Comparison of conditions V=0 and A center dot n=0 on conductor boundaries in A,V-A-Psi formulations

Drago et al. (1994) proposed a Coulomb gauge A,V-A-Ψ formulation for quasi-static transient eddy current problems using V=0, instead of A·n=0, as gauge condition on the conductor boundary. Further numerical experimentation using either V=0 or A·n=0 on different parts of the conductor boundary, however, has shown that, rather unexpectedly, J·n=0 is numerically more weakly enforced where V=0 is used. Hence, in spite of some advantages of V=0, A·n=0 still seems to be a better choice, after all. A theoretical explanation for this unexpected drawback of V=0 has been found and is presented in detail