WR-05波导中的TRL校准怎么设置

NIST的Dylan Williams和其他人写了一些关于毫米波和亚毫米波波导校准的论文。
解决了TRL校准的错误。
您没有提供有关您使用的校准标准和测试端口连接器的大量信息。
他们有精密定位销，环吗？
你的结果有多重复？
由于您使用2 3/4垫片来覆盖频带，因此垫片之间以及垫片和波导测试端口之间的未对准误差和孔径大小以及形状差异将导致校准误差在上下波段之间不同。
如果端口没有与校准垫片相同的机械尺寸和公差，则校准误差会更高。
直通连接行为将与两个3/4波形垫片连接不同。
肯



     以下为原文

  Dylan Williams of NIST and others wrote a few papers on mm wave and sub-mmwave waveguide calibrations.  Errors of TRL calibrations were addressed.  

You did not provide much information regarding the calibration standards and test port connectors that you used.  Do they have precision alignment pins, rings?  How repeatable are your results?

Since you are using 2 3/4 shims to cover the frequency band, misalignment errors and apperture size and shape differences between the shims and also between the shims and waveguide test ports will cause the calibration errors to be different betweent the upper and lower bands. 

If the ports were not made to the same mechanical dimension and tolerances as the calibration shims, calibration errors will be higher.  The thru connection behavior will be different from the two 3/4 wave shim connections.

Ken

波导端口以及cal标准都具有相同的精密法兰。
这些允许在波导的较宽壁的两侧使用两个额外的对准销。
我在Dylan Williams论文中可以找到的所有信息都与仅在110 GHz频率下的晶圆校准有关。
您能否告诉我在WR-05 140-220 GHz中校准的TRL校准等级如何？
假设使用两个垫片，延迟值分别为7.4042和4.9149。
特别是校准参考Z0和Testport参考平面的设置。
谢谢。



     以下为原文

  Both the waveguide ports as well as the cal standards have the same precision flanges. These allow for using two additional alignment pins on both sides of the broader wall of the waveguide. All the information I can find on Dylan Williams papers is to do with an on wafer calibration down at up to 110 GHz only. Could you tell me how the TRL Calibration Class Assignments should look like for calibration in WR-05 140-220 GHz? Assuming that two shims are used with delay values of 7.4042 and 4.9149, respectively. Especially settings for Calibration Reference Z0 and Testport Reference Plane. Thanks.

我不知道你在哪里以及如何获得参考资料。
我有Dylan Williams的以下论文，这些论文涉及毫米波和亚毫米波矩形波导校准。
D. Williams，“具有不完美测试端口的矩形波导矢量网络分析仪校准”，76 ARFTG Conference 2010年11月30日D. Williams，“500 GHz-750 GHz矩形波导矢量网络分析仪校准”，IEEE交易太赫兹科学
和技术，卷。
1，NO.2，2011年11月D. Williams，“Sub0mm波散射参数校准与不完美电气端口的比较”，IEEE与太赫兹科学与技术交流，Vol。
2，NO.1，2012年11月同样在类似主题J. Stenarson，K. Yhland，另一篇论文“对TRL和LRL校准wageguide宽度误差修改的影响”，第79届会议ARFTG 2012年6月为TRL Typeical类分配将是：直通
- 直通标准反射 - 短线/匹配：shim1，垫片2垫片1和垫片2，用适当的最小和最大频率定义。
注意，最小频率不必是截止频率。
截止频率由“连接器”定义定义。
与参考阻抗相同 - 由连接器阻抗定义。
对于波导，它是“1”。
您可以访问以下网址下载最新的PC校准套件编辑器：http：//na.tm.agilent.com/pna/apps/并选择19 MB Jun 2013版本。
新的cal kit编辑器修复了许多错误并且更容易使用。



     以下为原文

  I don't know where and how you get references.  I have the following papers by Dylan Williams that addressed mm wave and sub-mm wave rectangular waveguide calibrations.

D. Williams, "Rectangular-waveguide vector network analyzer calibrations with imperfect test ports", 76 ARFTG Conference  Nov. 30, 2010
D. Williams, "500 GHz-750 GHz Rectangular-Waveguide Vector Network Analyzer Calibrations", IEEE Transactions on Terahertz Science and Technology, Vol. 1, NO.2, Nov. 2011
D. Williams, "Comparison of Sub0mm wave Scattering parameter Calibrations with Imperfect Electrical Ports", IEEE Transactions on Terahertz Science and Technology, Vol. 2, NO.1, Nov. 2012

Also another paper on similar topic
J. Stenarson, K. Yhland, "Influence of wageguide width erros on TRL and LRL calibrations", 79th ARFTG Conference June 2012

Typeical class assignment for TRL will be:

Thru - Thru standards
Reflect - short
Line/ Match:  shim1, shim 2

shim 1 and shim2 defined with the appropriate min and max frequencies.  Note that min frequency need not be the cut off frequency.  Cut off frequency is define by the "connector" definition.  The same with reference impedance - defined by connector impedance.  For waveguide, it is "1".


You can download the latest PC cal kit editor by going to:
http://na.tm.agilent.com/pna/apps/  

and select the 19 MB Jun 2013 version.

The new cal kit editor fixed many bugs and is much easier to use.

> {quote：title = walezy写道：} {quote}>校准不会返回预期的结果，因为当其中一个垫片留在两个端口之间时，LogM类型的绘图上的S21和S12图表不平坦为0 dB
法兰。
当两个端口通过S21从零到零的背对背连接时，S12如预期的那样是平的。
你能发布S21和S12曲线图吗？
我怀疑你会在扫描的一部分看到它平坦而在另一部分不平坦。
它会根据垫片而改变。
如果垫片不是完全相同的（实际上是相同的而不是设计相同但是制作方式不同，或者连接方式不同）那么在频率上会有一个步骤改变使用哪条线。
在TRL cal中，无论标准的质量如何，直通和线路应该在它们是有效校准标准的区域内看起来平坦。
但由于你使用2行，每行只覆盖部分频率范围而另一部分由另一行调用。
你看到的波纹很可能是由于线路中的误差引起的实际真实残余波纹等。直通是平坦的事实只是卡尔的一个神器（你说它是平的，但是因此cal使它平坦，但是
仅适用于该阶段）。
尝试放入两个垫片（现在你有一个新线），你可能会看到整个乐队的波纹。
但这只是一个猜测，直到你发布了S21的糟糕情节。



     以下为原文

  > {quote:title=walezy wrote:}{quote}
>The calibration does not return the expected results as the S21 and S12 plots are not flat at 0 dB on LogM type of plot when one of the shims is left in between two port flanges. When the two ports are connected back to back with zero length thru the S21, S12 are flat as expected.

Can you post a plot of the S21 and S12 trace.  I am suspicious that you will see it flat on part of the sweep and not flat on another part.  And it will change depending on the shim.  

If the shims are not identical-identical (really identical rather than designed identically but made differently, or connected differently) then there will be a step at the frequency that changes which line is used.

In TRL cal, regardless of the quality of the standards, the thru, and the lines should look flat over the regions where they are the active cal standard. But since you use 2 lines, each line only covers part of the freqency range and the other part is cal'd by the other line.  The ripple you see is likely the actual true-residual ripple due to errors in the line, etc.   The fact that the thru is flat is just an artifact of the cal (you say it is flat but so the cal makes it flat, but only for that phase).  Try putting in both shims (now you have a new line) and you will likely see ripple across the band. 

But this is just a guess until you post of plot of the S21 badness.