This study aimed at establishing age-related norms for various aspects of timing: i) estimation of target intervals of 0.5 and 1 second in a rhythm-production context, ii) preferred and fastest tapping rates, and iii) detection and error correction of a phase-shift perturbation (PS) in otherwise isochronous sequences at 3 tempi from 1.11 to 3.33 Hz. Ten participants were musically trained (experts) and forty were not (non experts). The non experts were divided into 4 age groups: 19-35, 36-55, 56-75 and 76+ years old. Each just-detectable PS (JND) was estimated through the last 5 reversals of an adaptive staircase procedure (Levitt, 1971). Compensation functions (CF’s) were evaluated for PS below, at and above JND. The mean preferred and fastest rates were both fairly constant across age groups, fluctuating around 2 and 5 Hz respectively, except for the eldest group (76+) for whom they decreased by about 25 to 50%. The mean inter-tap intervals (ITI’s) produced by all but the eldest participants were quite close to the target durations of 0.5 and 1 second, though they suggest a slight underestimation of a 1-second interval. In contrast, the ITI’s obtained for the eldest group showed on average a 56% and 31% overestimation of 0.5 and 1 second respectively. JND’s did not vary significantly with age; however, they were lower for non experts (9 to 15 % of inter-onset interval or IOI) than experts (5 to 13 % IOI). Synchronization did not affect significantly PS sensitivity. Most participants aged between 19 and 75 exhibited normal CF’s (Repp, 2000). However, some of the eldest participants established a new post-shift baseline or failed altogether to correct for the timing error within the post-shift period of 5 tones, in particular for shifts that had not been reliably detected. The elderly participants performed best at the slowest 1.11 Hz tempo. The results are overall consistent with a slowing of the ‘internal clock’ starting from the mid-seventies onward. In old age, pre-attentive ‘automatic’ error detection and correction mechanisms may not be as efficient in sensorimotor synchronization; hence they might benefit from attentionally-driven cues and/or cognitive strategies such as those provided by musical expertise.