With molecular gases, this may be possible if you can sample at multiple wavelengths or wavelength bands This could be done possibly using filters in front of several photodiodes or with a prism/grating and photodiodearray. You might be better off just buying the optical system from someone like Ocean Optics, rather than designing it yourself. The emission spectra of most molecular gases are very broad and overlap highly so you might need to using multivariate techniques (discriminant analysis) to analyze the data even for yes/no questions. These can be implemented in a microcontroller and there are numerous reference in analytical chemistry literature (try Applied Spectroscopy, Analytical Chemistry, Journal of Chemometrics in any college library for references). Later Matt ----- Original Message ----- From: "Peter L. Peres" To: Sent: Wednesday, February 28, 2001 2:07 PM Subject: Re: [OT]: Spectrometry with spark or corona > Matt Mosley > >Actually I know a little about them. Spark and Arc sources were > >traditionally used for atomic analysis of solid metal samples, but suffer > >from sampling problems and matix inconsistency. The resulting spectra > >can > > Matt, thank you for the insight. I am just trying to make a yes/no > decision on the presence of a hefty (5% or more) percentage of gas in air. > I know that the broadened spectrum can be very hard to interpret but I do > not need to determine % and each component. It would be nice to be able to > tell which gas it is though. > > Peter > > -- > http://www.piclist.com hint: The PICList is archived three different > ways. See http://www.piclist.com/#archives for details. > -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.