We investigate which experiments are better suited to test the robust prediction that cosmic strings do not produce secondary Doppler peaks. We propose a statistic for detecting oscillations in the $C~l$ spectrum, and study its statistical relevance given the truth of an inflationary competitor to cosmic strings. The analysis is performed for single-dish experiments and interferometers, subject to a variety of noise levels and scanning features. A high resolution of 0.2 degrees is found to be required for single-dish experiments with realistic levels of noise. Interferometers appear to be more suitable for detecting this signal.