Superconducting-normalconducting-superconducting (SNS) transmons with 2-facet Al-shell nanowires are qubits compatible with magnetic fields above 10 mT. There are important correlations of the room temperature nanowire resistance with the chance of the qubit being measurable: at a resistance of 2−3 kΩ, the qubit is almost guaranteed to work. The chance of success halves every 2−3 kΩ increase. This information can be used to increase the yield. The flux noise power spectral density (PSD) of a model spin-1/2 fluctuator has been investigated as a function of the magnetic field using the Zeeman interaction. Not only the fluctuations parallel to the magnetic field contribute, but also the fluctuations perpendicular to the magnetic field. Cross-terms cancel out. The flux noise PSD of the SQUID is a linear combination of these spin PSDs when the spins are spatially uncorrelated. The magnetic field suppresses the parallel spin-axis noise PSD contribution as cosh^(-2)(μB/kBT). The magnetic field changes the perpendicular spin-axis PSD contribution due to the Larmor precession frequency peak 2fZ~μB, but does not influence the PSD contribution at frequencies higher that the Larmor precession frequency. When rotational asymmetry in the SQUID geometry is present, the PSD contributions of the perpendicular and parallel components can be separated. In our setup, a perpendicular coil is used to align the magnetic field with the transmon plane. The alignment procedure of maximizing the resonator frequency vs. the perpendicular coil field has been verified. To measure the flux noise, the perpendicular coil is first used to change the flux bias by large amounts. Then a dedicated flux bias is used to make a fine-grained sweep over the flux without flux-jumps, to calibrate the magnetic field at the SQUID. We have found a signal of the flux noise at zero field and at field. A flux noise amplitude of A~1000 μΦ_0 has been found at zero magnetic field.