Agilent 3458A不确定度计算

我不是统计专家，但我四处询问，这是我收到的回复：3458A确实很奇怪，不幸的是答案并不那么简单。
这是一个多方面的答案。
- 如果我要为这个产品计算扩展的MU，我会使用rms噪声的RSS（Root Sum Square）精度规格，尽管从规格可以看出RMS噪声贡献者与精度规格相比将非常小。
- TC规范需要直接添加到精度规范中。
如果您处于±5℃的校准温度的稳定环境中，并且执行ACAL 24小时和±1℃的ACAL温度，则不添加TC规范 - 线性度测量不是精度测量的因素，除非您在相关测量中进行相对测量。
相同的范围。
在这种情况下，您只能使用线性规范而不是我们已解决的其他组件。
例如，如果我在10V范围内，我想确定相对于10V测量的9V测量的测量不确定度。
然后我会使用线性度规范并忽略精度规格。
另一方面，如果我想要绝对MU，我只会使用上面的精度和噪声规格。
- 大多数人忘记的另一个因素是可追溯性错误的绝对准确性。
在这种情况下，我相信你将它添加到准确度规范中。我还附上了一份旧的白皮书，讨论了测试限制的来源，并且还包含统计部分.WHERE 3458A TEST LIMITS来自.zip



     以下为原文

  I am not a statistics expert but I asked around and this is the response I received:

The 3458A is a strange one indeed, unfortunately the answer isn’t as straightforward.  It is a multifaceted answer.

-      if I were to calculate the expanded MU for this product I would RSS(Root Sum Square) the accuracy spec with the rms noise, although from the specifications you can see that the RMS noise contributor will be very small compared to the accuracy spec.
-     The TC specification needs to be added directly to the accuracy specification. If you are in a stable environment ±5degC of cal temperature, and performed ACAL 24 hours and ±1degC of ACAL temp then the TC specification is not added
-     The linearity measurements are not a factor for accuracy measurements unless you are doing relative measurements in the same range. In this case, you would only use linearity specs and not the other components that we have addressed. For example, if I’m on the 10V range and I want to determine the measurement uncertainty of a 9V measurement relative to a 10V measurement. I would then just use the linearity specifications and ignore the accuracy specs. On the other hand, if I want the absolute MU I would only use the accuracy and noise specifications from above.
-     One more factor that most people forget about is the absolute accuracy of the traceability error. In this case, I believe that you add this to the accuracy specifications.

I have also attached an old white paper that discusses where the test limits came from and also contains a statistics section.WHERE 3458A TEST LIMITS COME FROM.zip

附件

• WHERE 3458A TEST LIMITS COME FROM.zip41.9 KB
  

非常感谢你！
这真的很有帮助还有一件事 - 我们将3458A送到安捷伦进行校准。
在提出规格时，您是否将所用校准标准的不确定性考虑在精度规格中？



     以下为原文

  Thank you very much! This was really helpful

One more thing - we send our 3458A to Agilent to get calibrated.  When coming up with specifications, do you factor the uncertainty of the calibration standards used into the accuracy spec?