怎么使用N5242A进行绝对功率测量

嗯，最明显的答案是它可能是由谐波和/或次谐波引起的。
功率传感器和VNA完全不同地测量功率。
功率计基本上可以看到应用于它的每个信号，而VNA只能看到它被调谐到的信号。
如果您发布频谱输出的宽带图，我们可以更好地了解发生了什么。
选择一些差异最大的CW频率。
然后得到从该CW频率的约1/4到该频率的约5或6倍的信号的频谱图。



     以下为原文

  Well, the most obvious answer is that it could be due to harmonics and/or sub-harmonics.  A power sensor and a VNA measure power completely differently.  A power meter sees basically every signal applied to it while a VNA only sees the signal it is tuned to. If you post a wide band plot of the spectral output, we would have a better idea of what is going on.  Pick some CW frequency where the discrepancy is worst.  Then get a spectral plot of the signal from about 1/4 of that CW frequency to about 5 or 6 times that frequency.

谢谢jvall的回复。
我已经在进行功率测量的端口处安装了低通滤波器。
目前我使用的低通滤波器具有高于15Ghz的高抑制（CW为10.7Ghz），所以我想我可以过滤任何谐波。
而且，在整个感兴趣的频带中，测量的差异几乎是一致的。
我不知道功率传感器频率范围内的噪声基底是否可以加到0.2dB ...... @。@



     以下为原文

  Thank you jvall for your response.

I already have lowpass filter at the port where power measurement is done.
Currently i'm using lowpass filter with high suppression above 15Ghz (CW at 10.7Ghz) so I think I'm ok filtering any harmonics.

Also, difference in the measurement is almost consistent throughout the frequency band of interest. 

I don't know if noise floor across the frequency range of power sensor can be added up to 0.2dB... @.@

你在做一个单独的1）源cal，然后2）接收器cal？
或者你正在做一个引导功率校准（使用复选框在S参数校准期间选择“校准源和接收器”。如果你正在做一个源和接收器校准，独立于S参数校准，那么接收器校准是
不匹配校正，并且VNA接收器（和信号源）与功率传感器不匹配会产生一些误差。使用引导功率校准将使其比功率传感器更精确。而且，并非一无所获，
如果您将功率从参考功率（通常为0 dBm）更改为其他功率，功率传感器并不完全准确。功率测量中的不匹配效应的典型误差为0.3-0.7 dB，如果您不执行引导功率
cal。最后，你如何比较功率计measuremetns和PNA接收器测量？你怎么知道每个功率源的功率是相同的？事实上，为了一个好的校准，差异应该在0.02的数量级
dB（正确地说明功率计不匹配可能导致插入的事实
也可以阅读！）您可以发布用于校准和测量的设置图，显示如何连接功率计，LPF和VNA接收器以进行比较。



     以下为原文

  Are you doing a separate 1) source cal, and then 2) receiver cal? or are you doing a guided power cal (using the checkbox to select "cal source and receiver" during the S-parameter cal.

If you are doing a source and receiver cal, independ of the S-parameter cal, then the receiver cal is not match corrected, and the VNA receiver (and source) being not as well matched as a power sensor, will give some error.

Using the guided power cal will allow it to be more accurate than a power sensor.

And, not for nothing, the power sensor isn't perfectly accurate if you change the power from the reference power (usually 0 dBm) to some other power.

Typical error for mismatch effects in power measurements is 0.3-0.7 dB, if you don't do a guided power cal.

And, finally, how are you comparing the power meter measuremetns to the PNA receiver measuerments?  How do you know the source power for each is the same?

In fact, for a good calibraition, the difference should be in the order of 0.02 dB (properly accounting for the fact that power meter mismatch may cause an incorrect reading too!)
 
you might post a drawing of the setup for calibration and measurement, showing how the power meter, LPF and VNA receiver are connected for the comparison.

仅供参考PNA固件A.09.30中引导功率校准功能首次亮相，因此如果您的固件早于此版本，那么您将无法在S参数校准向导中看到Dr_Joel所引用的“校准源和接收器功率”复选框。



     以下为原文

  FYI the Guided Power Cal feature first debuted in PNA Firmware A.09.30, so if your firmware is older than that then you won't see in the S-parameter cal wizard the "Calibrate source and receiver power" checkbox to which Dr_Joel referred.

功率传感器 - 到 - 功率传感器的可重复性大约为0.15-0.2 dB，事实上，虽然规格更差。
如果使用引导功率校准进行校准，使用与PNA校准相同的功率计进行测量，则应得到几乎相同的答案（在这种情况下，由于不匹配，功率计误差可能高达0.03）。
但如果您使用一个功率计进行校准并使用另一个功率计进行测量，则误差可能超过.2 dB。
在您的测量情况下，我假设您正在设置PNA以生成CW频率，并使用USB传感器进行测量（如何使用Power-meter-as-receiver功能，PMAR？）以及测量外部功率传感器读数
功率计显示？
在这种情况下，USB功率计读数是否与外部功率计匹配相同的CW频率和功率水平？
如果是这样，如果PNA不匹配，我们会遇到一个问题;
如果没有，那么你有一个功率计问题。
对于ext功率计，您是否也设置了校准因子。
你当然是零并引用它吗？
（参考漂移可在几分钟内达到0.05 dB）。



     以下为原文

  Power sensor -to- Power sensor repeatability is on the order of 0.15-0.2 dB, in fact, though the spec is even worse.
If you calibrate using guided power cal, using the same power meter to measure as you do for PNA calibration, then you should get almost identical answers (the power meter error can be as much as 0.03 in such a case due to mismatc). 

But if you calibration with one power meter and measure with another, the error can be over .2 dB.

In your measuerment case, I presume you are setting the PNA to generate a CW freq, and measauring with the USB sensor (how? Using the Power-meter-as-receiver function, PMAR?) and also measuring on the external power sensor reading the power meter display?  In such a case, does the USB power meter reading match the external power meter for the same CW freq and power level?

IF so, we have one issue if the PNA doesn't match; if no, then you have a power meter issue.

For the ext power meter, are you also setting the cal factor.  And of course do you zero and reference it?  (reference drift can be 0.05 dB over a few minutes).

Joel博士，USB功率计和接收器数据匹配在0.02dB以内。
外部功率计/传感器和VNA读数之间存在差异。
在进行测量之前我做零和校准，但我没有使用传感器的校准因子。
也许我应该更多地了解功率传感器和校正因子的使用。
我担心我可能会打开一罐蠕虫............它几乎看起来像是朝向功率计/传感器的不稳定性。
Arggg ...........但是有一个问题..正如我在上一篇文章中所提到的，我的VNA源大约是-40dBm，从DUT到VNA的功率大约是10dBm。
是否可以将校准功率提高到0dBm？
或者我应该使用不同的功率水平？
谢谢。



     以下为原文

  Dr. Joel,

USB power meter and and receiver data match within less than 0.02dB. The difference comes between external power meter/sensor and VNA reading.
I do zero and cal before making measurement but i'm not using cal factor from the sensor.
Maybe i should look up more on power sensor and use of correction factor.

I'm afraid that i might be opening a can of worm........... it almost look like i'm heading towards instability of power meter/sensor. Arggg...........

One question though..

as i mentioned on previous post, my source from VNA is about -40dBm and power into VNA from DUT is about 10dBm.
Is it ok to power cal at 0dBm? or should i use different power level?

Thank you.

你当然可以在0 dBm下校准并测量10 dBm。
PNA-X的接收器线性非常好，我们的规格仅受测量技术的限制，而不受接收器的实际性能的限制。
但是，在+10 dBm处，您处于压缩接收器的边界线上，因此为了安全起见，在开始校准之前，请务必在通道中使用至少5 dB的接收器衰减。



     以下为原文

  You can certainly cal at 0 dBm and measure at 10 dBm.  The PNA-X's receiver linearity is so good that our specs are only limited by the measurement technique and not the actual performance of the receiver.  However, at +10 dBm you are on the borderline of compressing the receiver, so just to be on the safe side, be sure to use at least 5 dB of receiver attenuation in your channel before you start the calibration.

> {quote：title = daras写道：} {quote}>你当然可以校准0 dBm并测量10 dBm。
PNA-X的接收器线性非常好，我们的规格仅受测量技术的限制，而不受接收器的实际性能的限制。
但是，在+10 dBm处，您处于压缩接收器的边界线上，因此为了安全起见，在开始校准之前，请务必在通道中使用至少5 dB的接收器衰减。
Dr_joel提到我应该做出一个澄清。
在-40 dBm输入时，您必须使用至少10 dB的源衰减，这意味着您可能无法使用正常的导向功率校准向导在0 dBm下进行功率计校准。
因此Dr_joel再次提到，Cal All是要走的路。
我一直在使用它，以至于我甚至不再考虑它，这就是为什么我只是假设你可以在0 dBm校准:-)



     以下为原文

  > {quote:title=daras wrote:}{quote}
> You can certainly cal at 0 dBm and measure at 10 dBm.  The PNA-X's receiver linearity is so good that our specs are only limited by the measurement technique and not the actual performance of the receiver.  However, at +10 dBm you are on the borderline of compressing the receiver, so just to be on the safe side, be sure to use at least 5 dB of receiver attenuation in your channel before you start the calibration.

I should have made one clarification as Dr_joel mentioned.  at -40 dBm input, you will have to use at least 10 dB source attenuation and that means you will probably not be able to do the power meter calibration at 0 dBm, using the normal guided power calibration wizard that is.  So again as Dr_joel mentioned, Cal All is the way to go.  I've been using it so much that I don't even think about it anymore, which is why I just presumed you could calibrate at 0 dBm :-)

Joel博士和daras，感谢您的支持！
我在功率测量方面取得了很好的进展。
现在的差异实际上小于0.05dB !!!!
我做了几个不同的测量，差异保持在0.05dB以下。
在不同的衰减器设置下使用引导式校准可以大大改善测量。
我知道这是网络分析仪讨论板，但我有关于功率传感器校正因子的最后一个问题。
我是否在零/卡之前应用校正因子？
或零/卡后？
还是重要的？
我正在评估PNA-X来取代我们目前的测试装置（它有超过10台设备），到目前为止绝对功率测量是最大的挂机。
我小组中的人对进度感到兴奋，并且所有积分都会回复给您。
：） 谢谢！！



     以下为原文

  Dr. Joel and daras,

Thank you for all your support!!! I was able to make good progress on power measurement. Now the difference is actually less than 0.05dB!!!! I made several different measurement and the difference remained under 0.05dB.

Using guided cal at different attenuator setting improved measurement a lot. 

I know this is network analyzer discussion board but i have one last question about correction factor on power sensor.

Do I apply correction factor before zero/cal? or after zero/cal? or does it matter?

I'm evaluating PNA-X to replace our current test set (which has more than 10 equipments) and so far absolute power measurement was the biggest hang up.

People in my group are excited about the progress and all credits goes back to you. :)

Thank you!!

> {quote：title = need2learn写道：} {quote}> Joel博士和daras，>>感谢您的支持！
我在功率测量方面取得了很好的进展。
现在的差异实际上小于0.05dB !!!!
我做了几个不同的测量，差异保持在0.05dB以下。
>您可能已接近该级别的残余误差限制。
>在不同的衰减器设置下使用引导式校准可以大大改善测量。
>>我知道这是网络分析仪讨论板，但我有关于功率传感器校正因子的最后一个问题。
>>我在零/卡之前应用校正因子吗？
或零/卡后？
还是重要的？
>您应用两种不同的校正因子。
对于零/校准，校准是使用功率计的50 MHz输出完成的，因此通常会打印一个“REF”校正因子（应该与50 MHz频率值相同），然后执行“校准”
零和校准的一部分，然后将校准因子更改为为要测量的频率指定的值。
它有点棘手，对于+ - .25 dB，它并不重要，但是小于0.05 dB ......你必须要小心一点，在100瓦输出时相差约1.2瓦
！
>我正在评估PNA-X取代我们目前的测试装置（有超过10台设备），到目前为止绝对功率测量是最大的挂断。
>>我小组中的人对进度感到兴奋，并且所有积分都会回复给您。
:)>嗯，你做了证明它的工作。
如果您需要任何其他帮助，请告诉我们（如果您愿意，可随时发送PM或电子邮件，其中包含您公司的详细信息和产品测试详情）



     以下为原文

  > {quote:title=need2learn wrote:}{quote}
> Dr. Joel and daras,
> 
> Thank you for all your support!!! I was able to make good progress on power measurement. Now the difference is actually less than 0.05dB!!!! I made several different measurement and the difference remained under 0.05dB.
> 
You're probably getting close to the limits of the residual errors, at that level.

> Using guided cal at different attenuator setting improved measurement a lot. 
> 
> I know this is network analyzer discussion board but i have one last question about correction factor on power sensor.
> 
> Do I apply correction factor before zero/cal? or after zero/cal? or does it matter?
> 
You apply two different correction factors. For zero/cal, the cal is done using the 50 MHz output of the power meter, so there is usually printed a "REF" correction factor (should be the same value as the 50 MHz freq value), then do the "cal" portion of the zero and cal, then change the cal factor to the value that is specified for the frequency to be measured.
It's all a little tricky, and for +-.25 dB, it doesn't matter much, but for less than 0.05 dB...you have to be a little more careful, and amounts to about 1.2 watts difference at 100 watts output!.

> I'm evaluating PNA-X to replace our current test set (which has more than 10 equipments) and so far absolute power measurement was the biggest hang up.
> 
> People in my group are excited about the progress and all credits goes back to you. :)
> 
well, you did the work to prove it.  Let us know if you need any other help (feel free to send a PM or email with your company details and product test details if you like)

> {quote：title = need2learn写道：} {quote}> Joel博士，>> USB功率计和接收器数据匹配在0.02dB以内。
外部功率计/传感器和VNA读数之间存在差异。
据此，我推断USB功率计和外部功率计/传感器不一致，对吗？
>在进行测量之前我做零和校准，但我没有使用传感器的校准因子。
>也许我应该更多地了解功率传感器和使用校正因子。
我有8485D;
在我的传感器上，10 GHz的校准因子为94.7％;
如果不考虑，这将导致0.236 dB的误差。
校准因子的不确定性为2％，因此表示0.09 dB的不确定性（忽略其他不确定性）。
因此，忽略校准因子会给您带来很大的误差，但即使使用它，两个传感器之间的相对功率精度也会达到0.125 dB（每个传感器的不确定度为0.09 dB）。
也就是说，即使采用最佳测量技术（正确地进行测量），您也不能指望两个传感器比0.125 dB更接近。
具有内置因素的u***功率传感器通过要求您输入被测信号的频率来使用它们;
当与PNA一起使用时，会自动发生这种情况。
>正如我在上一篇文章中提到的，我的VNA源大约为-40dBm，DUT的VNA功率大约为10dBm。
> 0校准电源是否可以接受？
或者我应该使用不同的功率水平？
>是的，因为更改源的功率电平不会影响B（端口2）接收器的功率测量精度。
但是，由于功率变化为-40 dBm，因此必须改变衰减器状态。
如果是这样，由于衰减器偏移不确定性，源功率（以及R1或端口1参考功率的校准）的误差将在0.25-0.5dB的量级。
在这种情况下，您有两种改进测量的选择：1）将port1源衰减器设置为10 dB，这允许校准-5 dBm（合理功率）以及将功率设置为-40 dBm（功率范围在
PNA约为衰减器设置的+10到-30 dBc）。
2）使用cal all，即使测量通道的所需功率范围具有不同的衰减器设置，它（即使在单通道测量中）也将在一个衰减器设置中允许功率校准。
它通过在校准时添加衰减器偏移的附加测量来实现此目的。
如果您有其他问题，请回复。
我可以肯定地说，很多大型着名的放大器测试公司都使用PNA-X引导功率校准进行生产放大器测试。
通过适当的设置，我很有信心它可以产生最精确的功率测量，甚至比功率计更精确（除非DUT正好是50欧姆，输出功率正好是0 dBm）。



     以下为原文

  > {quote:title=need2learn wrote:}{quote}
> Dr. Joel,
> 
> USB power meter and and receiver data match within less than 0.02dB. The difference comes between external power meter/sensor and VNA reading.

From this, I infer that the USB power meter and the external power meter/sensor do NOT agree, correct?

> I do zero and cal before making measurement but i'm not using cal factor from the sensor.
> Maybe i should look up more on power sensor and use of correction factor.

I have an 8485D; on my sensor the cal factor at 10 GHz is 94.7%; this would result in a 0.236 dB error if not accounted for.   The uncertainty of the cal factor is 2%, so that represents an uncertainty of  0.09 dB (ignoring other uncertainties).  Thus, ignoring the cal factor gives you a quite big error, but even using it, your relative power accuracy between two sensors is something like 0.125 dB (each sensor having an uncertainty of 0.09 dB).  That is, even with the best measurement techniques (doing everthing right) you cannot expect two sensors to agree more closely than 0.125 dB.  

The u*** power sensor, which has built-in factors, uses them by requiring that you enter the freqeuncy of the signal being measured; when used with the PNA, this happens automatically.

> as i mentioned on previous post, my source from VNA is about -40dBm and power into VNA from DUT is about 10dBm.
> Is it ok to power cal at 0dBm? or should i use different power level?
> 
Yes, as changing the power level at the source won't affect the power measurement accuracy of the B (port 2) receiver.  However, since the power change is to -40 dBm, it is liklely that you must change attenuator states.  If so, the error in source power (and in calibration of the R1 or port1 reference power) will be on the order of 0.25-0.5 dB, due to attenuator offset uncertainty.  In such a case, you have two choices for improved measurements:

1) set the port1 source attenuator to 10 dB, which allows calibration of -5 dBm (a reasonable power) as well as setting power to -40 dBm (power range on the PNA is about +10 to -30 dBc of the attenuator setting).
2) use cal all, which (even on a single channel measurement) will allow a power cal in one attenuator setting even if the desired power range for the measurement channel has a different attenuator setting. It does this by adding an additional measurment of the the attenuator offset at the time of calibration.

post back if you have other questions.  I can say, with good conviction, that many of the large-scale well-known amplifier test companies use the PNA-X guided power calibration for their production amplifier test.  And with proper setup, I have high confidence that it produces the most accuract power measurements, even more accurate than a power meter (unless the DUT is exactly 50 ohms, and the output power is exactly 0 dBm).