The so-called t1 noise, which arises due to random instabilities in the spectrometer hardware, remains the primary source of noise that limits the sensitivity of most 2D NMR experiments, particularly in the expanding group of solid-state NMR methods that utilize dipolar-recoupling. In this communication we revisit the relationship between the signal intensity and the t1 noise produced. It is shown that since the latter scales linearly with the signal strength, the use of a conventional relaxation delay of 1.3T1 may prove far from optimal. In cases where the fluctuations occur on a shorter timescale than the recycle delay, a considerably faster repetition rate should be used to maximize the time sensitivity in a 2D experiment than what is used to maximize the sensitivity in 1D. This is demonstrated with the acquisition of 1H{13C} Dipolar-mediated Heteronuclear Multiple-Quantum Correlation (D-HMQC) type spectra in which the sensitivity could be nearly doubled by choosing a very short relaxation delay corresponding to 0.2T1.