We present high precision negative ion thermal ionisation mass spectrometry (N-TIMS) Os isotope measurement results for the DROsS reference material, to investigate the limits on the precision of TIMS-based 186Os/188Os measurements. We use analytical conditions previously highlighted to optimise precision, present a new flexible data processing protocol, and measure 184Os on a Faraday Cup equipped with a 1012 Ω amplifier. Despite analytical protocols that minimise uncertainty from counting statistics and Johnson-Nyquist noise, the intermediate measurement precision of our approach does not significantly improve on previous high precision Os isotope measurements, with the exception of 184Os/188Os. This is probably due to uncertainties in measured amplifier gain factors, which are greater when using mixed arrays of 1011 and 1012 Ω amplifiers than when using 1011 Ω amplifiers alone, though Faraday Cup deterioration could also contribute. We propose that multi-dynamic Os isotope measurements could largely eliminate both of these uncertainties. Our 184Os/188Os measurement results are the most precise yet, yielding 184Os/188Os = 0.0013036 ± 0.0000007 (2s, n = 38). Additionally, we average our data with published data to recommend the following composition for DROsS: 186Os/188Os = 0.1199319 ± 0.0000024, 187Os/188Os = 0.1609227 ± 0.0000022, 189Os/188Os = 1.219709 ± 0.000010, 190Os/188Os = 1.983793 ± 0.000011.

Supporting Information