进行校准与高值衰减器连接的PNA时需要特殊设置吗？

这里有几个问题。
首先，Cal All对校准中选择的所有端口执行源功率校准，默认情况下，您将获得2端口校准。
我将在这里继续讨论并假设您没有将通道中的端口功率分离，并将端口2功率设置得足够低以容纳40 dB衰减器。
换句话说，你试图在端口2上实现-30到0 dBm，损耗为40 dB，这意味着端口2上的源（在Cal All执行源功率校准后）必须输出+10到+40
在校准参考平面上获得-30到0 dBm的功率dBm，这就是为什么在衰减器就位的情况下光源不平整，并且在没有它或使用较小的衰减器进行校准时工作正常。
第二个问题是通过40 dB焊盘进行2端口s参数校正。
即使光源可以输出+ 40 dB，完全开路或短路的反射信号在到达端口2反射计时也会小于-80 dBm。
所以即使在最好的情况下你也会有一个非常嘈杂的S22，并且取决于你的DUT的反向隔离，也是一个非常嘈杂的S12。
通过完整的12项2端口纠错，这些噪声S22和S12参数会对您尝试进行的S21测量产生重大影响。
在这样的设置中，你最终测量的是40 dB衰减器的S22而不是DUT的S22。
解决方案是执行仅前向传输校准或我们称之为“增强响应”校准。
该方法校正输入匹配（S11）并用响应校正校正传输（S21）。
下图显示了在Cal All向导中必须遵循的过程，以便选择EnhResp方法。
请注意，此校准需要定义的Thru标准，如果您使用的是ECal，则ECal thru按定义使用。
如果您不使用ECal，则需要零或冲洗校准设置。
！HTTPS：//dl.dropbox.com/u/87949221/CalAll_EnhancedResp.jpg！



     以下为原文

  there are a couple of issues here. first, Cal All performs a source power calibration for all the ports selected in the calibration and by default you are getting a 2-port cal.  I am going to go on limb here and assume you did not uncouple the port powers in your channels and set the port 2 power to be low enough to accommodate the 40 dB attenuator.  In other words you are trying to achieve -30 to 0 dBm on port 2 with 40 dB of loss which means that the source on port 2 (after the source power calibration performed in Cal All) will have to put out +10 to +40 dBm of power to get -30 to 0 dBm at the calibration reference plane and that is why the source is going unleveled with the attenuator in place and works fine when you calibrate without it or with a smaller attenuator.

the 2nd problem is with the 2-port s-parameter correction through a 40 dB pad.  even if the source could put out +40 dB, the reflected signal of a perfect open or a short would be less than -80 dBm by the time it reached the port 2 reflectometer.  So even in the best case scenario you would have a very noisy S22 and depending on the reverse isolation of your DUT, also a very noisy S12.  with a full 12 term 2-port error correction, these noisy S22 and S12 parameters would have a significant impact on the S21 measurement you are trying to make.  In a setup like this what you end up measuring in essence is the S22 of the 40 dB pad and not the S22 of the DUT.  The solution is to perform a forward transmission only calibration or what we call the "Enhanced Response" calibration.  this method corrects the input match (S11) and the corrects the transmission (S21) with a response correction.  the following diagram shows the process you have to follow in the Cal All wizard in order to select the EnhResp method.  Note that this calibration requires a defined Thru standard and if you are using an ECal, then the ECal thru is used as defined thru.  if you are not using an ECal, then you need a zero or flush-thru calibration setup.

!https://dl.dropbox.com/u/87949221/CalAll_EnhancedResp.jpg!

Dara解决了这个问题，但要简洁：1）如果你将衰减（或增益）加到超过10 dB的端口，你应该使用功率/衰减器下的“功率偏移”：偏移和限制选择。
对于你的情况，使端口2的偏移= -40 dB。
2）你必须解开港口的权力。
使端口2功率=（端口1功率+近似增益 - 10 dB）并且，您无法真正通过40 dB衰减器进行测试，因此您不应使用完整的2端口校准。
如果要测量所有4个s参数，则必须使用前面提到的“微波元件测量手册”第6章中所述的高功率设置。
但是：如果您的DUT是高增益而非高功率，则无需使用40 dB衰减器。
增益有多高，输出功率有多高？



     以下为原文

  Dara covered the issue, but to be succinct:
1) if you add attenuation (or gain) to a port in excess of 10 dB, you should use the "power offsets" in under the power/attenuators:offsets and limits selection.  for you case, make the offset for port 2=-40 dB.

2) You must uncouple the port powers.  Make port 2 power = (port 1 power + approximate gain - 10 dB)

And, you cannot really cailbrate through a 40 dB attenuator, so you should not use a full 2 port cal.  If you want to measure all 4 s-parameters, you must use a high power setup as described in chapter 6 of the previously referred book, Handbook of Microwave Component Measurements.

BUT: if you DUT is high gain and NOT high power, no need to use the 40 dB attenuator.  How high is the gain, how high is the output power?

感谢Daras，drkirkby，bhokkan和Joel博士的回复;
对此，我真的非常感激。
回复drkirkby：我只有一个基本的2端口PNA N5222A（没有暴露的其他端口，如R，B等），并且拥有最新的PNA固件A.09.85。
我已经阅读了应用笔记“Agilent AN 1287-6使用网络分析仪来表征高功率组件”，我正在按照应用笔记使用配置1。
！HTTPS：//lh6.googleusercontent.com/-1hpQaU8U6zQ/UU6HKaAi3-I/AAAAAAAAAAk/xcIYRFPSyO0/s412/Config01.png！
回答Daras和bhokkan：是的，我没有解开港口。
我需要测量的只是输入VSWR（S11），增益平坦度（S21），增益线性度（不同频率下的功率扫描以找到压缩点）和一些脉冲测量（相位误差测量）。
由于我没有测量S22或S12，这是否意味着前向传输校准（增强响应校准）就足够了？
顺便说一下，我还在使用机械校准套件（虽然我之前已经购买了ECal）。
我将尝试增强响应校准。
回答Joel博士：如果我只需要测量输入VSWR，增益平坦度，增益线性度和一些脉冲测量值，那么前向传输是否足够？
我的放大器DUT是一个50 dB增益放大器，我希望它能从-30 dBm驱动到最大0 dBm。
这意味着我的最大输出功率约为50 dBm（60 W）。
再次提前感谢。
编辑：Engineer007于2013年3月23日11:20 PM编辑：Engineer007于2013年3月23日下午11:45



     以下为原文

  Thanks Daras, drkirkby, bhokkan and Dr Joel for the replies; I really appreciate it.

In reply to drkirkby:
I only have a basic 2-ports PNA N5222A (without other ports like R, B, etc exposed) and am having the latest PNA firmware A.09.85. I've read the app note "Agilent AN 1287-6 Using a Network Analyzer to Characterize High-Power Components" and I'm using Configuration 1 as per the app note.

!https://lh6.googleusercontent.com/-1hpQaU8U6zQ/UU6HKaAi3-I/AAAAAAAAAAk/xcIYRFPSyO0/s412/Config01.png!

In reply to Daras and bhokkan:
Yes, I did not uncouple the ports. What I need to measure is just the input VSWR (S11), gain flatness (S21) , gain linearity (Power sweep at different frequencies to find the compression point) and some pulse measurements (phase error measurements). Since I did not have to measure S22 or S12, does it mean that forward transmission cal (Enhanced Response cal) is sufficient here? By the way, I'm still using a mechanical cal kit (although I have put a purchase for an ECal sometime ago). I will try out the Enhanced Response cal.

In reply to Dr Joel:
If I only need to measure input VSWR, gain flatness, gain linearity and some pulse measurements, would forward transmission cal be sufficed here? My amplifier DUT here is a 50 dB gain amplifier and I expect to drive it from -30 dBm to 0 dBm max. This means I would have a max output power of about 50 dBm (60 W).

Thanks again in advance.

Edited by: Engineer007 on Mar 23, 2013 11:20 PM

Edited by: Engineer007 on Mar 23, 2013 11:45 PM

> {quote：title = Engineer007写道：} {quote}>回复Joel博士：>如果我只需要测量输入VSWR，增益平坦度，增益线性度和一些脉冲测量值，那么前向传输速率是否足够？
我的放大器DUT是一个50 dB增益放大器，我希望它能从-30 dBm驱动到最大0 dBm。
这意味着我的最大输出功率约为50 dBm（60 W）。
你的计算有问题，因为50 dB增益放大器的0 dBm会给你50 dBm，即100 W，而不是60 W.我想到的是衰减器特性随温度变化的可能性。
如果您在室温下测量其注意力，然后根据该值进行计算，如果高功率水平下的耗散会显着改变温度，则可能会引入误差。
几年前，当我大约15或16岁时，我建造了一个1千瓦的放大器，用于144 MHz的业余无线电使用。
这是基于一对Eimac 4CX250B阀门。
我从来没有一个额定功率为1 kW的推进器假负载，所以对于测试我有一个RG58同轴电缆的卷轴，这是足够的损耗，反射功率是微不足道的。
这是因为我施加的电源时间太长，在这种情况下，同轴电缆会熔化。
这种情况发生了很多次，但我只是习惯于寻找熔化的地方，切断坏的位置，并在同轴电缆上贴上新的N插头！
那些日子很有趣！
戴夫。



     以下为原文

  > {quote:title=Engineer007 wrote:}{quote}
> In reply to Dr Joel:
> If I only need to measure input VSWR, gain flatness, gain linearity and some pulse measurements, would forward transmission cal be sufficed here? My amplifier DUT here is a 50 dB gain amplifier and I expect to drive it from -30 dBm to 0 dBm max. This means I would have a max output power of about 50 dBm (60 W).

There's something wrong with your calculations, as  0 dBm into a 50 dB gain amplifier would give you 50 dBm, which is 100 W, not 60 W. 

Something that crossed my mind is the possibility of the characteristics of the attenuator changing with temperature. If you measure its attenation at room temperature, then make calculations based on that value, you could introduce an error if dissipation at high power levels changes the temperature significantly. 

Years ago, when I was about 15 or 16, I built a 1 kW amplifier for amateur radio use at 144 MHz. That was based on a pair of Eimac 4CX250B valves. I never had a propor dummy load rated at 1 kW, so for testing I had a reel of RG58 coax, which was sufficiently lossy that the reflected power was insignificant. That was unless I kept the power applied for too long, in which case the coax would melt. That happened many times, but I just used to look for where it had melted, cut the bad bit off, and stick a new N plug on the coax! Those days were fun!

Dave.

抱歉，是我的错。
你是对的，它是100W，而不是60W。



     以下为原文

  Sorry, my mistake. You're right, it's 100W, not 60W.

> {quote：title = Dr_joel写道：} {quote}>低端噪声是由耦合器滚降引起的（开始时约为200 MHz，在10 MHz时下降到35 dB超额损耗）。
>>是逐个扫描的跟踪噪声，还是“固定”的尖峰（不会改变扫描到扫描）。
>>如果它是固定的，它会“嵌入”校准中，为了解决这个问题，你必须在校准期间降低BW。
>>如果它改变扫描到扫描，它只是通道中的噪声（这里的源功率是多少？如果它是-30 dBm，那么它意味着B接收器的功率为-70 dBm; 1 kHz
IF BW的本底噪声约为-100 dBm，因此信噪比仅为30 dB，由于滚降而在低频处变得更差。>>您可以使用平均值来消除此频率而不改变IF BW或降低IF
BW。如果您的通道当前为-30 dBm，请提高吸收功率并查看噪声是否降低。>>我认为40 dB焊盘仅在端口2上，对吗？您还可以在0 dBm（或更高）进行校准
），然后降低cal后的功率，这将减少误差项中的噪声.Spacecase



     以下为原文

  > {quote:title=Dr_joel wrote:}{quote}
> The low end noise is due to the coupler roll-off (starts about 200 MHz and rolls down to be 35 dB excess loss at 10 MHz).
> 
> Is the trace noise on a sweep-by-sweep basis, or is the spikyness "fixed" (doesn't change sweep to sweep). 
> 
> If it is fixed, it is "embedded" in the calibration, and to cure this problem you will have to lower the BW during cal all.
> 
> If it changes sweep-to-sweep, it is just noise in the channel (what is the source power here? If it is -30 dBm, then it means the power at the B receiver is -70 dBm; at 1 kHz IF BW the noise floor is about -100 dBm, so only 30 dB signal to noise, getting worse at the low frequencies due to the roll-off.
> 
> You can use averaging to remove this without changing IF BW, or low the IF BW.  If your channel is currently at -30 dBm, raise the soruce power and see if the noise is reduced.
> 
> I presume the 40 dB pad is only on port 2, right?

You could also calibrate at 0 dBm (or higher), then lower the power after cal, this will reduce the noise in the error terms.

Spacecase

感谢SpaceCase，将尝试更高的功率水平。



     以下为原文

  Thanks SpaceCase, will try at a higher power level.

> {quote：title = Dr_joel写道：} {quote} >>我假设40 dB打击垫只在端口2上，对吧？
Joel博士，Yup，40 dB打击垫仅在端口2上。顺便说一句，我没有回复一周，因为我没有PNA，因为它只是借给我。
现在我得到了一个新的PNA-X，我今天开始使用它。
但是，当我尝试使用增强型响应Cal进行Cal All时，我今天还有另一个*大问题*。
当我在校准结束时尝试保存校准集时，整个PNA应用程序崩溃了。
崩溃是非常一致的，它只发生在我修改calkit以执行增强响应校准时（如果它是一个完整的两端口校准，则没有问题）。
我之前在使用修改过的calkit的PNA上做过这个，之前从未见过这个问题。
我可能会忽略这些潜在的东西吗？
我正在使用最新的PNA软件版本A.09.85.08。



     以下为原文

  > {quote:title=Dr_joel wrote:}{quote}
> 
> I presume the 40 dB pad is only on port 2, right?

Dr Joel,

Yup, the 40 dB pad is only on port 2.

By the way, I didn't reply for a week plus because I don't have a PNA with me since it was only loaned to me. Now that I got a new PNA-X, I started to use it today.

However, I have another *big issue* today when I'm trying to do a Cal All using Enhanced Response Cal. When I tried to save the calset at the end of the calibration, the whole PNA application crashed. The crash was pretty consistent and it only happened when I modify the calkit to do Enhanced Response Cal (there is no issue if it's a full two port cal). I've done this before on a PNA using modified calkit and have never seen this issue before. Is there any potential stuff that I may overlook here? I'm using the latest PNA software version A.09.85.08.

> {quote：title = Engineer007写道：} {quote}>然而，当我尝试使用增强型响应Cal进行Cal All时，我今天还有另一个*大问题*。
当我在校准结束时尝试保存校准集时，整个PNA应用程序崩溃了。
崩溃是非常一致的，它只发生在我修改calkit以执行增强响应校准时（如果它是一个完整的两端口校准，则没有问题）。
我之前在使用修改过的calkit的PNA上做过这个，之前从未见过这个问题。
我可能会忽略这些潜在的东西吗？
我正在使用最新的PNA软件版本A.09.85.08。
您可以在启动Cal All向导之前向我们发送设置的状态文件吗？
另外，您正在使用的PNA-X的型号是什么？编辑：daras于2013年4月24日上午9:18



     以下为原文

  > {quote:title=Engineer007 wrote:}{quote}
> However, I have another *big issue* today when I'm trying to do a Cal All using Enhanced Response Cal. When I tried to save the calset at the end of the calibration, the whole PNA application crashed. The crash was pretty consistent and it only happened when I modify the calkit to do Enhanced Response Cal (there is no issue if it's a full two port cal). I've done this before on a PNA using modified calkit and have never seen this issue before. Is there any potential stuff that I may overlook here? I'm using the latest PNA software version A.09.85.08.

Can you send us a state file of your setup prior to starting the Cal All wizard? Also, what is the model number of the PNA-X you are using?

Edited by: daras on Apr 24, 2013 9:18 AM

我有好消息和坏消息。
坏消息是，遗憾的是，代码中存在问题并且我能够与您的州进行验证，A.09.85.08版本确实在校准过程结束时崩溃。
好消息是，该问题已在目前正在进行质量保证的新版本中得到解决。
我将向您发送指示，以下载新固件的alpha版本以进行试用。
官方发布将在2-3周内发布，所以希望在此之前您可以使用alpha版本。



     以下为原文

  I have good news and bad news.  The bad news is that unfortunately, there is a problem in the code and I was able to verify with your state, the A.09.85.08 version does indeed crash at the end of the calibration process.  The good news is that the problem has already been resolved in a new release that is currently undergoing QA.  I will PM you with instructions to download an alpha version of the new firmware to try out.  the official release will be in 2-3 weeks, so hopefully you can use the alpha version until then.

Dara，我已通过私信发送了有关alpha版本的反馈，请查看。
谢谢。



     以下为原文

  Dara, I've sent you a feedback on the alpha release via private message, please check it out. Thanks.

这可能值得一读。
+使用矢量网络分析仪（VNA）进行射频放大器测试的基础知识+ Taku Hirato http://www.home.agilent.com/upload/cmc_upload/All/Basics_of_RF_Amplifier_Test_with_VNA_pdf.pdf戴夫



     以下为原文

  This might be worth a read. 

+Basics of RF Amplifier Test with the Vector Network Analyzer (VNA)+ by Taku Hirato

http://www.home.agilent.com/upload/cmc_upload/All/Basics_of_RF_Amplifier_Test_with_VNA_pdf.pdf

Dave

> {quote：title = drkirkby写道：} {quote}>这可能值得一读。
>> +使用矢量网络分析仪（VNA）进行射频放大器测试的基础知识+ Taku Hirato >> http://www.home.agilent.com/upload/cmc_upload/All/Basics_of_RF_Amplifier_Test_with_VNA_pdf.pdf >> Dave是的，我们已经完成了
一些主动器件测量研讨会，这是一个强调使用ENA的研讨会（尽管所有基本思想同样适用于ENA和PNA）。
然而，制作良好放大器测量的最佳参考是Dr_Joel的_ * book * _的第6章。



     以下为原文

  > {quote:title=drkirkby wrote:}{quote}
> This might be worth a read. 
> 
> +Basics of RF Amplifier Test with the Vector Network Analyzer (VNA)+ by Taku Hirato
> 
> http://www.home.agilent.com/upload/cmc_upload/All/Basics_of_RF_Amplifier_Test_with_VNA_pdf.pdf
> 
> Dave

yes we have done a number of Active Device Measurement seminars and this is one with emphasis on using the ENA (although all the basic ideas apply equally to both ENAs and PNAs).  However, the best reference for making good amplifier measurements is in chapter 6 of Dr_Joel's _*book*_.

如果将去除器的去嵌入器大于10 dB，如果将衰减器的S22设置为0，则会获得更好的结果。（这是S11面向VNA的端口，s22是面向DUT的端口）。



     以下为原文

  If you de-embed attenutators greater than 10 dB, you get better results if you set S22 of the attenuator to 0. (that's where S11 is the port facing the VNA, and s22 is the port facing the DUT).

我尝试了PNA中的夹具去嵌入功能;
但这似乎只适用于进行频率扫描的测量通道。
我有其他通道进行功率扫描（增益与功率输入） - 它在这些通道中不起作用。
与一些进行脉冲测量的通道相同。
这是一个限制吗？



     以下为原文

  I tried the fixture de-embed feature in the PNA; but this only seem to work for measurement channels that do frequency sweep. I have other channels that do power sweep (gain versus power input) - it doesn't work in those channels. Same with some channels that do pulse measurement. Is this a limitation?

夹具去嵌入应适用于所有扫描类型。
对于固定CW频率的扫描类型，校正基本上是在所有功率电平下应用的单频校正。
当你说它不起作用时，你看到的是什么让你有理由相信它不起作用？
如果你能详细说明，我们可以提供更多帮助。谢谢，



     以下为原文

  Fixture de-embedding should work in all sweep types.  for sweep types at a fixed CW frequency, the correction is basically a single frequency correction applied at all the power levels.  When you say it doesn't work, what are you seeing that gives you reason to believe it isn't working?  We can be more help if you can elaborate a little bit.

Thanks,