The basic principle of the improved passive flux sampler has been reported in a companion paper. In the present paper, the performance of different types of the sampler as a function of angle of incidence of airflow was studied in a wind tunnel. The following types of sampler were examined: a basic straight sampler (220 mm long); straight samplers (220 mm long) with openings that had been modified in one of several different ways, all aimed at reducing turbulence around the openings and thus achieving nearer-to-perfect performance as angle of incidence changed; a cranked sampler, designed to reduce the distance between the planes of the two openings to 80 mm (type useful in confined spaces); and two types of sampler (dual curved and angular) for both of which the two openings are in the same plane, albeit offset transversely (useful for space-boarded buildings). The dependency on the angle of incidence of the different sampler types was compared by measuring the pressure drop across the sampler, and hence calculating a normalised air velocity through each sampler's precision central orifice for each angle of incidence. Ideal performance would yield a cosine relationship between air velocity through the sampler and angle of incidence. Any deviations from such a cosine relationship imply a potential under- or over-estimation of ammonia flux rate, the magnitude of which depends on the relative persistence of the relevant angle of incidence during the exposure period of the sampler. The closest approach to ideal performance as angle of incidence changes was achieved by the straight sampler modified by the addition of solid tear-drop shapes around both its openings. However, the overall length of the straight sampler means that it is not suitable for all sampling scenarios. For confined spaces e.g. complex geometries in the roof ridges of livestock buildings, the cranked sampler is more suitable, accepting that it does not so closely approach ideal performance as the angle of incidence changes. Likewise, the dual curved and angular samplers are more suitable for ammonia fluxes through space-boarded walls or similar: in both the above scenarios, non-ideal performance is mitigated by the fact that, with defined air pathways in the roof ridge, and with a space-boarded wall of finite thickness, respectively, the range of angles of incidence likely to be seen by the cranked, dual curved and angular samplers is generally limited to angles near to 0 or 180°.