如何进行带连接天线的全双端口校准

> {quote：title = igouzouasis写道：} {quote}>嗨，>>我在我的实验室遇到问题，想和你讨论。
如果你能给我任何帮助，我将不胜感激。
>>假设您想要根据S11和S21测量薄介电片的反射和透射特性。
我们的实验室只有一个VNA和一些宽带微带天线（非常指示性），设计用于我们想要测量电介质的频率。
>>最简单的想法是在两个天线就位的情况下执行完整的2端口校准，一个连接到每个端口，然后将片材插入其间并进行S11和S21测量。
>>但是如何进行连接天线的完整2端口校准？
我在应用笔记中找不到任何信息。
>>谢谢> Yannis我认为你不能做你想象的事情。
要进行完整的2端口校准，您需要在每个端口上放置3个标准，如果端口现在是天线，我看不出这是怎么回事。
一个明显的例外是，如果你可以使天线成为波导以同轴转换，而不是微带线。
在这种情况下，您需要一个波导校准套件。
也就是说，之前已经使用过测量两个天线之间的介电特性。
你会遇到天线需要在远场的问题，为此你可能需要相当多的样本。
有许多方法可以测量材料的介电性能。
请查看http://cp.literature.agilent.com/litweb/pdf/5991-2171EN.pdf http://cp.literature.agilent.com/litweb/pdf/5989-2589EN.pdf和Keysight 85071E软件
http://cp.literature.agilent.com/litweb/pdf/5988-9472EN.pdf根据您的频率，将其置于波导中可能是最好的方法。
我从未尝试过这种方法，但我想到的一种方法是将薄片放在APC7插头上，背后有无限金属导体。
我确信人们可以从中确定材料，尽管人们可能需要EM模拟器来确定S11测量的介电特性。
如果材料可以被认为是无限厚的，那么使用同轴探头是很常见的，但这对于像你这样的薄板来说是行不通的。
我想知道是否还有其他人想过一个带有薄片和无限大小地平面的同轴探头。
我怀疑我是第一个想到它的人，我相信它可以产生有用的信息。
人们甚至可以考虑在两个APC7连接器之间放一张薄片，并测量S11和S21。
同样，虽然人们可能很容易获得一些数据，但将其转换为任何有用的数据会更成问题。
戴夫



     以下为原文

  > {quote:title=igouzouasis wrote:}{quote}
> Hi,
> 
> I have an issue at my lab and would like to discuss it with you. I'd be grateful if you could provide me with any help.
> 
> Suppose you want to measure the reflection and transmission properties of a thin dielectric sheet in terms of S11 and S21. We have only a VNA in our lab and a few broadband microstrip antennas (not very directive), designed to operate at the frequencies we want the dielectric to be measured. 
> 
> The simplest of the ideas would be to perform a full 2-port calibration with two antennas in place, one connected to each port, then insert the sheet in between and make S11 and S21 measurement. 
> 
> But how is a full 2-port calibration with antennas connected performed? I couldn't find any info in application notes.
> 
> Thanks
> Yannis

I don't think you can do what you envisage doing. To do a full 2-port calibration you would need to put 3 standards on each port, and if the port is now an antenna, I can't see how that is possible. An obvious exception being if you can make the antenna a waveguide to coax transition, rather than a microstrip one. In that case you would need a waveguide calibration kit. 

That said, measuring the dielectric properties between two antennas has been used before. You get the problem that the antennas need to be in the far field, and for that you are likely to need quite a bit sample. 

There are a number of ways of measuring the dielectric properties of materials. Take a look at 

http://cp.literature.agilent.com/litweb/pdf/5991-2171EN.pdf
http://cp.literature.agilent.com/litweb/pdf/5989-2589EN.pdf

and Keysight's 85071E software

http://cp.literature.agilent.com/litweb/pdf/5988-9472EN.pdf

Depending on your frequency, putting it in a waveguide might well be the best way. 

I've never tried it, but one method I have thought about is to put a thin sheet against and APC7 plug, with an infinite metal conductor behind it. I'm sure one could determine something about the material from that, although one would probably need an EM simulator to determine the dielectric properties from an S11 measurement. Using a coaxial probe is quite common if the material can be considered infinitely thick, but that is not going to work for a thin sheet like you have. 

I wonder if anyone else has thought about a coax probe with a thin sheet an infinite size groundplane. I doubt I am the first to think of it, and I'm fairly confident that it could yield useful information. 

One could even think about putting a thin sheet between two APC7 connectors, and measuring both S11 and S21. Again, although one could probably get some data easy enough, converting it to anything useful would be more problematic. 

Dave

戴夫非常感谢你的信息。
如果没有忽略Dave的方法，有没有办法用微带天线进行校准？
这些精确的天线将成为实验设置的一部分，无需更改。
GRL是否可能通过？



     以下为原文

  Dave thank you a lot for your info. 

Without ignoring Dave's methodolody, is there a way to do the calibration with the microstrip antennas in place?

These exact antennas will be part of an experimental setup and need not to be changed.

Is GRL a possible way through?

> {quote：title = igouzouasis写道：} {quote}>戴夫非常感谢你的信息。
>>不忽略Dave的方法，有没有办法用微带天线进行校准？
>>这些精确的天线将成为实验设置的一部分，无需更改。
> GRL是否可能通过？
可能。
我们的材料专家本周已经出版，但我们的材料测量软件确实支持GRL方法，我不知道天线类型是否有限制。



     以下为原文

  > {quote:title=igouzouasis wrote:}{quote}
> Dave thank you a lot for your info. 
> 
> Without ignoring Dave's methodolody, is there a way to do the calibration with the microstrip antennas in place?
> 
> These exact antennas will be part of an experimental setup and need not to be changed.
> 
> Is GRL a possible way through?

possibly. Our materials expert is out this week, but our materials measuerment software does support the GRL method and I don't know if there is limitation on the antenna type.

谢谢乔尔。
请记住，我不想测量ε和μ。
我只想在一定频率范围内对板坯进行S11和S21测量。
为此，我需要通过执行校准来消除天线施加的任何误差。
然后，我将知道S11和S21测量仅归因于板坯。



     以下为原文

  Thanks Joel.

Please bear in mind that I don't want to measure ε and μ. I just want to take S11 and S21 measurements of the slab over a range of frequencies. 

To do this, I need to remove any error imposed by the antennas by performing the calibration. Then, I'll know that the S11 and S21 measurements will be attributed to the slab only.

只是一个对此知之甚少的人的想法，但喜欢挑战性的问题！
你可以a）校准用普通的校准套件哄骗。
b）制作两个*相同的*微带天线，连接到同轴电缆。
c）将两个天线放置一些适当的距离，以便可以假设远场条件。
我们假设它是X米。
c）测量S21和S11。
（显然，S22 = S11，S12 = S21）d）假设S21的一半相移和S21幅度的一个*四分之一由一个天线和自由空间的X / 2的距离引起。
（我认为它需要是幅度的1/4，因为功率随着距离的平方而下降，但我建议你仔细检查我的逻辑是否正确）。
S11的值不需要任何这样的校正。
e）在两个天线之间插入厚度为T的电介质e）将天线之间的间距增加到T + X. f）再次测量所有4个S参数。
g）考虑解嵌电介质的问题，其中电介质插入两个夹具之间，两个夹具都知道它们的S参数。
固定装置恰好是一个天线加上X / 2米的自由空间。
如果设备上有电连接器，我只考虑S参数有任何意义，所以我不确定考虑天线和X / 2米自由间距有4个S参数是多么有效。
通常我只会考虑天线有S11，但如果你把它与自由空间结合，*也许*也可以有S21，S12和S22。
也许合理的是考虑S22 = 0的天线加上X / 2米的空间，因为在自由空间接口处没有反射，我假设它是端口2.所以天线+ X / 2米的
自由空间有* S11你测量天线* S22 = 0 * S12 = S21，通过测量两个相同的天线，其间有自由空间计算得出。
无论如何，只是我的想法。
（我甚至不能使用太多的啤酒来提出这样一个疯狂的想法）DaveEdited：drkirkby于2014年10月16日下午1:45



     以下为原文

  Just a thought from someone who knows very little about this, but enjoys challenging problems! 

Could you 

a) Calibrate to coax with a normal cal kit. 

b) Make two *identical* microstrip antennas which connect to coax. 

c) Place the two antennas some suitable distance apart, such that far field-conditions could be assumed. Let's assume it is X metres. 

c) Measure S21 and S11. (Obviously S22=S11 and S12 =S21)

d) Assume that half the phase shift of S21, and one *quarter* of the magnitude of S21 is caused by one antenna and a distance of X/2 of free space. (I'm thinking it needs to be 1/4 the magnitude, as power falls with the square of distance, but I suggest  you double-check if my logic is right). The values of S11 don't need any such correction. 

e) Insert the dielectric of thickness T between the two antennas 

e) Increase the spacing between the antennas to T + X. 

f) Measure all 4 S-parameters again. 

g) Consider the problem one of deembedding the dielectric, where the dielectric is inserted between two fixtures, both of which you know the S parameters of. The fixtures happen to be one antenna plus X/2 metres of free space. 

I've only ever considered S-parameters having any meaning if a device has electrical connectors on it, so I'm not sure how valid it would be to consider an antenna and X/2 metres of free spacing having 4 S-parameters. Normally I would only consider the antenna having S11, but if you combine it with free space, *maybe* it can have S21, S12 and S22 too. 

Maybe it is reasonable to consider S22=0 of the antenna plus X/2 metres of space, as there would be no reflection at the free space interface, which I'm assuming is port 2. So the antenna + X/2 metres of free space has

* S11 you measure of the antenna
* S22 = 0
* S12 = S21, computed from measuring two identical antennas with free space in between. 

Anyway, just my thoughts. (I can't even use the excuse I have had too many beers to come up with such a crazy idea)

Dave

Edited by: drkirkby on Oct 16, 2014 1:45 PM

嗨，戴夫，谢谢你的努力，我认为没有*相同的*天线。
实际上，由于制造公差，天线是不同的。
一旦我们试图测量S21测量中的微小偏差，我们就不能假设两个天线是相同的。
我们系统所需的灵敏度非常高，这使得事情变得复杂，因为天线不是传输线的正确替代品。
但是，我们正努力在受控，无干扰的环境中进行测量，在这种环境中，一切尽可能保持稳定。
作为第一步，我希望专家向我保证，根据我们的限制，GRL是一种正确的方法。
如果其他实验室成员设置了相同或类似的设置，我将非常感激分享建议和想法。
谢谢



     以下为原文

  Hi Dave and thanks for your effort,

I think that there's not such thing as *identical* antennas. In practice, antennas are different because of fabrication tolerances. And as soon as we're trying to measure slight deviations in S21 measurements, we can't assume that the two antennas are identical. 

The sensitivity required for our system is very high and this makes things complicated as antennas are not a proper substitute of a transmission line. However, we're striving to perform the measurements in a controlled, interference-free environment, where everything is kept as stable as possible.

As a first step, I want the experts to assure me that GRL is a proper way to go through based on our limitations. 

If the same or similar setup has been .. set up by other lab members, I'd be grateful to share advice and ideas.

Thanks

> {quote：title = igouzouasis写道：} {quote}>嗨Dave，谢谢你的努力，>>我认为没有*相同*天线这样的东西。
实际上，由于制造公差，天线是不同的。
一旦我们试图测量S21测量中的微小偏差，我们就不能假设两个天线是相同的。
很公平。
我想值得说一下“轻微”是​​什么，以及VNA的频率，型号和选项，你的校准套件等，因为否则会有很多未知数。
希望Keysight的材料专家能够为您提供帮助。
如果是我认为的那样，我上周见过她。
戴夫



     以下为原文

  > {quote:title=igouzouasis wrote:}{quote}
> Hi Dave and thanks for your effort,
> 
> I think that there's not such thing as *identical* antennas. In practice, antennas are different because of fabrication tolerances. And as soon as we're trying to measure slight deviations in S21 measurements, we can't assume that the two antennas are identical. 

Fair enough. 

I guess it would be worth saying what "slight" is, along with the frequency, the model and options of the VNA, your cal kits etc, since otherwise there are a lot of unknowns.  

Then hopefully Keysight's materials expert will be able to help you. If it is who I think it is, I met her last week. 

Dave

轻微的偏差意味着0.1dB。
频率为60GHz。
我们实验室的VNA似乎没有安装选件100（它是经过翻新的Agilent HP E8361C）。
我们应该能够将其发送到keyight，以便在那里安装选项吗？
任何有关校准方法的更新都将受到欢迎。
谢谢



     以下为原文

  Slight deviations would mean 0.1dBs.
Frequency at 60GHz.

The VNA of our lab doesn't seem to have the option 100 installed (it's a refurbished Agilent HP E8361C). Should we be able to send it to keysight, to have the option installed there? 

Any update on the calibration method would be welcome.

Thanks

要迁移，sourceId：117435



     以下为原文

  to be migrated, sourceId: 117435

那个研讨会真的很有用！
谢谢你。
据我所知，选项010是计算被测板坯介电特性的软件。
那么GRL校准本身呢？
我认为选项100支持它。因此，选购100可以在购买后安装在仪器中，还是应该从一开始就安装？



     以下为原文

  That seminar was really informative!
Thank you for that.

As far as I understood, option 010 is the software that calculates the dielectric properties of the slab under test.
What about the GRL calibration itself? I think it's supported in option 100.

So, can option 100 be installed in an instrument after its purchase or it should have been installed from the beginning?

我认为软件包和分析仪的选项之间存在一些可以理解的混淆... 85071E材料测量软件（基本软件包）可用于该软件的选项：85071E-100自由空间校准许可证（增加GRL校准功能）85071E
-200拱形反射率许可证（NRL拱形反射率，此处不需要）85071E-300谐振腔体许可证（SPDR和分离圆柱谐振器，此处不需要）85071E-400高级功能升级（导入和显示s参数等，不需要）
当与GRL校准一起使用时，E8361C需要一个选项：E8361C选项-010时域转换功能基本的85071E可以进行自由空间测量和显示介电特性，但这并不是您所需要的。
据我了解，您只需要对样品s参数进行自由空间测量。
难以进行良好的校准......如果您想使用85071E软件的GRL功能来帮助校准，那么您需要三件事：基本的85071E软件+ 85071E-100选项+ E8361C选项-010你可以
然后在85071E GRL的帮助下进行校准，然后测量您的deive并直接从分析仪保存s参数数据。
85071E-100需要在分析仪上安装E8361C-010：支持的分析仪您可以通过单击帮助，关于网络分析仪检查分析仪上的已安装选项，可以通过订购E8361CU-010将时域功能添加到您的分析仪（
“U”意味着升级）当然，如果您想在没有GRL的情况下尝试校准，您可以在没有任何额外软件的情况下完成此操作。
只需定义一个TRL或LRM校准....需要一些额外的工作:)



     以下为原文

  I think there is some understandable confusion between the options for the software package and the analyzer...

85071E Materials Measurement Software (the basic software package)
available options to this software:
85071E-100 Free space calibration license (adds GRL calibration capability)
85071E-200 Arch Reflectivity License  (NRL Arch reflectivity, not needed here)
85071E-300 Resonant Cavity License (SPDR and Split Cylinder Resonators, not needed here)
85071E-400 Advanced Functionality Upgrade (import & display s-parameter, etc, not needed here)

One option is required for the E8361C when used with the GRL calibration:
E8361C option -010 time-domain transform capability

The basic 85071E can make a free space measurement and display dielectric properties, but this is not exactly what you need. As i understand it, you just need to make a free-space measurement of your samples s-parameters. The difficulty would be making a good calibration...

If you want to use the GRL capability of the 85071E software to help with calibration, then you need three things: 
The basic 85071E software + 85071E-100 option + E8361C option -010

You could then calibrate with the help of the 85071E GRL, then measure your deive and save the s-parameter data directly from the analyzer. 

The 85071E-100 requires E8361C-010 to be installed on the analyzer: Supported Analyzers

you can check installed options on your analyzer by clicking on Help, About Network Analyzer
the time-domain capablity can be added to your analyzer by ordering E8361CU-010  (the "U" means an upgrade)

of course, if you want to try the calibration without GRL, you can do this without any extra software. Just define a TRL or LRM calibration....some additonal work required  :)

确实，雷，这是我正在寻找的信息。
非常感谢你指出这一点。
真有帮助。
关心Yannis



     以下为原文

  Indeed, Ray, that was the info I was looking for.

Thanks a lot for pointing that out. Really helpful.

Regards
Yannis