Nowadays, optically pumped magnetometers (OPMs) are gaining popularity in magnetoencephalography (MEG). However, like all sensors that work with biomagnetic signals, they are highly demanding to a low level of external magnetic field in the experimental environment. In the current paper, we proposed the experimental comparison of OPMs signal quality by recording their signal in two different magnetically-shielded rooms (MSRs) with different types of active suppression. Firstly, we conducted experiments in MSR in absence of additional local active shielding (only with one embedded into the walls). Next, we built the adaptive system of active suppression (ASAS), lightweight and portable, in another MSR and tested OPMs there. For this purpose we used sensors of two generations (Gen-1 and Gen-3) of commercial OPMs manufactured by QuSpin Inc. and tested them with standard signals generated by phantom and test coil. The results showed that MSR with installed ASAS in it enable OPMs to work optimally and give reliable results, while OPMs tested in the MSR containing SQUID system and active shielding embedded into the MSR walls, show unstable signal with low reproducibility. This study will undoubtedly facilitate research groups in building their own OP-MEG systems and understand the nature of false sensor deterioration.