The Total Electron Content (TEC) of the ionosphere can be studied with a linear combination of measurements, known as ionospheric observable, from Global Navigation Satellite Systems (GNSSs). Two types of measurements are commonly used to derive the ionospheric observable, the pseudorange and the carrier phase measurements. The pseudorange measurement is absolute but noisy while the carrier phase measurement is accurate but ambiguous. The key to use the carrier phase measurements as precise unambiguous pseudoranges is to isolate the ambiguity term in the carrier phase measurement. We use three methods to retrieve the ionospheric observable, namely the Carrier to Code Levelling (CCL), the Undifferenced and uncombined Precise Point Positioning (UPPP) and PPP-Ambiguity Resolution (PPP-AR). The results achieved are: (1) By using data from an entire month within the last Solar Cycle maximum, the PPP-AR ionospheric observable, which is the most accurate, is found to be over one order of magnitude more accurate than the least accurate CCL ionospheric observable. The observation error of the PPP-AR ionospheric observable ranges from 0.05 to 0.11 TECU (Total Electron Content Unit) while that of the CCL ionospheric observable is from 0.65 to 1.65 TECU. (2) The time-varying receiver Differential Code Bias (DCB), which presents clear Day Boundary Discontinuity and intra-day variability pattern, contributes the most part of the observation error of the ionospheric observable. This contribution is assessed by the short-term stability of the Between-Receiver DCB, which ranges from 0.06 to 0.17 TECU in a single day. (3) The remaining part of the observation errors presents a sidereal time cycle pattern, indicating the effects of the multipath. Further, the magnitude of the remaining part implies that the code multipath effects are much reduced.