Abstract

Examples of six oxygen concentrators (DeVO2, Dom 10, Econo 2, Hudson, Permox, and Roomate) were evaluated over a 9-28 day period to determine (1) the oxygen yield (% O2) over the flow range 1-4 l min-1; (2) 90% oxygen rise time (90% RT) from a cold start when they were operated at 2 l min-1; (3) accuracy and readability of the flow device; (4) static outlet pressure; (5) major components comprising the product gas (Hudson only); and (6) general characteristics. At an outlet flow of 2 l min-1 the mean % O2 generated by all models, except the Permox (which was lower, mean (SD) 90.5% (3.1%), were between 94% and 95% with a range of less than +/- 0.5%. The Dom 10, Econo 2, and Hudson consistently generated higher oxygen concentrations than the other models at flow rates greater than 2 l min-1. The 90% RT was less than 10.5 minutes for all models. Deviations of up to 22% were observed between predicted and measured flow rates in all models except the DeVO2, Hudson, and Permox. It was possible to set the orifice type flow devices fitted to the Permox and Roomate between indicated flow settings, resulting in cessation of flow. Spectral analysis of the output of one device showed that argon and oxygen were concentrated to similar extents, indicating that the maximal attainable oxygen yield for a molecular sieve concentrator is about 96%.