Formal errors of solutions too small - Interpretation of the cofactor matrix

[matteocappuccio]

Dear PRIDElab admins.
We are analyzing the outputs of a multi-constellation static GNSS processing, and we are wondering about the effectiveness of its estimated error. We compute the standard deviations of the XYZ components as SigmaX/Y/Z=Sigma0*Sx/y/z. Values are in the order of10e-5 millimiters....which is nonsense.
Are we doing some mistake about the dimensional interpretation of Sigma0 or of the cofactors?
Could you please clarify this for us?

We thank you for your kind response.

Matteo & Luca