Correcting for the dynamic response of a respiratory mass spectrometer.

Mass spectrometers produce distorted measurements of gas concentrations because of the time delays and rise times inherent in their responses. Three techniques for numerically correcting such distortion were applied to the acetylene step responses of a Perkin-Elmer MGA1100 mass spectrometer and to simulated data. The techniques investigated were 1) a simple time-delay correction, 2) an exponential model method that assumes a biexponential form for the peak of the impulse response, and 3) a Fourier transform method of deconvolution known as Wiener filtering. The time-delay correction produced an order of magnitude reduction in measurement error. The exponential model method improved on the time-delay correction, and the Wiener filter gave the most accurate corrections in all cases examined.