An analyzer should respond uniformly to all solvents that might be used. Any differences between individual solvent response factors should not reduce the margin of safety. Typically, this means that the calibration is based upon that solvent producing the lowest response, so the analyzer indicates the true concentration of this one solvent and it indicates readings higher than actual concentration for all others in use.
Sometimes calibrations are made using weighted averages of response factors based upon the expected ratio of solvents in a mixture. Although this approach is specifically excluded by some authorities, it is still in common use because the practical use of an analyzer with a wide range of response factors can not allow economical operation otherwise.
Significant errors occur when the mixture does not actually match the calculated response factor, either by a change to the solvent composition in the coating mixture, or by the effects of varying vaporization rates in different dryer zones. At a minimum, the potential error from miscalculation, changes to solvent mixture, or variations in vaporization, should be studied to ensure that the margin of safety is not greatly reduced by this method.
Use of analyzers with response factor ratios equal to the margin of safety must be excluded. Depending on the solvents, infrared analyzers and FID analyzers can have response factors with enough variation to exceed a 50% L.F.L margin of safety. Catalytic bead sensors that have aged or been poisoned can also have sufficient variation to exceed the margin of safety.