Yashchuk Valeriy V
Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA.
Rev Sci Instrum. 2009 Nov;80(11):115101. doi: 10.1063/1.3249559.
Drifting of experimental setups with change in temperature or other environmental conditions is the limiting factor of many, if not all, precision measurements. The measurement error due to a drift is, in some sense, in-between random noise and systematic error. In the general case, the error contribution of a drift cannot be averaged out using a number of measurements identically carried out over a reasonable time. In contrast to systematic errors, drifts are usually not stable enough for a precise calibration. Here a rather general method for effective suppression of the spurious effects caused by slow drifts in a large variety of instruments and experimental setups is described. An analytical derivation of an identity, describing the optimal measurement strategies suitable for suppressing the contribution of a slow drift described with a certain order polynomial function, is presented. A recursion rule as well as a general mathematical proof of the identity is given. The effectiveness of the discussed method is illustrated with an application of the derived optimal scanning strategies to precise surface slope measurements with a surface profiler.
随着温度或其他环境条件的变化，实验装置的漂移是许多（即便不是所有）精密测量的限制因素。在某种意义上，由漂移引起的测量误差介于随机噪声和系统误差之间。一般情况下，漂移的误差贡献无法通过在合理时间内进行的多次相同测量来平均消除。与系统误差不同，漂移通常不够稳定，无法进行精确校准。本文描述了一种相当通用的方法，用于有效抑制各种仪器和实验装置中由缓慢漂移引起的虚假效应。给出了一个恒等式的解析推导，该恒等式描述了适用于抑制用特定阶多项式函数描述的缓慢漂移贡献的最优测量策略。给出了该恒等式的递归规则以及一般数学证明。通过将推导得到的最优扫描策略应用于使用表面轮廓仪进行精确表面斜率测量，说明了所讨论方法的有效性。
