Abstract Isotope ratio mass spectrometry (with or without accelerator) relies on the principle that unknown samples and standards behave in the same manner in the instrument under the same conditions, allowing the isotope ratio of the unknown samples to be determined by comparison with the isotope ratios of the standards. The range of conditions within which this essential condition applies can be limited by many factors. We report here observations and tests (carried out at the Leibniz-Labor, Kiel, Germany) performed to determine the range of sample sizes/current outputs which can be safely analyzed in a given set. Since the fractionation effects are small under routine conditions, extreme situations (such as ten-fold ion current variations) had to be created where the effects would be seen more clearly. Experiments have shown that the 13C/12C variation as a function of current intensity is mainly due to a small difference in the position and direction of the 13C beam compared to that of the 12C and 14C beams at the exit of the recombinator. The cause of the 14C/12C fractionation versus current intensity is caused by side effects of the accelerator voltage stabilization which is using the 13C3+ beam.