆线用于TRL夹具。
另一种选择（我试过）是使用全新的单端口开放式自动夹具拆卸，例如安捷伦新的PLTS程序中提供的。
这允许在每个夹具的一个端口上仅打开固定装置。
我没有尝试过混合阻抗，但我会在本周晚些时候看看结果是否符合预期。



     以下为原文

  In fact, I think it is theorectically possible to do this, I'm not sure anyone has implemented the math to do this.

The theory of TRL relies on the fact that the S-parameters of some of the elements are well known, such as the S-parameters of the Thur are S11&S22=0, S21&S12=1; in more sophisticated imiplementations, the values can be non-ideal and known (that is, the impedance of the thru can be other than 50 ohms), but I've not see any solution that provides for different impedances on the input and output.

You say you want to use SMA 50 ohm connectors for the connection, but do you intend to make a TRL cal kit that is based on 50 ohms on one end of 33 ohms on the other?  How do you propose to construct the Thru standard or the line standard?  How to you intend to construct the DUT? Can you post of representative sketch of your layout (using descriptive text is so much worse than just a simple sketch).

In cases like this, it is often easier to consturct all 50 ohm standards, calibrate your fixtures in 50 ohms, and then use impedance transformation on non-50 ohm parts to create the 33 ohm based measurments.

In such a case you will have a discontinuity (discrete change in linewidth) at the interface of the 33 ohm DUT with the 50 ohm line used for the TRL fixture.

An alternative (which I have tried) is to use the brand new one-port open-only automatic fixture removal such as provided in Agilent's new PLTS program.  This allows de-embedding fixtures with only an open on one port of each fixture.  I haven't tried it with mixed impedances, but I will later this week to see if the results are as expected.

对于您的1/3输出DUT配置和不同的端口阻抗，TRL校准不是一个好选择，尤其是当您有50欧姆SMA连接器时。
尽量保持TRL线的阻抗匹配，同时进行阻抗变换非常棘手。
从理论上讲，这是可能的。
实际上，您可能会发现它非常困难。
你如何验证你的测量？
您可能最好使用未知的直通校准并创建一些33欧姆的标准。
肯



     以下为原文

  For your 1 in 3 out DUT configuration and with different port impedances, TRL calibration is not a good choice especially when you have 50 ohm SMA connectors.  Try to keep the impedance of the TRL lines match and doing impedance transformation the same time is very tricky. Theoretically, it is possible.  Practically, you may find it very difficult.  How do you verify your measurements?  You may be better off using unknown thru calibrations and create some 33 ohm standards.  

Ken