I. Overview

As a core component of electronic weighing instruments, load cells are listed as key managed measuring instruments in China. The current national standards and verification regulations for load cells in China are equivalent to the international recommendation R60 of the International Organization of Legal Metrology (OIML). Due to various reasons, the selection of sensors during the type evaluation and the description of certain clauses in the evaluation process of R60 are not clear enough, which leads to different laboratories in China adopting different methods for the type evaluation of load cells due to different understandings. As a result, load cells of the same family may lead to different evaluation results in different laboratories. As type evaluation involves the issuance of manufacturing enterprise licenses, it should be as standardized and unified as possible. The following is a discussion on individual clauses during the type evaluation process.

Ii. Selection and Coverage of Load Cells

When R60 and the verification regulations for load cells in China conduct type evaluation and select sensors for a family of load cells that include one or more groups of different scales and characteristics, the following three principles are adopted: First, sensors with the best metrological performance are given priority. That is, under the same measurement range, sensors with higher metrological performance can cover those with lower metrological performance. Second, the sensor with the smallest measurement range should be given priority. That is, among sensors with the same metrological performance, the sensor with the smallest measurement range is selected for the test, and the coverage is expanded from the smaller measurement range to the larger one at the specified ratio. Thirdly, the humidity test and additional tests for sensors equipped with electronic devices shall only be conducted on one of the sensors with the most stringent characteristics. Since the metrological performance of sensors depends on multiple parameters such as sensor level, maximum calibration division number, minimum calibration division value, distribution coefficient, and temperature range, it is difficult to determine which group has the best metrological performance or the strictest characteristics when several parameters other than the level of the sensors in each group of a family intersect. For instance, it is difficult to determine which group of sensors has better metrological performance between the 20t C-class nmax=3000 and Y=12000 and the 20t C-class nmax=4000 and Y=8000, which leads to inconsistent test schemes. As China has not yet formulated a national type evaluation outline for load cells, the type evaluation is conducted in accordance with the verification regulations. Therefore, when the application products encounter the above situations, it will also lead to inconsistencies in the test plans.

Iii. Calculation after the test of Load Cells

The verification regulations for R60 and load cells stipulate that after the test is completed, not only should the calculation be carried out according to n=nmax, but also according to n=nmax -500 and n=nmax -1000 (n≥500), and it is necessary to verify whether the condition vmin≤v is met. Then, the calculation and verification results should be filled in the summary table of test conclusions. As the specification itself does not provide a clear explanation of the above requirements, there are significant differences in the understanding among various technical institutions. One understanding is that since the maximum allowable error table is a stepped allowable error, the discontinuity at the critical point leads to the fact that sensors of the same level may fail the test at a smaller calibration division even if they pass the test at a larger calibration division. Therefore, recalculation should be carried out based on the smaller calibration division. This explanation seems to have some sense on the surface, but upon careful examination of C4.4 in R60 and the check items in the summary table, it can be seen that since vmin keeps increasing when calculated based on nmax-500 and nmax-1000, this means that the check items in the summary table will always be in line, so what's the point

Verification. In addition, since the test curve of the sensor is continuous and non-abrupt under normal circumstances, theoretically speaking, the error curve should also be continuous and non-abrupt. In this case, when a specific sensor (for example, C3) is calculated as qualified based on the nmax test, the probability of failure when calculated based on NMAX-500 and NMAX-1000 is extremely low (see Figure 1); The verification of the test results of numerous sensors with errors in a critical state did not find that this situation occurred, which also proved this point. In fact, even if there are sensors that pass the test but fail the calculation, it is often caused by sudden changes in the test data. This situation may be due to large errors brought about by factors such as the test equipment. Therefore, even if such a situation exists, it is very accidental. Such low-probability events can usually be ignored. Obviously, this understanding does not conform to the original meaning of R60.