pna x梳状滤波器发生器校准相位与功率测量似乎在不同校准的特定功率水平下获得不同的相位值读数

SMC +相位参考校准，提供了一种校准各个接收器的相位响应的方法，但不是绝对相位。
所以即使在校准之后，起始阶段也是随机的，但是表征了从一个频率到下一个频率的相位差。
当您使用SMC进行群延迟测量时，SC21相位响应被标准化为扫描中的一个点。
标准化为的点实际上是设置对话框中的设置之一（见下文）。
一旦相位响应被归一化，我们就可以使用以下公式计算群延迟：！https：//dl.dropboxusercontent.com/u/87949221/Forum/ThreadID_118944/GroupDelay.png！
所以，不 - 你没有得到绝对的群体延迟。
你得到的是归一化相位响应的群延迟。
但是，如果您对AM / PM测量感兴趣，那么起始相位值并不重要。
你所需要的只是相对于功率变化的相位变化。
此外，由于AM / PM是在单一频率下测量的，因此您甚至不需要第一层相位参考校准（除非您还想测量群延迟）。
如果您只对AM / PM感兴趣，那么只需进行正常的SMC校准，但只需启用通道中的相位测量并设置功率扫描。
！HTTPS：//dl.dropboxusercontent.com/u/87949221/Forum/ThreadID_118944/SMCPhase.png！



     以下为原文

  the SMC + Phase Reference cal, provides a way to calibrate the phase response of the individual receivers, but not the absolute phase.  so even after calibration, the starting phase is random, but the phase difference from one frequency to the next is characterized.  when you make the group delay measurements with SMC, the SC21 phase response is normalized to one of the points in the sweep. the point that you normalize to is actually one of the settings in the setup dialog (see below).  Once the phase response is normalized, we then compute the group delay by using the equation:

!https://dl.dropboxusercontent.com/u/87949221/Forum/ThreadID_118944/GroupDelay.png!

So, no - you don't get the absolute group delay.  what you get is the group delay of a normalized phase response.  However, if you are interested in an AM/PM measurement, then the starting phase value is not important.  all you need is the change in phase relative to change in power.  Also because AM/PM is measured at a single frequency, you don't even need the 1st tier phase reference calibration (not unless you also want to measure the group delay).  if you are only interested in AM/PM, then just do a normal SMC calibration, but just enable the Phase measurement in your channel and setup your power sweep.

!https://dl.dropboxusercontent.com/u/87949221/Forum/ThreadID_118944/SMCPhase.png!

嗨，是的，我也正在进行群延迟测量，并且在测试电缆更换和重新校准后我能够获得可重复的结果。
但是当我在起始功率的相位读数方面看到相位与功率测量的差异时，我有点担心我是否正确地进行了校准。
但当然，正如我所怀疑的那样，只有两个感兴趣的功率水平之间的相位差异才是最重要的。
我只是想确认单个相与功率读数本身是随机的并且可以变化，尤其是在测试电缆改变之后。
至于绝对群延迟，我认为agilent / Keysight FAE告诉我SMC + Cal Comb滤波器发生器校准将允许绝对群延迟测量，因为梳状滤波器是NIST可追踪的，但我可能是错的。
从http://literature.cdn.keysight.com/litweb/pdf/5991-1367EN.pdf上的这篇文章中可以看出，VMC技术将允许进行绝对群延迟测量，尽管您有可变元素，例如参考
/校准混频器和所涉及的IF滤波器，其行为可能根据从哪个制造商获得每个元件而变化。
谢谢，rok



     以下为原文

  Hi,

Yes, I'm also taking group delay measurements, and I'm able to get repeatable results after the test cable change and re-cal. But when I saw the difference in the phase vs power measurement in terms of the starting power's phase reading, I was a bit concerned whether I did the cal correctly. But of course as I suspected, only the differences in phase between 2 power levels of interest is what matters. I just wanted to confirm that the individual phase vs power reading by itself is  random and can vary, especially after a test cable change.

As for absolute group delay, I thought an agilent/Keysight FAE told me that the SMC+Cal Comb filter generator cal would allow for absolute group delay measurements because of the Comb Filter being NIST traceable, but I could be wrong. From this article at http://literature.cdn.keysight.com/litweb/pdf/5991-1367EN.pdf, it does state that the VMC technique will allow for absolute group delay measurements, although you have variable elements such as the reference/calibration mixer and the IF filter involved, of whose behavior might vary depending on which manufacturer each of those elements were procured from. 

Thanks,
rok

是的 - 乔尔博士是对的。
我见过不同的人使用“群延迟”，“绝对群延迟”和“相对群延迟”的不同定义，这就是为什么我总是回到我们实际使用的等式和测量过程。
所以我们实际上是采取相位响应并将其归一化到一个点（由用户选择），然后将群延迟方程应用于该数据。



     以下为原文

  Yes - Dr. Joel is correct.  I've seen different people use different definitions of "group delay", "absolute group delay" and "relative group delay", which is why I always go back to the equation and the measurement process we actually use.  so we are in fact taking a phase response and normalizing it to a point (selected by the user) and then applying the group delay equation to that data.

> {quote：title = daras写道：} {quote}>所以，不 - 你没有得到绝对的群延迟。
你得到的是归一化相位响应的群延迟。
>！https：//dl.dropboxusercontent.com/u/87949221/Forum/ThreadID_118944/SMCPhase.png！
我认为daras是错误的。
我们在SMC中有绝对的群延迟测量。
我们没有绝对相位测量，主要是因为在转换器上，要知道从输入到输出的相位转换，你必须包括LO相位的影响，我们不在SMC中测量（但我们测量它
PNA的NVNA混频器模式）。
在SMC中，我们通过锁相到输出信号（在软测量中）来克服这一点，以在某个测量频率点建立参考相位点，但您必须启用相位并设置相位点。
更换电缆将改变LO的绝对相位，无需重新锁定，输出的相位将随LO变化而变化（应该如此）。



     以下为原文

  > {quote:title=daras wrote:}{quote}
> So, no - you don't get the absolute group delay.  what you get is the group delay of a normalized phase response. 
> !https://dl.dropboxusercontent.com/u/87949221/Forum/ThreadID_118944/SMCPhase.png!

I think daras mispoke.  We DO have absolute group delay measuerments in SMC.  We don't have absolute phase measurements, primarily because on a converter, to know the phase transfer from input to output you have to include the effects of the LO phase, which we don't measure in SMC (but we do measure it in the NVNA mixer-mode of the PNA).  In SMC we overcome this by phase-locking onto the output signal (in a softwre sense) to establish a reference phase point at some measurement frequency point, but you have to enable the phase and set the phase point.

Changing a cable will change the absolute phase of the LO and without re-locking, the phase of the output will shift by the LO change (as it should).

是的 - 乔尔博士是对的。
我见过不同的人使用“群延迟”，“绝对群延迟”和“相对群延迟”的不同定义，这就是为什么我总是回到我们实际使用的等式和测量过程。
所以我们实际上是采取相位响应并将其归一化到一个点（由用户选择），然后将群延迟方程应用于该数据。



     以下为原文

  Yes - Dr. Joel is correct.  I've seen different people use different definitions of "group delay", "absolute group delay" and "relative group delay", which is why I always go back to the equation and the measurement process we actually use.  so we are in fact taking a phase response and normalizing it to a point (selected by the user) and then applying the group delay equation to that data.

是的 - 乔尔博士是对的。
我见过不同的人使用“群延迟”，“绝对群延迟”和“相对群延迟”的不同定义，这就是为什么我总是回到我们实际使用的等式和测量过程。
所以我们实际上是采取相位响应并将其归一化到一个点（由用户选择），然后将群延迟方程应用于该数据。



     以下为原文

  Yes - Dr. Joel is correct.  I've seen different people use different definitions of "group delay", "absolute group delay" and "relative group delay", which is why I always go back to the equation and the measurement process we actually use.  so we are in fact taking a phase response and normalizing it to a point (selected by the user) and then applying the group delay equation to that data.

是的 - 乔尔博士是对的。
我见过不同的人使用“群延迟”，“绝对群延迟”和“相对群延迟”的不同定义，这就是为什么我总是回到我们实际使用的等式和测量过程。
所以我们实际上是采取相位响应并将其归一化到一个点（由用户选择），然后将群延迟方程应用于该数据。



     以下为原文

  Yes - Dr. Joel is correct.  I've seen different people use different definitions of "group delay", "absolute group delay" and "relative group delay", which is why I always go back to the equation and the measurement process we actually use.  so we are in fact taking a phase response and normalizing it to a point (selected by the user) and then applying the group delay equation to that data.

是的 - 乔尔博士是对的。
我见过不同的人使用“群延迟”，“绝对群延迟”和“相对群延迟”的不同定义，这就是为什么我总是回到我们实际使用的等式和测量过程。
所以我们实际上是采取相位响应并将其归一化到一个点（由用户选择），然后将群延迟方程应用于该数据。



     以下为原文

  Yes - Dr. Joel is correct.  I've seen different people use different definitions of "group delay", "absolute group delay" and "relative group delay", which is why I always go back to the equation and the measurement process we actually use.  so we are in fact taking a phase response and normalizing it to a point (selected by the user) and then applying the group delay equation to that data.