The carrier phase (CP) measurements should be unambiguous to achieve the enhanced accuracy of Global Navigation Satellite System (GNSS) services. However, all the real-time CP measurements are manipulated by different errors, namely, satellite and receiver clock error, ionosphere delay, troposphere delay, integer ambiguity, etc. Even after removing or minimizing all these errors, the CP measurements are influenced by multipath error and receiver hardware residual or bias. In this research article, a double difference (DD) method is proposed to estimate residual or bias in CP observations of Navigation with Indian Constellation (NavIC). The proposed method abolishes all the common errors and provides bias. However, the CP measurement with a bit of bias value can be considered a high-quality signal. Here, we investigate the quality of NavIC satellites L5 (1176.45 MHz) signal with real-time measurements obtained through zero base length (ZBL) and short base length (SBL) experiments conducted at Kurnool (15.79 N, 78.07 E), India. The results show that the CP measurement bias of the L5 signal is within the range ±0.338 and ±1.688 m, respectively, for stationary and moving receivers. The bias was enhanced to 3.71 m for the SBL measurements. The results also demonstrate that the bias is low when carrier to noise (C/No) is high, and its range is nearly the same for all the days. This research would be helpful to estimate the bias that improves accuracy of precise point positioning (PPP) applications.