Repeatability and reproducibility of fire tests for cigarette ignition of upholstered furniture composites

An analysis has been made to compare the repeatability and reproducibility of three tests for cigarette ignition of upholstered furniture composites: ASTM E1352, ASTM E1353 and NIST 851 mock-up. The first two of these tests are traditional methods designed to assess the potential of upholstered furniture components to being ignited by cigarettes while the last one, NIST 851 mock-up, is a new method designed to assess the propensity of cigarettes to ignite upholstered furniture composites. The traditional methods, ASTM E1352 and ASTM E153 use a single cigarette for each determination and can be run in two ways: to obtain a numerical output of char length on the substrate (following the standard) or to obtain a pass/fail result for the substrate (practical use); the precision was analysed in both fashions. The new method, NIST 851 mock-up, uses 144 cigarettes for each determination (48 each on three substrate composites), and assesses the fraction of substrates that have been ignited (char length exceeding 10 mm), i.e. by a combination of pass/fail data. The analysis for actual char length was made according to ASTM E691 guidelines, while the analysis of the pass/fail was made according to a modification applicable to binary data. The precision of the test methods was as follows (in descending order): NIST 851 > ASTM E1353 P/F > ASTM E1352 P/F > ASTM E1353 > ASTM E1352. (C) 1998 John Wiley & Sons, Ltd.