SMC+相位和VMC测量之间的比较

相位偏差将是相同的（受校准方法的相对精度限制），但是由于LO的相位可能存在固定的相位偏移。
在SMC中，所有相位迹线在一个点处被归一化以消除由于LO而产生的相位偏移，这是提供稳定相位显示所需的（否则相位显示将针对每次扫描处于不同的固定偏移）。
在VMC中，LO的相位偏移归一化为R通道混频器的相位响应。
因此，如果您在R通道混频器LO路径中添加了一条延迟线，重新校准，您仍然会看到测试混频器的VMC相位似乎发生偏移，即使测试混频器没有变化，并且校准发生在
更改。
这是因为recirpocal cal方法不会消除测试或参考混频器的LO路径延迟的变化。
解释原因很复杂......但是我的书第7章对此进行了解释！
总结：如果你需要匹配相位偏差，两者都可以。
如果您需要匹配绝对相位（不太可能，除非这是一个多通道雷达应用程序），您将需要使用VMC ....并且您必须在与测试混音器完全相同的设置上重新测量“黄金”混音器以获得
公平的比较。
如果您在一个系统上测量“黄金”混频器，并且在另一个系统上测试，则无法确保LO路径相同，并且任何差异都将反映在固定相位（平坦w.r.t.freq）偏移中。
如果您不关心平坦相位偏移，请使用SMC +相位。
最后，如果您可以为VMC中的参考和测试混频器提供相同的LO（不使用嵌入式LO），那么VMC相位迹线上的噪声通常比SMC小约6倍（这在SMC中通过添加固定
平均）。
最后，我的经验是，VMC校准混频器的相位表征不如SMC的相位参考校准方法，因此您可能会因校准问题而看到更多错误。
宽带混频器尤其如此。
当然，这些方法只适用于FIXED LO ......如果你想要扫过LO，那就是另一种讨论。



     以下为原文

  The phase deviation will be the same (limited by the relative accuracy of the calibration methods), but there may be a fixed phase offset due the the phase of the LO.

In SMC, all the phase traces are normalized at one point to remove the phase offset due to the LO, which is needed to provide stable phase displays (otherwise the phase display would be at a different fixed offset for each sweep).

In VMC, the phase offset of the LO is normalized to the phase response of the R-channel mixer.  So, if you added a delay line into the R channel mixer LO path, recalibrate, you will still see the phase of the VMC for the test mixer appear to shift, even though no change was made to the test mixer and calibrate occured after the change.  This is because the recirpocal cal method does not remove changes to the delay of the LO path of the test or reference mixer.  It's complicated to explain why... but it is explained in chapter 7 of my book!

To conclude:  If you need to match phase deviation, both will work.  If you need to match absolute phase (unlikely, unless this is a multichannel radar application) you will need to use VMC....AND you must re-measure the "golden" mixer on the exact same setup as the test mixer to get a fair comparison. If you measured the "golden" mixer on one system, and the test on another, you cannot ensure the LO path is identical, and any difference would be reflected in a fixed phase (flat w.r.t. freq) offset.  If you don't care about flat phase offset, use SMC+phase.

Finally, if you can provide the same LO to both the reference and test mixer in VMC (not using an embedded LO), then the noise on the phase trace of VMC is usually about 6x smaller than SMC (this is fixed in SMC by adding averaging).

And finally, finally, my experience is the phase characterization for the calibration mixer for VMC is not as good as the phase reference calibration method for SMC, so you may see more error due to problems with calibration.  this is especially true with wideband mixers.

And, of course, these methods only work with FIXED LO...if you want swept LO, that's a different discussion.

Dr.Joel，我们打算对混合器进行表征，这些混合器是相位匹配的。
我们想要测量每个混合器和混合器之间相位响应最小的混合器。
这将用于多通道接收器。我们可以使用SMC +阶段应用程序吗？
目前我们正在使用VMC通道，然后，我们将连接的第一个DUT的相位标准化，而不进行任何校准。
然后我们连接下一个DUT，并看到相位偏差与标准化迹线相关。
这是一个正确的方法吗？
RGDS



     以下为原文

  Dr.Joel,

We intend to characterize mixers and these mixers are phase matched. We want to measure the phase response of  each of the mixers and group mixers that have the least devaition between one another. This is to be used in a multichannel receiver.Can we do this with the SMC+phase application?

Currently we are using a VMC channel and then, we normalize the phase of the 1st DUT that is connected without doing any cal. Then we connect the next DUT and see the phase deviation wrt to the normalized trace. Is this a correct method?

Rgds

> {quote：title = nsm写道：} {quote}> Dr.Joel，>>我们打算对混合器进行表征，这些混合器是相位匹配的。
我们想要测量每个混合器和混合器之间相位响应最小的混合器。
这将用于多通道接收器。我们可以使用SMC +阶段应用程序吗？
>这是一个常见的请求，但通常是定义的。
您使用的是固定LO，还是扫描RF和IF？
如果是这样，您可以使用SMC。
如果你正在扫除LO，那么我们还有更多要谈论的......你是否在扫除LO？
>目前我们正在使用VMC通道，然后，我们将连接的第一个DUT的相位标准化，而不进行任何校准。
然后我们连接下一个DUT，并看到相位偏差与标准化迹线相关。
这是一个正确的方法吗？
>此方法可以很好地工作，但如果不使用VMC纠错，混频器端口上的不匹配将导致明显的偏差。
让我解释。
如果PNA的端口阻抗为50欧姆（或经过校准），其中一个混频器为52欧姆，则频率扫描会显示纹波。
如果另一个混频器有48欧姆，它也会显示纹波，但相位相反。
所以比较是有效的（a-b）。
但如果PNA不是50欧姆（或未校准），则假设PNA的阻抗为52欧姆。
第一个混音器将不显示纹波，第二个将显示2x纹波。
差异与上面的情况相同，但在这种情况下，一个看起来平坦而另一个看不平。
因此，为了进行比较，两者都可行。
在测试端口电缆上添加一些焊盘（6 dB是好的）将有助于减少未校准的不匹配效应。



     以下为原文

  > {quote:title=nsm wrote:}{quote}
> Dr.Joel,
> 
> We intend to characterize mixers and these mixers are phase matched. We want to measure the phase response of  each of the mixers and group mixers that have the least devaition between one another. This is to be used in a multichannel receiver.Can we do this with the SMC+phase application?
> 
This is a common request but often pooly defined.  Are you using a fixed LO, and sweeping the RF and IF?  If so, you can use SMC.  If you are sweeping the LO, then we have much more to talk about...Are you sweeping the LO or not?
>  Currently we are using a VMC channel and then, we normalize the phase of the 1st DUT that is connected without doing any cal. Then we connect the next DUT and see the phase deviation wrt to the normalized trace. Is this a correct method?
> 
This method will work well, but mismatch on the mixer ports will cause an apparent defiation if VMC error correction is not used.  Let me explain. If the PNA has a port impedance of 50 ohms (or is calibrated), and one of the mixers is 52 ohms, it will show a ripple on the frequency sweep.  If another mixer has 48 ohms it will also show a ripple, but opposite phase.  So the comparison will be valid (a-b). 

But if the PNA is not 50 ohms (or is not calibrated) then suppose the PNA has an impedance of 52 ohms.  The first mixer will show no ripple, the second will show 2x ripple. The difference is the same as the case above, but in this case one looks flat and other does not.

Thus, for comparison purposes, either will work. Adding some pads to the test port cables (6 dB is good) will help reduce uncalibrated mismatch effects.

Dr.Joel，很抱歉包含这些信息。
我们使用固定LO进行测量。
那么在SMC +相位测量对话框中，相位参考点是否定义了迹线必须归一化的点，以消除由于LO引起的相位偏移？
因此，对于我的测量应用，如果我要对我的混频器（DUT）上的所有测量值保持这种共性，那么相位差将给出指示混频器上的相位如何匹配的情况.Rgds



     以下为原文

  Dr.Joel,

I am sorry to include the information. We are using a fixed LO for our measurements.

so in the SMC+phase measurements, the dialog box, does the phase reference point define the point at which the trace has to be normalized for removing the phase offset due to the LO? 

SO, for my measurement application if I were to keep this common for all measurements on my mixers(DUTs), then difference in phase will give me the indication of how the phases are matched on the mixers..

Rgds

> {quote：title = nsm写道：} {quote}>所以在SMC +相位测量，对话框中，相位参考点是否定义了迹线必须归一化的点，以消除相位偏移
LO？
这是原始数据规范化的地方。
关闭校正后，它应该返回零。
通过校正，它将固定，但不是零，因为校正可能由于响应跟踪和不匹配项而具有一些相位偏移。
>那么，对于我的测量应用，如果我要对我的调音台（DUT）上的所有测量保持这种共性，那么相位差将给出指示如何在混音器上匹配相位。>它将显示相对
搅拌机之间的相位响应。
但是，如果你在绝对阶段受到影响，那么由于LO会产生偏移。
考虑一下：对混频器进行SMC测量，将其保存在存储器中。
然后改变LO信号的相位。
输出的阶段 - 将改变（所有点将被LO阶段偏移，或者如果你是高端LO则为负）。
相对相位响应（相位纹波与频率）不会改变。
在SMC中，归一化不会显示LO变化，即使改变LO的相位也会看起来相同。
在VMC中，如果您改变LO进入测试混频器的相位（但仅保留参考），您将看到VC21走线上的相移。
如果是我，做这个比较，我会使用VMC进行比较，因为LO的相位噪声效果被参考混频器取消，因此可以进行更清晰的测量。
我会做它没有修正，并在测试电缆的末端使用一个6 dB的焊盘，就在它们连接到测试混频器的地方。
这假设您可以轻松访问参考混音器。
使用端口3驱动测试混频器和端口4以驱动参考混频器。
这将很容易并且运作良好。



     以下为原文

  > {quote:title=nsm wrote:}{quote}

> so in the SMC+phase measurements, the dialog box, does the phase reference point define the point at which the trace has to be normalized for removing the phase offset due to the LO? 
This is where the raw data is normalized.  With correction off, it should return zero. With correction on, it will fixed, but not zero as the correction can have some phase offset due to response tracking and mismatch terms.
> SO, for my measurement application if I were to keep this common for all measurements on my mixers(DUTs), then difference in phase will give me the indication of how the phases are matched on the mixers..
> 
It will show the relative phase response between mixers.  But, if you are intereseted in the absolute phase, then there will be an offset due to the LO.

Consider this: take an SMC measurement of a mixer, save it in memory.  Then change the phase of the LO signal.  The phase of the output -will- change (all the points will be offset by the LO phase, or the negative if you are high-side LO).  The relative phase response (phase ripple vs. frequency) will not change.  And in SMC, the normalization will not show the LO change, the will look the same even if you change the phase of the LO.

In VMC, if you change the phase of the LO going to the test mixer (but leave the reference alone) you will see a phase shift on the VC21 trace.

If it were me, doing this comparision, I would use VMC to do the compare as the phase noise effect of the LO is canceled by the reference mixer and so cleaner measurments are possible.  I would do it uncorrected and use a 6 dB pad on the end of the test cables, just where they connect to the test mixer.  This presumes you have easy access to a reference mixer.  Use port 3 to drive the test mixer and port 4 to drive the reference mixer.  It will be easy and work well.

Dr.Joel，我使用的是N5222A fw ver 9.80.2,2源，4端口NA我们正在尝试比较VMC和SMC +相位测量。
我们想要取消参考混频器和校准混频器/滤波器，SMC +阶段将简化我们的测试。
我们此时的目的是纯粹比较混合器的相位输出，我们希望为实验室中的人员建立一个程序，使用SMC +阶段简化，尽管步骤数量更多。
请找到我所做的测量。
我附上两个混音器的csv文件。
这有SMC +阶段和VMC结果。
VMC测量清楚地显示相对于混合器（SL no 32387）的大约11至12度的相位差。
然而，SMC +相位结果未显示任何相位差异。
这令人费解。
有什么我想念的吗？
我使用sl没有32387的混频器作为校准混频器。我使用它作为SMC +相位和VMC测量的参考DUT。
谢谢大家的支持。
RGDS
32387.csv22.0 KB32383.csv22.0 KB



     以下为原文

  Dr.Joel,
I am using the N5222A fw ver 9.80.2, 2 source,4 port NA
We are trying to compare the VMC and SMC+phase measurements. We want to do away with the reference mixer and cal mixer/filter and the SMC+phase will simplify our testing. Our intention at this point is to purely compare the phase output of the mixers and we want to establish a procedure for the people in our lab to use the SMC+phase for its simplicity albeit the number of steps are more.
Please find the measurements that I have done. I am enclosing the csv files of two mixers. This has the SMC+phase and the VMC results. The VMC measurement clearly shows a phase difference of around 11 to 12 deg with respect to the mixer (SL no 32387). However the SMC+phase result does not show any difference in phase. This is puzzling. Is there anything that I am missing?
I have used the mixer with sl no 32387 as the calibration mixer.I am using this as the reference DUT  for the SMC+phase and VMC measurement. 

Thank you all the support.

Rgds

附件

• 32382.csv22.3 KB
  
• 32387.csv22.0 KB
  
• 32383.csv22.0 KB
  

感谢csv情节，但老实说，我们在excel中加载它们，然后创建图表，看看你的意思是很多工作。
您可能会创建一个小报告，以便下次为我做这件事。
我对数据感到困惑：SC21中有图表，然后是每个csv中的tr3 / mem和tr4 / mem。
你能告诉我tr3和tr4的轨迹是什么（它们看起来像是相对于其他混音器，但即使是大数据看起来很奇怪（你说32387是用于相位比较的参考调音台，但它的tr3 / mem数据不是
我正如我所料，当您使用此调音台作为校准混音器时，您是否也在校准期间使用低通滤波器（我认为您在降频转换模式下使用）。测量期间滤波器是否打开？
并且，您是否在VMC测量中使用滤波器？您是否校准了VMC和SMC测量？这些都是每个校准的重新校准吗？内存轨迹是什么？编辑：Dr_joel于2013年9月16日上午7:51



     以下为原文

  Thanks for the csv plots, but honestly, they are a lot of work on my part to load them in excel, then create charts, to see what you mean.  You might create a little report that does that for me next time.

I'm confused by data: there is plots of SC21, and then tr3/mem and tr4/mem in each csv.  Can you tell me what are the tr3 and tr4 traces (they look like they are relative to some other mixer, but even the magn data looks odd (you say 32387 is the reference mixer for phase comparision, but it's tr3/mem data is not 1,0 as I would expect).
when you used this mixer as the calibration mixer, did you also use a low pass filter ( I presume you used in in down conversion mode) during the cal.  Is the filter on during the measurement?


And, do you use the filter in the VMC measuerments?  Did you calibrate both VMC and SMC measurements?

Are these the calibrated resutls on each?

What is the memory trace from?

Edited by: Dr_joel on Sep 16, 2013 7:51 AM

Dr.Joel，道歉。
我附上了一份带有图表的文字文件供参考。
如果这些数据有意义，请告诉我。
我已经完成了SMC +相位校准以及VMC校准。但是对于VMC校准，我没有使用任何滤波器。
测量在下转换模式下完成，我没有对幅度数据进行归一化。
我只看相位数据。
SMC +相位数据未显示任何相位变化。
但是，VMC测量显示了相位的变化。
RGDS



     以下为原文

  Dr.Joel,

My apologies. I have attached a word document with the plots for reference. Please let me know if this data makes sense.

I have done a SMC+phase calibration and also the VMC calibration.However for the VMC calibration, I have not used any filter. The measurement is done in the down conversion mode

I have not normalized the magnitude data. I am only looking at the phase data.

The SMC+phase data does not show any variation in phase. However, the VMC measurement shows the variation in phase.

Rgds

附件

• Mixer Phase tracking measurement.docx37.3 KB
  

感谢更清晰的文件。
一些观察结果：两个VMC测量都具有相似的相位响应，但是它们被固定相位偏移，并且这个固定相位可能是由于两个混频器之间的LO路径延迟的微小差异。
两种SMC测量都具有相似的相位响应。
SMC测量人为地消除了固定的相位偏移，因此它们看起来彼此叠加。
如果固定相位偏移很重要，则不能使用SMC。
如果它只是相对相位（偏离相位）那么你可以使用它。
VMC数据中的纹波可能是由于在混频器上没有使用滤波器而导致混频器表征不良造成的。
您必须有一个滤波器，以便在混频器表征期间摆脱图像信号，否则它将是校准数据中的错误。
VMC cal-mixer-characterization依赖于反射上/下转换，图像（RF + LO）必须通过滤波器移除以避免它导致错误。
只有在VMC校准混频器表征期间，您才需要在VMC或SMC测量期间使用滤波器。
那么，最后一个问题，在您的word文档中，您担心VMC在两个测试混频器之间的“标准化”阶段（相对于参考混频器）显示固定偏移？



     以下为原文

  Thanks for the clearer document.  Some observations:

Both VMC measurements haves similar phase responses, but they are offset by a fixed phase, and this fixed phase is likely due to a slight difference in the LO path delay between the two mixers.

Both SMC measuerments ahve similar phase responses.  The SMC measurement artificially removes the fixed phase offset so they appear to land on top of each other.  If the fixed phase offset is important, you cannot use SMC.  If it is only relative phase (deviation from phase) then you can use it.

The ripple in the VMC data is likely caused by a bad mixer characterization due to not using a filter on the mixer.  You must have a filter to get rid of the image signal during the mixer characterization, or it will be an error in the cal data.  The VMC cal-mixer-characterization relies of reflected up/down conversion, an the image (RF+LO) must be removed by a filter to avoid it causing errors.

You don't need a filter during either VMC or SMC measuerments, only during the VMC cal-mixer-characterization.

So, final question, in your word document, is your concern that the VMC shows a fixed offset in the 'normalized' phase (relative to the reference mixer) between the two test mixers?

Dr.Joel，我有一些澄清。
当您参考两个混频器之间的LO路径延迟时，由于混频器的结构，您是指内部的路径延迟吗？
在外部，连接是相同的，当我连接不同的混音器时它们不会改变。
VMC测量显示REF混频器（DUT）的偏移。
这也是我试图使用SMC测量的。
如果我必须解释SMC +相位测量结果，我测量的所有混频器都与线性相位有相同的偏差，这就是我无法看到偏移的原因。
如果是这样，我们如何使用SMC +相来比较混合器的相位响应？
有可能吗？
RGDS



     以下为原文

  Dr.Joel,

I have a few clarifications. When you refer to the LO path delay between the two mixers, are you referring to the path delay  internally due to the construction of the mixer? Externally, the connections are the same and they do not change when I connect the different mixers.

The VMC measurements shows an offset wrt the REF mixer(DUT). This is what I am trying to measure using the SMC as well.

If I have to interpret the SMC+phase measurement results, all the mixers that I have measured have equal deviation from linear phase and is that the reason I am unable to see the offset.

If so how do we use SMC+phase for comparing the phase responses of mixers? Is it possible at all?

Rgds

乔尔博士，我有一些澄清。
在VMC测量中显示的固定偏移量是我尝试使用SMC +相位测量的。这些数据是否表示混频器的相位匹配？
是否不可能使用SMC +阶段进行类似的测量？
当您参考LO路径延迟时，您是指由于混频器结构而导致的内部路径？
从外部开始，连接对于不同的混合器是相同的。
关于SMC +相位数据的解释，这意味着我测量的不同混频器与线性相位具有相同的偏差，因为相位超过了相位？
RGDS



     以下为原文

  Dr Joel,

I have a few clarifications.

The fixed offset that shows in the VMC measurement is what I am trying to measure using SMC+phase.Does this data not indicate the phase matches of the mixers? Is it not possible to do similar measurements using SMC+phase? 

When you refer to the LO path delay are you referring to the internal  path due to the construction of the mixer? Since externally, the connections are the same to the different mixers.

As to the interpretation of the SMC+phase data, soes it mean that the different mixers that I have measured have the same deviation from linear phase since the phase traces over lap?

Rgds

Joel博士，当你提到固定的LO偏移量时，你指的是由于混频器结构导致的延迟吗？
与混频器的外部连接是相同的。
这是否偏移，并不表示不同混频器之间输出的相位不匹配？
我试图使用SMC +阶段应用程序进行类似的测量。
至于SMC +相位结果的解释，是否意味着我测量的混频器与线性相位具有相似的偏差？
RGDS



     以下为原文

  Dr Joel,

When you refer to the fixed LO offset, are you referring to the delay due to the construction of the mixer? The external connections to the mixers are the same. Does this offset, not indicate the phase mismatch at the output between the different mixers? I was trying to do a similar measurement using the SMC+phase application.

As to the interpretation of the SMC+phase results, does it mean that the mixers I have measured have similar deviation from linear phase?

Rgds

Joel博士，当你提到固定的LO偏移量时，你指的是由于混频器结构导致的延迟吗？
与混频器的外部连接是相同的。
这种偏移是否表明不同混频器之间输出的相位不匹配？
我试图使用SMC +阶段应用程序进行类似的测量。
至于SMC +相位结果的解释，是否意味着我测量的混频器与线性相位具有相似的偏差？
RGDS



     以下为原文

  Dr Joel,
When you refer to the fixed LO offset, are you referring to the delay due to the construction of the mixer? The external connections to the mixers are the same. Does this offset not indicate the phase mismatch at the output between the different mixers?

I was trying to do a similar measurement using the SMC+phase application.

As to the interpretation of the SMC+phase results, does it mean that the mixers I have measured have similar deviation from linear phase?

Rgds

Joel博士，当你提到固定的LO偏移量时，你指的是由于混频器结构导致的延迟吗？
与混频器的外部连接是相同的。
这是否偏移，并不表示不同混频器之间输出的相位不匹配？
我试图使用SMC +阶段应用程序进行类似的测量。
至于SMC +相位结果的解释，是否意味着我测量的混频器与线性相位具有相似的偏差？
RGDS



     以下为原文

  Dr Joel,

When you refer to the fixed LO offset, are you referring to the delay due to the construction of the mixer? The external connections to the mixers are the same. Does this offset, not indicate the phase mismatch at the output between the different mixers? I was trying to do a similar measurement using the SMC+phase application.

As to the interpretation of the SMC+phase results, does it mean that the mixers I have measured have similar deviation from linear phase?

Rgds

> {quote：title = nsm写道：} {quote}> Joel博士，>>当你提到固定的LO偏移量时，你指的是由于混频器的构造造成的延迟吗？
与混频器的外部连接是相同的。
这是否偏移，并不表示不同混频器之间输出的相位不匹配？
我试图使用SMC +阶段应用程序进行类似的测量。
>如果混频器的内部结构在LO路径中有轻微的延迟差异，您将看到不同混频器之间的偏移（所有频率处的固定相位偏移）。
该偏移是由于混频器的LO路径中的相移或延迟引起的。
LO路径中的任何延迟差异都不会反映在SMC测量结果中，但会在VMC中>>对于SMC +相位结果的解释，是否意味着我测量的混频器具有与线性相位相似的偏差？
>> Rgds是的，相似但不完全相同。
SMC仅显示相位偏差而非绝对相位。
这是因为输出的相位取决于RF路径的相位响应和LO路径的相位响应。
由于LO是固定的，LO路径的相位再现在VMC中显示为固定偏移;
但它在SMC中被标准化了。



     以下为原文

  > {quote:title=nsm wrote:}{quote}
> Dr Joel,
> 
> When you refer to the fixed LO offset, are you referring to the delay due to the construction of the mixer? The external connections to the mixers are the same. Does this offset, not indicate the phase mismatch at the output between the different mixers? I was trying to do a similar measurement using the SMC+phase application.
> 
If the internal construction of the mixer has a slight delay difference in the LO path, you will see the offset (fixed phase offset at all frequencies) between different mixers.  This offset is due to the phase shift or delay in the LO Path of the mixer.  Any delay difference in the LO path will not be reflected in the SMC measuremetns, but will in the VMC
> As to the interpretation of the SMC+phase results, does it mean that the mixers I have measured have similar deviation from linear phase?
> 
> Rgds

Yes, similar but not identical.  SMC shows only deviation from phase not absolute phase.  This is because the phase of the output depends upon BOTH the phase repsone of the RF path, AND the phase response of the LO path.  Since the LO is fixed, the phase repsonse  of the LO path shows up as fixed offset in VMC; but it is normalized away in the SMC.