Alpha emissivity measurements are important in the semiconductor industry for assessing the suitability of materials for use in production processes. A recently published round-robin study that circulated the same samples to several alpha counting centers showed wide center-to-center variations in measured alpha emissivity. A separate analysis of these results hypothesized that much of the variation might arise from differences in sample-to-entrance window separations. XIA recently introduced an ultra low background counter, the UltraLo-1800 ("UltraLo"), that operates in a fundamentally different manner from the proportional counters used at most of the centers in the original study. In particular, by placing the sample within the counting volume, it eliminates the sample-to-entrance window separation issue noted above, and so offers an opportunity to test this hypothesis. In this work we briefly review how the UltraLo operates and describe a new round-robin study conducted entirely on UltraLo instruments using a set of standard samples that included two samples used in the original study. This study shows that, for LA ("Low Alpha" between 2 and 50 α/khr-cm2) sample measurements, the only remaining site-to-site variations were due to counting statistics. Variations in ULA ("Ultra-Low Alpha"<2 α/khr-cm2) sample measurements were reduced three-fold, compared to the earlier study, with the measurements suggesting that residual activity variations now primarily arise from site-to-site differences in the cosmogenic background. © 2014 Elsevier B.V.