Methods for inspection of standard samples or sample uniformity

SJ/Z 3206.9-1989 Standard sample or sample uniformity test method SJ/Z3206.9-1989 standard download decompression password: www.bzxz.net
This standard is applicable to the uniformity test of cotton block, powder standard samples (referred to as standard samples) or samples manufactured in large quantities. The test adopts emission spectrum analysis technology. This method is formulated based on data statistical theory.

Guiding technical document of the Ministry of Electronics Industry of the People's Republic of China Standard sample or sample uniformity test method SJ/Z3206.9-89
This method is applicable to the uniformity test of mass-produced rod-shaped, powdered standard samples (referred to as standard samples) or samples. The test adopts emission spectrum analysis technology. This method is formulated based on data statistics theory. 1 Reference standard
GB9259-88 Terminology of emission spectrum analysis 1.1
1.2SJ/Z3206.7--89 General rules for the preparation of standard samples for spectrum analysis 1.8SJ/Z3206.14-89 General rules for error and experimental data processing methods for spectrochemical analysis 1.4SJ/23206.8-89 General rules for sampling and sample preparation methods for emission spectrum analysis 2 Description
2.1 This method uses the measurement data of a certain number of samples randomly selected to determine the uniformity of the overall standard sample. 2.2 The test instruments and equipment should not cause systematic errors in the measurement data, otherwise they must be calibrated. 2.3 This method only calculates the measurement data of one component to be measured, and other components can be processed in the same way. 2.4 The analytical line or analytical line pair used should not be interfered by other spectral lines or affected by the background. 2.5 The confidence level adopted by the method depends on the product technical conditions that the standard sample to be tested should meet. 2.6 The data used in the calculation can be the content of the component to be tested, or other measured values ​​representing the content. 3 Sampling
3.1 For rod and block standard samples, 30 to 50 representative samples are taken from different parts of each standard sample, numbered and labeled, and at least 15 samples are selected for inspection by drawing lots or using a random number table (see Appendix A). 3.2 For powder standard samples, 30 to 50 samples are randomly taken from the standard sample packaging bottle, numbered and labeled, and at least 15 samples are selected for inspection by drawing lots or using a random number table (see Appendix Λ). 3.8 If there are less than 15 samples, all samples need to be inspected. 4 Inspection method
4.1 Using the best spectral analysis conditions, take m samples, and record or read the analysis results of the measured components n times for each sample. Generally, n should be more than 4 times. The data are listed in Table 1. Approved by the Ministry of Electronics Industry of the People's Republic of China on February 10, 1989 and implemented on March 1, 1989
Average value
Note, in the table:
Tm-2 Xmis
tm=Tm/ng
mNumber of samples,
n--Number of repeated measurements,
T-Sum of each column,
B-Sum of each row,
t--Average value of each column,
G-Total. bzxZ.net
SJ/Z3206.9--89
Table 1 Uniformity Test Data
B,-X.+Xa..+Xml
B,=X+X.+..+Xm2
B,-X,n+Xn+....+mm
i=lj=1
If repeated measurement data cannot be obtained for some reason, the corresponding degrees of freedom should be appropriately changed during calculation, or test data should be supplemented.
4.2 Calculation
4.2.1 Calculate the correction values ​​of T, t, B, and G in Table 1 CF=G*/mn
4.2.3 The sum of squares of all measurement data St
(Xi,)+-G/mn
i=lj=l
Wherein, X- is a single measurement value.
4.2.4 Sample square sum Sm
Sm=［(Ti+T+.+T2)/nG\/mn
4.2.5 Sum of squares Ss
4.2.6 Standard deviation S
SJ/Z3206.9-89
S,-［（Bi+B+...+B2)/m1-G\/mn(4)
S=(St-S=-S.)/(m-1)(n-1)) ★
4.2.7 Select the confidence level α of the test. This method adopts 95%. The q value related to the number of samples and the degree of freedom f of the measurement error can be found from the q table (see Appendix B). f=(m-1)(n-1)
4.3 Determine uniformity
4.8.1 Use the T test method. The test statistic is d = q (m, f, α) s / Vn || tt || 4.3.2 If the difference between the average values ​​of any two samples is d = 1t: -t| is greater than the critical value of the test statistic, that is, d>dT, it means that the standard sample is not uniform at the 95% confidence level, when d;
Tip: This standard content only shows part of the intercepted content of the complete standard. If you need the complete standard, please go to the top to download the complete standard document for free.