PNA-X中增强的响应校准

嗨Rohit，有许多设置会影响增强响应校准（ERC）的质量。
我在你的信息中注意到的第一个红旗是你提到的PNA-X本底噪声为-45 dB。
我认为你的意思是dBm，或者你是说噪声基底比校准期间接收器测量的信号低45 dB？
在合理的IFBW（如1 KHz）下，PNA-X的本底噪声约为-100 dBm。
这意味着如果您以-5 dBm的默认功率水平进行校准并考虑到65 dB的路径损耗，则使用B接收器测量的信号（使用ERC在端口2上进行的唯一测量）将是
大约-70 dBm，仍然比噪声基底高30 dB。
所以第一个问题是你的频道的功率水平和IFBW是多少？
您可以通过增加功率，降低IFBW，增加扫描或点平均或所有3的组合来提高校准期间的信噪比。当增加功率时，如果您使用的话，您需要注意的一件事是
ECal，是为了将功率保持在ECal的压缩点以下（对于大多数较新的型号EC，大约-3 dBm）。
为了超越这种水平的诊断，我们需要更多地了解你的意思“......通过结果并不好（相位和增益）”。
测量是否有噪声（快速随机波纹从扫描变为扫描并随着平均值的增加而下降）或者您是否看到纹波不会从扫描变为扫描并且不会平均？
增益和相位的平均值是否正确或是否存在偏移？
在您的DUT到位后发布一些图片或仪器状态（.csa文件），并在扫描并保持通道保持后将有所帮助。
我可以提出一个建议，但没有任何附加信息，你应该把50 dB衰减器尽可能靠近DUT的输出。
输出端口校准参考平面上的长电缆和50 dB焊盘可能会导致某些不匹配，而这些不匹配将无法通过ERC进行校准。



     以下为原文

  Hi Rohit,

There are a number of settings that will impact the quality of the enhanced response calibration (ERC).  the first red flag that I notice in your message is that you entioned the PNA-X noise floor being -45 dB.  I assumed you meant dBm or are you saying that the noise floor is 45 dB below the signal being measured by the receiver during calibration?  At a reasonable IFBW (like 1 KHz), the noise floor of the PNA-X is around -100 dBm.  This means that if you calibrate at the default power level of -5 dBm and you take into account your 65 dB of path loss, the signal that you are measuring with the B receiver (the only measurement made on port 2 with ERC) will be around -70 dBm, which is still 30 dB above the noise floor.  So the first question is what is the power level and IFBW of your channel?  

You can improve the signal to noise ratio during calibration by either increasing the power, decreasing the IFBW, adding sweep or point averaging or a combination of all 3.  when increasing power, the one thing you need to pay attention to if you are using an ECal, is to keep the power below the compression point of your ECal (for most newer model ECals that is about -3 dBm).

To go beyond this level of diagnosis, we need to know a little more about what you mean by "... the thru results are not good both (phase and gain)".  Are the measurements noisy (fast random ripples that change from sweep to sweep and go down with increased averaging) or are you seeing ripple that does not change from sweep to sweep and does not average out?  Are the average values of Gain and Phase correct or is there an offset?  posting some pictures or your instrument state (.csa file) with your DUT in place and after taking a sweep and putting the channel on hold will help.  

One recommendation I can make without having any additional information that you should put the 50 dB pad as close to the output of your DUT as possible.  A long cable at the output port calibration reference plane and the 50 dB pad could cause some mismatch that will not get calibrated out with ERC.

嗨Daras，谢谢你的回复。
首先，我很遗憾在发布上一个帖子时犯了一个错误。
你写的噪声基底大约-100 dbm（1Khz IFBW）不是-45dBm。
我正在校准我的系统@ 0 dbm功率和1Khz IFBW。我还在校准期间启用了100点平均。
由于我从PNA产生0 dBm并且系统中存在7-8 db的正向路径损耗，因此达到Ecal的功率应该在-7到-8 dBm左右。
模型（N4433a），我建议压缩水平为-7 dBm ......我应该进一步降低功率（比如-5 dbm）或者这没关系？
还有一个问题让我想到这个压缩值应该考虑在内，无论我对Ecal做什么校准......对吧？
我将在明天发布通过响应（校准）的图像，因为PNA今天不适用于我...但它看起来像噪音对我来说，我没有尝试进一步增加平均值并看到响应噪声是否降低
或者保持稳定......为了增加你的推荐点...在50 db衰减器和Dut输出之间，我有一个灵活的4米电缆+2个开关和2个半刚性电缆，每个0.5米。我猜你的是
问题然后我的问题是我无法避免这种电缆，因为它的ATE设置和DUT是保持一点远离ATE。
此致，罗希特。



     以下为原文

  Hi Daras,


Thanks for reply.
First of all I am sorry I made a mistake while posting the previous thread. You are write the Noise floor is around -100 dbm (for 1Khz IFBW) not -45dBm. I am calibrating my system @ 0 dbm power and 1Khz IFBW.I have also enabled 100 point averaging during calibration.

Since I am generating 0 dBm from PNA and a forward path loss of 7-8 db is there in the system so the power reaching to Ecal should be around -7 to -8 dBm. The Model (N4433a) which I have suggest a compression level of -7 dBm ...should I decrease my power further (say -5 dbm) or this is alright??
                           One more question came to mind this compression value should be taken in account irrespective of any calibration I do with my Ecal...right??
                           I will post the images of the thru response (of calibration) by tomorrow as PNA is not available with me today...but it looks likes noise to me, I have not tried further increasing the averaging and see the response whether the noise reduces or remains stable...
                          To add to your recommendation point...between 50 db attenuator  and Dut output, I do have a flexible 4 meter cable + 2 switches and 2 semi rigid cable of 0.5 m each.I guess thi smight be the problem then but my problem is I can't avoid this cables as its an ATE setup and DUT is to be kept little away from the ATE.

Regards,
Rohit.

> {quote：title = Rohit写道：} {quote}>由于我从PNA产生0 dBm并且系统中存在7-8 db的正向路径损耗，因此达到Ecal的功率应该在-7到-7左右 -
8 dBm。
模型（N4433a），我建议压缩水平为-7 dBm ......我应该进一步降低功率（比如-5 dbm）或者这没关系？
我认为如果你知道你有足够的路径损耗（可能想在校准期间用功率计检查它，然后连接ECal），这应该没问题。
>还有一个问题让我想到这个压缩值应该考虑在内，无论我对Ecal做什么校准......对吧？
是的 - 每当您使用ECal进行任何校准时，您都必须注意功率水平。
>为了增加你的推荐点......在50 db衰减器和Dut输出之间，我确实有一个灵活的4米电缆+2个开关和2个半刚性电缆，每个0.5米。我想这是问题然后我的问题
我无法避免这种电缆，因为它的ATE设置和DUT要远离ATE。
>就这样我脑子里有正确的照片 - 在你的设置中，DUT输出端的第一件事是一根4米长的电缆端接到一个开关（我假设它是一个高功率开关 - 对吗？）那是
在这种情况下，有点不幸的设置。
你碰巧知道第一个开关的回波损耗吗？



     以下为原文

  > {quote:title=Rohit wrote:}{quote}
> Since I am generating 0 dBm from PNA and a forward path loss of 7-8 db is there in the system so the power reaching to Ecal should be around -7 to -8 dBm. The Model (N4433a) which I have suggest a compression level of -7 dBm ...should I decrease my power further (say -5 dbm) or this is alright??

I think this should be fine if you know you have enough path loss (might want to check it with a power meter during calibration before hooking up the ECal).

>  One more question came to mind this compression value should be taken in account irrespective of any calibration I do with my Ecal...right??

Yes - whenever you use an ECal for any calibration, you have to be careful about the power levels.

>                           To add to your recommendation point...between 50 db attenuator  and Dut output, I do have a flexible 4 meter cable + 2 switches and 2 semi rigid cable of 0.5 m each.I guess thi smight be the problem then but my problem is I can't avoid this cables as its an ATE setup and DUT is to be kept little away from the ATE.
> 

Just so I have the correct picture in my head - in your setup, the first thing at the output of your DUT is a 4 meter cable terminated into a switch (which I am assuming is a high power switch - right?)  That is a bit of an unfortunate setup in this case.  Do you happen to know the return loss of the first switch?

你好Daras，谢谢你的回复。:) Daras写道：就这样我的头脑中有正确的图片 - 在你的设置中，DUT输出的第一件事是一根4米长的电缆端接到一个开关（我是
假设是一个高功率开关 - 对吗？）在这种情况下这是一个不幸的设置。
你碰巧知道第一个开关的回波损耗吗？
Rohit的回复：我们在Tx模式下以DUT的脉冲模式工作时，DUT的平均功率约为20 W，占空比为10％（最大值）。
我的开关具有超过100W的平均功率能力，所以我认为这不是一个问题。就规格而言，请仔细看看。
Daras，还有一个问题，当我在脉冲模式（在DUT的TX模式下）进行测量时，我是否应该仅使用我的Ecal或CW模式在脉冲模式下进行校准。
因为目前我在CW模式下进行增强响应并使用该calset文件在脉冲模式下进行测量....这种方法是正确的吗？



     以下为原文

  Hello Daras,


Thanks for your reply.:) 

Daras Wrote:
Just so I have the correct picture in my head - in your setup, the first thing at the output of your DUT is a 4 meter cable terminated into a switch (which I am assuming is a high power switch - right?) That is a bit of an unfortunate setup in this case. Do you happen to know the return loss of the first switch?

Rohit's Reply:
Well as we operate in pulse mode for DUT in Tx mode, The average power of DUT will be around 20 W with 10% duty cycle (max). My switch has more than 100W average power capability so I don't think that's an issue.As far as the specs goes its attached,kindly have a look.

Daras, one more question,as I take my measurement in Pulse mode (in TX mode of DUT) should I do calibration also in Pulsed mode only using my Ecal or CW mode. Because currently I am doing enhanced response in CW mode and using that calset file to do measurement in Pulsed mode....is this method right??

附件

• Capture.PNG18.1 KB
  

> {quote：title = Rohit写道：} {quote}> Rohit的回复：>我们在Tx模式下以DUT的脉冲模式工作时，DUT的平均功率约为20 W，占空比为10％（最大值）。
我的开关具有超过100W的平均功率能力，所以我认为这不是一个问题。就规格而言，请仔细看看。
>我并不关心开关的功率处理，但是想知道它的回波损耗是什么，看起来在你运行它的频率下它有大约20 dB的回波损耗，这也不算太糟糕。
然而，它前面的4米电缆仍然可能会导致一些问题，如果没有通过模拟器，很难量化。
> Daras，还有一个问题，当我在脉冲模式（在DUT的TX模式下）进行测量时，我是否应该仅使用我的Ecal或CW模式在脉冲模式下进行校准。
因为目前我在CW模式下进行增强响应并使用该calset文件在脉冲模式下进行测量....这种方法是正确的吗？
我们通常会告诉正在进行宽带脉冲测量的人在CW模式下进行校准，主要是因为它可以更容易地进行功率校准。
如果以脉冲方式进行功率校准，则必须具有峰值功率计或在功率计损耗表中设置偏移，以获得由平均功率传感器测量的脉冲信号的占空比。
s参数校准可以在脉冲或CW中完成，在您的情况下，因为您只进行ERC（无功率校准），所以在脉冲或CW中校准没有区别。



     以下为原文

  > {quote:title=Rohit wrote:}{quote}
> Rohit's Reply:
> Well as we operate in pulse mode for DUT in Tx mode, The average power of DUT will be around 20 W with 10% duty cycle (max). My switch has more than 100W average power capability so I don't think that's an issue.As far as the specs goes its attached,kindly have a look.
> 

I wasn't concerned about the power handling of the switch, but wanted to see what its return loss was and it looks like at the frequencies you are operating it has about 20 dB return loss, which is not too bad.  however, the 4 meter cable in front of it could still cause some problems, which are hard to quantify without running through a simulator.

> Daras, one more question,as I take my measurement in Pulse mode (in TX mode of DUT) should I do calibration also in Pulsed mode only using my Ecal or CW mode. Because currently I am doing enhanced response in CW mode and using that calset file to do measurement in Pulsed mode....is this method right??

We normally tell people who are doing wideband pulse measurements to calibrate in CW mode, mainly because it makes it easier to do the power calibration.  if you do the power calibration in pulse you either have to have a peak power meter or setup an offset in the power meter loss table for the duty cycle of the pulsed signal measured by an average power sensor.  The s-parameter calibration can be done in either pulsed or CW and in your case, since you are only doing an ERC (no power calibration), it makes no difference to calibrate in pulsed or CW.

还有几个问题;
1）我认为衰减器在端口2上，你正在进行S21测量，对吗？
正如daras所说，最好将50 dB焊盘放在最接近DUT的位置。
2）你说15 dB的路径损耗，是在端口1或端口2上或在它们之间划分3）你使用什么IF BW。
你有65 dB的损耗，你说45 dB的S / N;
所以我猜你没有使用窄的IF BW或更多的平均值。
4）你正在做高增益，这意味着你在测试端口设置的功率非常低。
你有什么力量？
你的衰减器级别是多少？
您可以考虑反转端口1耦合器，以在较低功率级别设置下改善S11。
在我的书第6章第6.3.1.1节中有一节。
请放心，我们有许多客户在放大器和转换器上进行高增益（> 100 dB）低输入功率测量，获得非常好的结果，您只需要进行适当的设置。
5）你的DUT增益是什么，S11匹配？
和输出功率？



     以下为原文

  A couple more questions;

1) I presume the attenuator is on port 2, and you are doing S21 measurements, correct?

As daras said, best to put the 50 dB pad closest to the DUT.

2) you say 15 dB of path loss, is that on port 1 or port 2 or divided between them

3) What IF BW are you using.  you have 65 dB of loss, you say 45 dB of S/N; so I guess you are not using a narrow IF BW or more averaging.

4) you are doing high gain, which likley means you are setting the power very low at the test port. What power are you setting? what is your attenuator level?   you might consider reversing the port 1 coupler to improve S11 at lower power level settings.  There is a section on this in my book, chapter 6, section 6.3.1.1.


Be assured, we have lots of customers doing high gain (>100 dB) low input power measurments on amplifiers and converters getting very good results, you just have to get the proper settings. 

5) what is your DUT gain, S11 match? and output power?

还有几个问题;
1）我认为衰减器在端口2上，你正在进行S21测量，对吗？
答：是的。
我们正在使用3号端口和
4因此，衰减器在端口4上并进行S43测量。
正如daras所说，最好将50 dB焊盘放在最接近DUT的位置。
2）你说15 dB的路径损耗，是在端口1或端口2上或在它们之间划分Ans：15 dB是总路径损耗（前向加返回）3）你使用什么IF BW。
你有65 dB的损耗，你说45 dB的S / N;
所以我猜你没有使用窄的IF BW或更多的平均值。
Ans：1 Khz 4）你正在做高增益，这意味着你在测试端口设置的功率非常低。
你有什么力量？
你的衰减器级别是多少？
您可以考虑反转端口1耦合器，以在较低功率级别设置下改善S11。
在我的书第6章第6.3.1.1节中有一节。
Ans：我正在测量@ 0 dBm。
请放心，我们有许多客户在放大器和转换器上进行高增益（> 100 dB）低输入功率测量，获得了非常好的结果，您只需要进行适当的设置即可。
5）你的DUT增益是什么，S11匹配？
和输出功率？
Ans：DUT增益：Tx模式下的52 dB（+/- 1 dB）回波损耗为15 dB



     以下为原文

  A couple more questions;

1) I presume the attenuator is on port 2, and you are doing S21 measurements, correct?
Ans:  Yes. We are using port 3 & 4 so, attenuator is on port 4 and doing S43 measurement.

As daras said, best to put the 50 dB pad closest to the DUT.

2) you say 15 dB of path loss, is that on port 1 or port 2 or divided between them

Ans: 15 dB is total path loss (Forward plus return)

3) What IF BW are you using. you have 65 dB of loss, you say 45 dB of S/N; so I guess you are not using a narrow IF BW or more averaging.

Ans: 1 Khz

4) you are doing high gain, which likley means you are setting the power very low at the test port. What power are you setting? what is your attenuator level? you might consider reversing the port 1 coupler to improve S11 at lower power level settings. There is a section on this in my book, chapter 6, section 6.3.1.1.

Ans: I am doing measurement @ 0 dBm.

Be assured, we have lots of customers doing high gain (>100 dB) low input power measurements on amplifiers and converters getting very good results, you just have to get the proper settings. 

5) what is your DUT gain, S11 match? and output power?

Ans: DUT Gain : 52 dB (+/- 1 dB) in Tx mode
        Return Loss of 15 dB