During a product upgrade by our client, resonances were introduced into the scanning mirror carriage of the spectrometer, affecting its performance. It was determined that the resonance was excited at each turnaround of the carriage, driven by a voice coil under closed loop control. The client was struggling to identify the cause. Instrumentation with accelerometers and analysis of the mirror feedback signal indicated a resonance with a significant longitudinal component at 4.6kHz, 0.6g pk-pk. FEA modal analysis identified a matching mode in the carriage. Fixing the problem was made more difficult due to the need to not increase mass, and maintain CG close to voice coil axis, however a solution was found using a truss style of design for the carriage.