在8720ES中没有L0，L1，L2作为短标准

> {quote：title = linan0827写道：} {quote}>嗨，全部，>>我们有一台8720ES网络分析仪。
我们想在校准分析仪中输入一些校准标准的自定义系数。
对于开放标准，菜单中有选项，您可以在分析仪中键入C0，C1，C2等。
然而，对于短标准，它在那里是空的，只能输入的是偏移延迟和偏移损失。
但是自定义校准套件的指令明确表明这些标准存在L0，L1，L2系数。
>>>我们想知道为什么8720ES不可能输入短标准系数，或者只输入偏移延迟和偏移丢失信息是否足够。
>>非常感谢您的解释。
问候，>> NL搞笑，我打算问几乎同样的问题，虽然我有8720D，而不是8720ES。
我有固件7.74。
（如果你陈述你的固件版本可能会有用。）几周前Joel写的一些东西，大约6 GHz以上，短路的电感不容忽视。
我正在输入6 GHz校准套件（85032B），因此对我来说不是问题，因为HP 85032B的手册列出了L0 = L1 = L2 = L3 = 0。
但我担心如果校准套件的工作频率更高，会发生什么。
我正在认真考虑购买新的Agilent 85032F 9 GHz校准套件。
查看PNA上85032F的定义，电感系数不为零。
http://na.tm.agilent.com/pna/caldefs/PNA/85032F.htm所以人们可能会期望，如果在8720D上使用它们，它们也应该是非零的。
有趣的是，85032F是我的8720D的固件。
我刚看了8720D的编程手册，我找不到任何日期，所以它可能比我的固件版本更旧。
查看该手册，没有用于输入L0，L1，L2或L3的GPIB命令。
因此，这些似乎不是你可以发送到GP-IB的东西，但不是在前面板上设置的。
我期待您对帖子的回复。
戴夫



     以下为原文

  > {quote:title=linan0827 wrote:}{quote}
> Hi, All,
> 
> We have a 8720ES network analyzer. We would like to input some custom coefficients of the calibration standards into our analyzer. For open standard there is the option in the menu so that you can type C0, C1, C2, etc into the analyzer. However for short standard it is empty there, only thing you can input is the offset delay and offset loss. But the instruction of the custom cal kits explicitly indicate there are those L0, L1, L2 coefficients for those standards.
> 
> We would like to know why it is not possible for 8720ES to have those coefficients of the short standard to be input, or is it sufficienct to only input offset delay and offset loss information.
> 
> Thank you very much for the explanation.
> 
> Regards,
> 
> NL

Funny, I was going to ask virtually the same question, although I have an 8720D, not an 8720ES. I have firmware 7.74. (It might be useful if you state your firmware version.)

From something Joel wrote a few weeks ago, above about 6 GHz, the inductance of the short can't be ignored. I was entering a 6 GHz cal kit (85032B), so it was not an issue to me, as the manual for the HP 85032B lists L0=L1=L2=L3=0. But I was concerned what would happen if the cal kit worked to a higher frequency. 

I'm giving serious consideration to buying a new Agilent 85032F 9 GHz cal kit. Looking at the definitions for a 85032F on a PNA, the inductance coefficients are non-zero. 

http://na.tm.agilent.com/pna/caldefs/PNA/85032F.htm

so one might expect that if used on an 8720D, they should be non-zero too. It's interesting that the 85032F is in fhe firmware for my 8720D. 

I just looked in the programming manual for the 8720D, on which I can't find any date, so it might be older than the firmware version I have. Looking at that manual, there are no GPIB commands for entereng L0, L1, L2 or L3. Hence these don't appear to be things you can send down the GP-IB, but not  set on the front panel. 

I look forward to what responses you get to your post. 

Dave

在区域8753代码中，短路电感的影响非常小，因此在延迟中被吸收。
该代码迁移到8720，当客户询问时，响应可能是“如果你担心L值很重要，你应该升级到8510.（第一个8720A的性能大大低于
跟踪噪音并没有很大的稳定性，但是成本却大大降低了。）随着8720的改进，代码没有变化，因此今天仍然如果校准套件具有L值，你不能忽略它们，但必须计算
与过量电感相关的延迟，并将其添加到套件的延迟，以获得8720的良好延迟值。但是！如果您担心质量测量，我可以建议您升级到新的PNA



     以下为原文

  in the orgional 8753 code, the effect of the inductance of the short was very small and so absorbed in the delay.  That code migrated to the 8720 and at the time if a customer asked about, the response might be "if you are that concerned that the L values matter, you should upgrade to the the 8510.  (the first 8720A had substantially inferior performance with high trace noise and not great stability, but was substantially lower cost).

As the 8720 improbed, no change was made to the code, and thus it remains today that if the cal kit has L values, you cannot ignore them, but must compute the delay that is associated with the excess inductance, and add it to the delay of the kit to get a good delay value for 8720. 

But! if you are that concerned about quality measurments, may I recommend that you upgrade to the new PNA

> {quote：title = Dr_joel写道：} {quote}>在orgional 8753代码中，短路电感的影响非常小，因此在延迟中吸收。
该代码迁移到8720，当客户询问时，响应可能是“如果你担心L值很重要，你应该升级到8510.（第一个8720A的性能大大低于
跟踪噪音并没有很大的稳定性，但是成本却大大降低了。>>> 8720改进后，代码没有变化，因此今天仍然如果校准套件有L值，你不能忽视它们，但必须
计算与过量电感相关的延迟，并将其添加到套件的延迟，以获得8720的良好延迟值。是否有一个公式可以为此提供，我们可以使用而不必回到第一原则
并试图去除它？我认为这个延迟将取决于频率，就像短路的电感一样，所以我们可能需要一个用户校准套件用于一个频率范围，另一个用于更高的频率范围。>但是！如果你
是关注质量测量的，我可以重新考虑一下
修改你升级到新的PNA我怀疑运行8753s，8720s等的人的问题是成本。
我从VNA中获得的金额为0.00美元。
我想建立一家公司，在那里我会从他们那里赚到钱，但是在这一刻，这对我来说都是支付，没有任何进展。（几个月前我被裁掉了）。
我根本无法证明PNA的成本。
我非常希望能够提供PNA。



     以下为原文

  > {quote:title=Dr_joel wrote:}{quote}
> in the orgional 8753 code, the effect of the inductance of the short was very small and so absorbed in the delay.  That code migrated to the 8720 and at the time if a customer asked about, the response might be "if you are that concerned that the L values matter, you should upgrade to the the 8510.  (the first 8720A had substantially inferior performance with high trace noise and not great stability, but was substantially lower cost).
> 
> As the 8720 improbed, no change was made to the code, and thus it remains today that if the cal kit has L values, you cannot ignore them, but must compute the delay that is associated with the excess inductance, and add it to the delay of the kit to get a good delay value for 8720. 

Is there a formula you can give us for this, which we can use without going back to first principles and trying to derrive it? 

I assume this delay is going to be frequency dependant, like the inductance of the short, so perhaps we need one user cal kit for one frequency range, and another for a higher freqenecy range.  

> But! if you are that concerned about quality measurments, may I recommend that you upgrade to the new PNA

I suspect the issue for people running 8753s, 8720s etc is cost. The amount of money I have earned from my VNAs is $0.00. I'm looking to set up a company, where I will make some money from them, but at the minute it is all payout for me, and nothing coming in. (I was made redundant a few months ago).  I simply can't justify the cost of a PNA. I'd *very* much like to be in the position of affording a PNA.

我只是为L与f创建多项式函数，然后计算电抗XL与f，然后计算值的相位，然后在感兴趣的频率图上，最小二乘法找到相位斜率并计算等效延迟。
留给读者的练习。编辑：Dr_joel于2012年11月15日下午6:07



     以下为原文

  I would simply create the polynomial functoion for L vs. f, then compute reactance XL vs f, then compute the phase of the value, then over the frequency of interest map a least squares find the phase slope and compuate the equivalent delay.

Left as an excercise to the reader.

Edited by: Dr_joel on Nov 15, 2012 6:07 PM

非常感谢您的回复。
这激发了我挖掘安捷伦的应用笔记，以找出用于模拟校准套件标准的真实模型。
幸运的是我找到了1287-11，这是非常详细的解释理论。
对于那些面临同样问题的人，我建议不仅要找到有效短路L的等效延迟，还要建议与此程序相关的有效损失。
实际上，短路模型是一个有损TL，它与一个非线性短路电感相结合。
当您计算由此电感引起的延迟并将其添加到有损TL时，您还会沿线引入更多损耗。
因此，在我看来，最好既要计算有效延迟，又要计算有效损失。
详细计算将留作练习，:-)



     以下为原文

  Thank you very much for your reply.

This inspired me to dig the agilent's app notes to find what's the real model used to model the cal kit standard. Luckily I found 1287-11, which is very detailed explaining the theory.

For those of you who are facing the same problem, I would like to suggest not only to find the equivalent delay of the effective short L, but also the effective loss associated with this procedure. Actually the model of the short is a lossy TL cascading a non-linear short inductance. When you are calculating the delay caused by this inductance, and add it to the lossy TL, you are also introducing much more loss along the line. So in my point of view it is better to both calculate the effective delay, as well as the effective loss.

Detailed calculation would be left as an exercise, :-)

> {quote：title = linan0827写道：} {quote}>非常感谢您的回复。
>>这启发我挖掘安捷伦的应用笔记，找出用于模拟校准套件标准的真实模型。
幸运的是我找到了1287-11，这是非常详细的解释理论。
我必须下载它。
听起来很有趣。
你的帖子让我思考（总是很危险...）我想知道是否有可能通过输入修改后的系数为C0，C1，C2和C3来“欺骗”8720使用L0，L1，L2和L3的值。
电容变为负值？
你可能认为我很生气，也许我是，但这是我的逻辑。
值L0，L1，L2和L3允许计算作为频率函数的电感。
给定L（f），更现代的VNA可能会计算出短路的电抗如何随频率变化。
由于电感电抗具有正号，因此这将是阻抗的虚部的正值。
由于我们拥有的古董不允许进入L0，L1，L2和L3，我想知道是否有可能欺骗VNA使用负电容值而不是正电感？
如果是这样，有可能找到系数C0，C1，C2和C3，它们给出负电容。
戴夫



     以下为原文

  > {quote:title=linan0827 wrote:}{quote}
> Thank you very much for your reply.
> 
> This inspired me to dig the agilent's app notes to find what's the real model used to model the cal kit standard. Luckily I found 1287-11, which is very detailed explaining the theory.

I'll have to download that. It sounds very interesting. 

Your post got me thinking (always dangerous...)

I wonder if it is possible to "trick" an 8720 into using values for L0, L1, L2 and L3, by entering modified coefficients as C0, C1, C2 and C3, such that the capacitance becomes negative? You probably think I am mad, and perhaps I am, but this is my logic. 

The values L0, L1, L2 and L3 allow one to compute the inductance as a function of frequency. Given L(f), a more modern VNA presumably works out how the reactance of the short varies with frequency. Since an inductance reactance has a positive sign, that will be a posiive value for the imaginary part of the impedance. 

Since the antiques we own don't allow entry of L0, L1, L2 and L3, I wonder if it is possible to trick the VNA into using a negative value of capactance rather than a positive inductance? If so, it might be possible to find coefficents C0, C1, C2 and C3, which give a negative capacitance. 

Dave

嗨，这真的是一个很好的思考，无论如何......实际上我的自定义校准套件（好吧它来自毛里......）短标准有明确的负L0值，这表明电容行为。
但是，如果在VNA的校准菜单中将标准定义为“短”，则无法输入C0 / C1 / C2 / C3系数，这仅在标准定义为“打开”时才有效。
因此，虽然您可以使用C0 / C1 / C2 / C3系数映射短标准，但是在我看来，没有这样的机制可以使用它们......或者您还有其他想法吗？



     以下为原文

  Hi,

It is really a good thinking,anyhow...

Actually my custom cal kits' (OK it is from maury...) short standard has explicitly negative L0 values, which is indicating a capacitance behavior.

However, if you define your standard to 'Short' in VNA's cal menu, you will be unable to put C0/C1/C2/C3 coefficients, which is only active when the standard is defined as 'Open'. So although you can map short standard with C0/C1/C2/C3 coefficients, there is no such a mechanism so that you can use them, in my point of view...

Or you have other thinking?

> {quote：title = linan0827写道：} {quote}>嗨，嗨，>这真是一个很好的思考，无论如何......我不知道好的思考。
这可能完全不可行。
如果这个想法根本无处可去，我不会感到惊讶。
但有时它有助于+开箱即用+。
>实际上我的自定义校准套件（确定它来自maury ...）短标准具有明确的负L0值，这表示电容行为。
我怀疑电感是否定的。
记住这些是多项式的4个系数。
您需要将4个值粘贴到多项式中，并将其作为频率的函数进行评估。
我很确定如果你这样做，那么校准套件设计工作的所有频率的整体电感都是正的。对于没有设计套件的频率，它可能是负的，但这不是问题。
>但是，如果在VNA的校准菜单中将标准定义为“短”，则无法输入C0 / C1 / C2 / C3系数，这仅在标准定义为“打开”时才有效。
因此，尽管您可以使用C0 / C1 / C2 / C3系数映射短标准，但是在我看来，没有这样的机制可以使用它们...您必须将其定义为打开，以便
可以输入C0，C1，C2和C3。
但是标准定义中的文本可以设置为显示短（F）或短（M），即使标准定义为开放。
>或者你有其他想法？
我希望我已经解释了我的想法。
在本质上。
1）绘制校准套件的电感与频率的关系图 - 即L（f）。
这应该总是积极的。
2）绘制电抗与频率的关系图，即X（f）。
这应该总是正3）看看是否有可能找到系数C0，C1，C2和C3，它们给出相同幅度的电抗，但与步骤2中计算的电阻符号相反.4）将菜单中的文本更改为
显示短而不是开放。
我不是*说这会起作用 - 只是值得一看。
人们需要更多的VNA知识，而不是我有更好的想法。
我想这值得一试。
如果这个想法至少值得继续下去，或者是否完全浪费时间，安捷伦的某些人可能会有更多的想法。
我怀疑任何人都会提前知道它是否有用，但有人可能知道它是否没有工作机会。
DaveEdited：drkirkby于2012年12月10日凌晨4点47分



     以下为原文

  > {quote:title=linan0827 wrote:}{quote}
> Hi,
Hi,
> It is really a good thinking,anyhow...
I don't know about good thinkhing. It might be totally unworkable. I would not be surprised at all if this idea leads nowhere at all. But sometimes it helps to +think out of the box+. 

> Actually my custom cal kits' (OK it is from maury...) short standard has explicitly negative L0 values, which is indicating a capacitance behavior.

I doubt the inductance is negative. Remember these are 4 coefficients of the polynomial. You would need to stick the 4 values into the polynomial and evaluate the inductance it as a function of frequency. I'm pretty sure if you do that, then the overall inductance would be positive at all frequencies for which the cal kit is designed to work It might be negative at frequencies for which the kit is not designed, but that is not an issue. 

> However, if you define your standard to 'Short' in VNA's cal menu, you will be unable to put C0/C1/C2/C3 coefficients, which is only active when the standard is defined as 'Open'. So although you can map short standard with C0/C1/C2/C3 coefficients, there is no such a mechanism so that you can use them, in my point of view...

You would have to define it as open, in order that C0, C1, C2 and C3 could be entered. But the text in the standard definitions can be set to display Short(F) or Short(M), even if the standard is defined as a open. 

> Or you have other thinking?

I hope I've explained my thinking. In essence. 

1) Plot a graph of inductance vs frequency for your cal kit - i.e. L(f). That should always be positive. 
2) Plot a graph of reactance vs frequency i.e. X(f). That should always be positive
3) See if it is possible to find coefficients C0, C1, C2 and C3 which give a reactance of the same magnitude, but opposite sign to that calculated in step 2.
4) Change the text in the menu to show Short instead of open. 

I'm *not* saying this will work - only that it might be worth a look at. One would need more knowledge of VNAs than I have to have a better idea. I guess it is worth a try. 

Perhaps someone from Agilent might have more idea if this idea is at least worth persuing, or whether it would be a complete waste of time. I doubt anyone would know for sure in advance if it would work, but someone might know if it has zero chance of working. 

Dave

Edited by: drkirkby on Dec 10, 2012 4:47 AM