The use of small rotors capable of very fast magic-angle spinning (MAS) in conjunction with proton dilution by perdeuteration and partial reprotonation at exchangeable sites has enabled the acquisition of resolved, proton detected, solid-state NMR spectra on samples of biological macromolecules. The ability to detect the high-gamma protons, instead of carbons or nitrogens, increases sensitivity. In order to achieve sufficient resolution of the amide proton signals, rotors must be spun at the maximum rate possible given their size and the proton back-exchange percentage tuned. Here we investigate the optimal proton back-exchange ratio for triply labeled SH3 at 40 kHz MAS. We find that spectra acquired on 60 % back-exchanged samples in 1.9 mm rotors have similar resolution at 40 kHz MAS as spectra of 100 % back-exchanged samples in 1.3 mm rotors spinning at 60 kHz MAS, and for (H)NH 2D and (H)CNH 3D spectra, show 10-20 % higher sensitivity. For 100 % back-exchanged samples, the sensitivity in 1.9 mm rotors is superior by a factor of 1.9 in (H)NH and 1.8 in (H)CNH spectra but at lower resolution. For (H)C(C)NH experiments with a carbon-carbon mixing period, this sensitivity gain is lost due to shorter relaxation times and less efficient transfer steps. We present a detailed study on the sensitivity of these types of experiments for both types of rotors, which should enable experimentalists to make an informed decision about which type of rotor is best for specific applications.