ARCHIVED REPORTS_1988_2 - Laserfiche WebLink

Home Browse Search

7.3.2.2 A newly prepared chamber working standard is rejected unless the relative response factors for all eight halogenated hydrocarbons of interest fall within +/- 10% of the historically established mean values. 7.5.2 The mean -square error due to lack of fit about the regression line is compared to the total mean -square error of the independent replicates about their individual means. The calibration is accepted if the F -ratio is less than the 95% rejection limit. 7 7.3.3 A working chamber standard is checked at least every three months for conformity to criteria 7.3.2.1 and 7.3.2.2. 7.3.3.1 A new standard is prepared as frequently as required as determined by the above mentioned criteria. 7.3.3.2 Any reports generated after the standard ceases to be demonstratively within the established tolerances shall contain a cautionary explanation. 7.4 The gas chromatograph is calibrated periodically. 7.4.1 Calibration factors are determined on the basis of the mean values of the previous calibration runs which meet the criteria of 7.4.3. 7.4.2 Each day a calibration check is performed using the Pfaudler chamber standard to span the instrument. 7.4.3 If the response for each compound of interest is within 10% of the established calibration value, the established calibration factors are retained. 7.4.4 The calibration check is repeated if the response of the instrument has changed by more than 10% from the established values. 7.4.5 If the response is still out of tolerance, a quality assurance report is submitted, remedial action is initiated, and new calibration factors calculated. 7.4.6 Blank samples shall be analyzed daily after the calibration is completed and, whenever necessary, between samples. 7.5 The linearity of the instrument is checked periodically. 7.5.1 A gas chromatographic multipoint linearity check is performed annually with standards of at least four different concentrations and four replicate runs for each concentration. The concentrations should include the anticipated range of sample concentrations above the limit of d6tection. 7.5.2 The mean -square error due to lack of fit about the regression line is compared to the total mean -square error of the independent replicates about their individual means. The calibration is accepted if the F -ratio is less than the 95% rejection limit. 7