Constant-time tensor correlation under magic-angle spinning conditions is an important technique in solid-state nuclear magnetic resonance spectroscopy for the measurements of backbone or side-chain torsion angles of polypeptides and proteins. We introduce a general method for the design of constant-time tensor correlation experiments under magic-angle spinning. Our method requires that the amplitude of the average Hamiltonian must depend on all the three Euler angles bringing the principal axis system to the rotor-fixed frame, which is commonly referred to as non-gamma encoding. We abbreviate this novel approach as COrrelation of Non-Gamma-Encoded Experiment (CONGEE), which exploits the orientation-dependence of non-gamma-encoded sequences with respect to the magic-angle rotation axis. By manipulating the relative orientation of the average Hamiltonians created by two non-gamma-encoded sequences, one can obtain a modulation of the detected signal, from which the structural information can be extracted when the tensor orientations relative to the molecular frame are known. CONGEE has a prominent feature that the number of rf pulses and the total pulse sequence duration can be maintained to be constant so that for torsion angle determination the effects of systematic errors owing to the experimental imperfections and/or T(2) effects could be minimized. As a proof of concept, we illustrate the utility of CONGEE in the correlation between the C' chemical shift tensor and the C(α)-H(α) dipolar tensor for the backbone psi angle determination. In addition to a detailed theoretical analysis, numerical simulations and experiments measured for [U-(13)C, (15)N]-L-alanine and N-acetyl-[U-(13)C, (15)N]-D,L-valine are used to validate our approach at a spinning frequency of 20 kHz.