Since Middleton showed the potential for a sputter source working from CO2, we have been building a source which can automatically change samples. The source produces a maximum beam of 25μA of C−, with typical operation between 10 and 20μA. Although beam generation from the source is very reliable, the mechanics of sample changing have given considerable problems. The changing of samples also involves considerable care in gas handling, and a computer control system has been written which ensures the correct sequence of the 16 operations required for sample change.