The dimensionless aerodynamic particle sizer (APS) response function (normalized particle velocity against particle Stokes number) first reported by Chen et al. (1985) is explored for much larger solid particles (diameters to 35 pm) over a similar range of instrument pressures (624-740 mm Hg) and flow rates (4.2-6.0 L/min). An essentially unique response function is found for low and intermediate Stokes numbers under a variety of operating conditions, including the use of argon as the carrier gas. For large particles, however, non-Stokesian drag effects introduce systematic differences among calibration sets so that a unique response function no longer applies. The largest differences are observed between calibrations performed in air and argon, although even in this case the sizing error amounts to < 12% for a 20-pm polystyrene latex sphere. For intermediate Stokes numbers, a direct consequence of this work is that a reference calibration (channel number against Stokes number) can be used under different ambient conditions by setting the APS to operate at the same nozzle velocity as used in the reference calibration. With the single-velocity method, the factory-supplied calibration relating channel number to aerodynamic diameter can be used for air over a reasonable range of ambient temperatures and pressures. The same calibration can be used with an argon carrier gas provided that the aerodynamic diameters reported by the APS software are adjusted by the square root of the gas viscosity ratio. For the singlevelocity mode of operation, a generalization of a correction proposed by John (1987, 1989) can be made and is shown to reduce by one half the sizing error introduced by non-Stokesian drag.

The experimental response of the phase Doppler particle analyzer (Aerometrics, Inc., Sunnyvale, CAI to monodisperse oleic acid droplets in the 2.3-24-pm range is presented for three optical configurations. Significant nonlinearities for droplets with diameters < 20 p m were seen when using the standard configuration of a 495-mm focal length receiving lens at an off-forward scattering angle of 0 = 30". Of particular concern are response oscillations < 7 p m that lead to multivalued regions. In order to reduce these nonlinearities, two improvements to the standard receiver configuration were studied as suggested by theoretical studies reported in the literature. First, the receiver focal length was shortened to 240 mm (at 0 = 30") to increase the integration of scattered light. --As expected, the instrument response did significantly improve over the standard case, both in terms of linearity and scatter, for droplets > 5 pm. A significant departure from linearity (undersizing by 1.7 p m for a 2.3-pm indicated droplet diameter) was observed below this size, although the multivalued region of response was greatly reduced. Second, the receiving lens (focal length of 240 mm) was configured at a collection angle of 70°, close to the Brewster angle for oleic acid droplets where interference from reflected light should be minimized. Surprisingly, the performance of this configuration was comparable to the standard case (even though a shorter focal length was used), and inferior to the 240-mm/3O0 configuration.