使用Smart Cal与Unguided Cal的结果不同

嗨，为了帮助理解为什么会出现结果差异，我有几个问题。
我猜你不是在做TRL类型的cal而是SOLT类型？
当您使用Guided SmartCal时，您为端口选择了哪些DUT连接器（例如，您选择了女性，男性还是其中一个）？
由于大多数校准套件中的Thru被定义为零长度（当性别连接器时为男性对女性），在这种情况下，当您说您的DUT是男性对男性或女性对女性时，默认情况下引导SmartCal将
给你一个Unknown Thru又名SOLR cal。
虽然PNA无引导校准就像ENA校准那样不会询问您在哪个端口上有性别，因此Unguided必须始终假设使用cal kit定义中指定的thru模型进行传统SOLT。
您是否正在使用安捷伦校准套件，其定义内置于PNA和ENA固件？
如果是这样，哪一个？
如果它们不是内置工具包，你可以发布你的PNA'.ckt'文件和ENA'.ckx'工具包的文件吗？



     以下为原文

  Hi, to help understand why that difference in results might occur, I have a couple questions.

I am guessing you were NOT doing TRL type of cal but rather SOLT type?  And when you did the Guided SmartCal, which DUT connectors did you select for the ports (e.g., did you pick both Female, both Male, or one of each)?  Since the Thru in most cal kits is defined as zero-length (Male-to-Female when gendered connectors), in that case when you say your DUT is Male-to-Male or Female-to-Female then Guided SmartCal by default will give you an Unknown Thru a.k.a. SOLR cal.  Whereas PNA Unguided calibration is like ENA calibration in that is does not ask you which gender you have on which port, so Unguided has to always assume to do traditional SOLT using the model of the thru specified in the cal kit definition.

Are you using Agilent cal kit(s) whose definitions are built-in to both the PNA and ENA firmware?  If so, which one(s)?  If they're not built-in kits, could you post your PNA '.ckt' file(s) and ENA '.ckx' files of the kits?

我的DUT是女性 - 女性。
因此，使用Smart Cal，未知通过cal。
使用的calkit是Agilent 85036B。



     以下为原文

  My DUT is female-female. And hence, unknown thru cal with the Smart Cal. The calkit used is the Agilent 85036B.

由于你有一个女性 - 女性DUT，你在校准期间使用的通过量是非零长度。
SmartCal未知通过cal算法考虑到了这一点。
Unguided cal并不是未知的，所以你的实际通过与cal kit定义中描述的通过越不同，你的Unguided cal就越不准确。
85036B / E校准套件定义中的直通定义为零长度（公母 - 假设将两个端口连接器直接连接在一起）。
这可以通过编辑校准套件的定义并查看Thru标准的参数：0 pSec的偏移延迟在PNA上观察到。
请注意，校准套件编辑UI在PNA固件A.10.00.00及更高版本（Windows 7）中进行了一些改进。
以下是用于查看或编辑校准套件定义的适用PNA帮助链接：固件A.09.90.xx及更低版本，固件A.10.00.00及更高版本。
所以这解释了你在PNA上做的两个cals之间存在某些差异的原因，也就是说我期望你在未知通过cal时看到的测量值更准确地显示了你的DUT的响应。
关于为什么你的ENA结果与这两者不同，我对此有所怀疑。
当ENA显示内置安捷伦试剂盒定义之一的校准标准的性别'（m）'或'（f）'时，它实际上是指您的*测试端口*的性别，即相反的
cal标准的性别。
因此，例如当您按下软键表示您正在测量“打开（m）”时，ENA假定您在该端口的末端有一个公连接器，并且您正在连接母打开。
这就是在PNA和ENA出现之前它在所有HP VNA上的表现（VNA开始时的大多数客户必须表明他们当时想要它）。
由于客户习以为常，ENA团队选择继续这样做。
如果当你做了你的ENA校准时，你正在考虑引用标准性别的按钮标签（正如PNA所做的那样），因此cal会有至少一小部分错误。
该套件中的开口和短裤的数学模型在两种性别之间存在一些差异。
（对于PNA的引入，人们认为更多人会直观地期望在标准标签中引用性别来指示标准的性别.PNA帮助中的一个示例位置指出的是Unguided Cal Multiple Standards对话框主题）。



     以下为原文

  Since you have a female-female DUT, the thru you used during your cal of course is nonzero length.  The SmartCal unknown thru cal algorithm takes that into account.  The Unguided cal does not do unknown thru, so the more your actual thru differs from the thru described in the cal kit definition, the less accurate your Unguided cal is.  The thru in the 85036B/E cal kit definition is defined as zero-length (male-female assumes connecting the two port connectors directly together).  This can be observed on the PNA by editing that cal kit's definition and looking at the parameters of the Thru standard: the Offset Delay of 0 pSec.  Note the cal kit editing UI underwent some improvements in PNA Firmware A.10.00.00 and higher (Windows 7).  Here are the applicable PNA Help links for viewing or editing a cal kit definition: Firmware A.09.90.xx and lower, Firmware A.10.00.00 and higher.  So that explains a reason for some difference between the two cals you did on the PNA, i.e. I'd expect the measurement you're seeing with the unknown thru cal is showing more accurately the response of your DUT.

Regarding why your ENA result differed from both of those, I have a suspicion about that.  When the ENA shows a gender '(m)' or '(f)' for a cal standard of one of the built-in Agilent kit definitions, it is actually referring to the gender of your *test port*, i.e. the opposite of the cal standard's gender.  So for example when you push the softkey saying you're measuring "Open (m)", the ENA assumes you have a male connector at the end of that port and you're connecting the female Open.  That is how it went on all HP VNAs before PNA and ENA came along (the majority of customers in the beginning days of VNAs must have indicated they wanted it that way back then).  The ENA team chose to continue with that since customers were accustomed to it.  If when you did your ENA cal you were thinking of the button labels referring to the standards' genders (as PNA does), that cal would have at least a small amount of error for that reason.  The mathematical models of the Opens and Shorts in that kit have some difference between the two genders.  (For the introduction of the PNA it was felt that more people would intuitively expect a gender referenced in a cal standard label to indicate the standard's gender.  One example place in PNA Help this is pointed out is the Unguided Cal Multiple Standards dialog topic).

同样使用ENA，默认情况下它可能不会执行Unknown Thru，因此您可能会遇到类似的错误。
如果确实有错误，则错误的大小可能取决于错误的原因（使用f-f直通而不是零长度或定义通过或未知通过）和原始硬件响应。
在具有良好源和负载匹配的VNA上发生错误不会引起太多波动;
在匹配不佳的VNA上产生相同的错误可能会导致很大的错误。
在几乎所有情况下，智能卡可以提供最好的校准。



     以下为原文

  Also with the ENA, it may not do Unknown Thru by default, so you can get similar errors.

If you do have an error, the size of the error can depend on both the cause of the error (using a f-f thru instead of a zero length or defined thru or unknown thru) and the raw hardware response.  Making an error on an VNA with good source and load match doesn't cause much ripple; making the same error on a VNA with poor match can cause a big error.

In almost all cases, smart cal gives the best possible cal.