We performed a bench (BS) and human (HS) study to test the hypothesis that estimation of tidal volume (V) during noninvasive helmet pressure support ventilation (nHPSV) would be possible using a turbine driven ventilator (TDV) coupled with an intentional leak single-limb vented circuit. During the BS a mannequin was connected to a lung simulator (LS) and at different conditions of respiratory mechanics, positive end expiratory pressure (PEEP) levels and leaks (30, 50 and 80 L/min). All differences were within the 95% limits of agreement (LoA) in all conditions in the Bland-Altman plot. The overall bias (difference between V measured by TDV and LS) was 35 ml (95% LoA 10 to 57 ml), 15 ml (95% LoA -40 to 70 ml), 141 ml (95% LoA 109 to 173 ml) in the normal, restrictive and obstructive conditions. The bias at different leaks flow in normal condition was 29 ml (95% LoA 19 to 38 ml). In the HS four healthy volunteers using nHPSV had a pneumotachograph (P) inserted through a mouthpiece to measure subject's V.The bias showed a scarce clinical relevance. In conclusions, V estimation seems to be feasible and accurate in all conditions but the obstructive one. Additional leaks seem not to affect V reliability.