Carrier phase and pseudorange disagreement as revealed by Precise Point Positioning solutions

In GNSS data reduction, carrier phase (phase) and pseudorange (code) data are complementary. An illustrative example of their interplay is provided, and then it is shown that frequency biases in phase data can be estimated by examination of the difference between the code and phase residuals in Precise Point Positioning (PPP) solutions. Apparent frequency biases, in some cases approaching 0.2 ns/day have been found, although many are an order of magnitude less. These frequency biases could be due to small design imperfections in the GPS receivers. We have also noted that PPP processing is sensitive to the relative weights given the pseudorange (code) and the phase in the sense that down-weighting the code by a factor of 10,000 is preferable to down-weighting by a factor of 10 billion; we think we understand the reason for this.