Hello Everyone I am researching on LAMBDA implementation to calculate the integer ambiguity. y = Aa + Bb y is the observed double difference for each epoch . A is the design matrix with normalized LOS vector B is a diagonal matrix of the wavelength . As a first step I calculate the float solution and I need help with this. The model used are popular ones referred by P.J. de Jonge ,CCCJM Tiberius and Teunissen This is a sample y vector DD of code and phase 1.e+5* 0.0004 -0.0009 -0.0007 -0.0000 -1.9138 -1.0779 0.0778 0.2645 Question : 1. Can I just take y * y' (i,e y * y transpose ) to calculate the variance Qy? Else please advice me on the method. 2.Is there a sample code with numerical examples including actual GPS data that calculate the float ambiguities and variance matrix? I tried but covariance matrix Qa had a huge value order of e^10! thanks