An unbalanced measure of the approval of the vacuum gauge

The relationship between the mV value measured by the thermocouple and the degree of vacuum is as shown. Because the thermocouple vacuum measurement range is wide, it can cover 3 orders of magnitude, and the mv value and the vacuum degree are generally non-linear. Generally, a full-scale calibration is required. At this time, the uncertainty of the calibration result is to evaluate the full scale.

Uncertainty Analysis Thermocouple vacuum gauges have large differences in the uncertainty of different pressure points. Single-point repetitive measurement not only has a large workload but is not representative of the full scale. That is, the merits of the entire thermocouple vacuum gauge calibration result cannot be evaluated from one or a few points. For a thermocouple vacuum gauge, since full-scale calibration is used, how to perform full-scale global uncertainty assessment is the problem we need to solve.

The use of thermocouple vacuum gauge calibration data for full-scale uncertainty assessment In China, thermocouple vacuum gauges are generally calibrated using a second-class method. The thermocouple vacuum gauge is connected to the second-class standard capacitance film vacuum gauge at the corresponding position of the comparison system. The second-class standard measured value is used as the standard value (the uncertainty of the standard value is the second-class standard uncertainty, see 3.2) The thermocouple vacuum gauge is used as a measurement result.

Comparing the measurement results with the standard values ​​(this is different from the mVP curve) in the full scale range, a series of comparison data can be obtained. The calibration model can be considered at this time: ear = A ugly million; (l) where: Ge i ith i thermocouple indicator; slice ith standard value. For this we can use linear least-squares fitting to evaluate the overall uncertainty1.

For the calculation of linear fitting, many softwares have been available recently, such as Excel has this function. However, taking into account the wide range of thermocouples, the logarithm, and the calculation of uncertainty, I chose Origin software. On the other hand, if the number of points taken is higher where the measurement values ​​and the standard values ​​are in poor agreement, the overall uncertainty can be increased.

In this way, there is uncertainty or a large or small arbitrariness. To solve this problem, we must try to uniform the points when reading data, that is, to make the data points as equal as possible, in general. It is sufficient to take the same number of points evenly for each level.

The fitting method of the origin software is to do fitting after measuring the yellow point. Logarithm fitting after taking the logarithm of the measurement point We took a thermocouple as an example and made a full-scale calibration and processed it. Since the full scale spans several orders of magnitude, if the linear fitting is performed directly, the fitted uncertainty band exhibits a double flare shape, and the uncertainty value is in the form of an absolute value (see Table 1), which is for a wide range of Vacuum gauges are obviously not suitable. We take the logarithm of the measured value and the standard value, and then perform a linear fit, as shown.

The data points thus taken are relatively evenly distributed over the fitted straight line, and the uncertainty band is parallel to the fitted straight line. Take the logarithm of the measurement results, and calculate the uncertainty back again, it seems very complicated, but we can do the following processing: Xi s = 10 axi (5) because Y two lgy, if you differentiate on Y, you can get: dY = Industry y (6) The relative uncertainty can be obtained directly, which is why the uncertainty band is parallel to the fitting line. The essence of taking the logarithm is to change the multi-range problem into a uniform calibration problem, and the relative uncertainty can be directly obtained by further derivation.

In actual work, there will be a high vacuum, indicating that the number is non-zero. For the digital instrument used in this example, the calibration model should be y, which is obviously too complicated. Since p. is very small, only a slight correction is needed in the first order of magnitude, and the impact on the evaluation of uncertainty can be ignored. In Origin software, only a straight line with a slope of 1 is fitted during fitting. The tables are the results of a linear fit of direct linear fit and b=1, respectively.

Machinery Clamps

Machinery Clamps,Metal Clamps,Metal Strap Clamp,Small Metal Clamps

Ningbo Metal Sharing Supply Chain Management Co., Ltd , https://www.zenlesmart.com