The resolution of TSI's 3071 Differential Mobility Analyzer (DMA) is smaller than ideally expected for nondiffusing particles, both at sheath airflow rates Q(s) in excess of 30 lit/min, and at aerosol over sheath airflow rate ratios Q(a)/Q(s) smaller than 10%. The first problem is partly due to a slight widening step following the laminarizing grid, which tends to destabilize the sheath airflow. The second can be associated in part to the recirculation vortex found by Chen et al. (1996) just upstream of the mixing region where the aerosol Bon joins the sheath airstream. It is also attributed here to mean Bow deceleration for sufficiently small values of Q(a)/Q(s), due to an increase in the how cross section following the mixing region. These problems are reduced in this study by replacing the DMA piece forming the outer electrode between the laminarization screen and the aerosol inlet into the analyzing region. The new part has a reduced inlet ID to avoid the initial step. It is slightly tapered, ending on an increased ID that permits accelerating and stabilizing the Bow following the mixing region. It is also longer than TSI's part, leading to a narrower inlet slit, which reduces the tendency for an unsteady recirculation bubble to form even at small Q(a)/Q(s) ratios. The performance is improved substantially at small Q(a)/Q(s) ratios and very substantially in the range of Q(s) from 30 to 80 lit/min as well as for Q(a)/Q(s) < 2%. Machining costs for the new part are of only a few hundred dollars. (C) 1998 American Association for Aerosol Research.
