a question about relaxation control
 

dear all,
recently I did some attempt on adjusting the relaxation factor in PHOENICS, seeing what will happen.
before doing this i was convinced that the relaxation should be very small for non-isothermal problems, but when i tried to make the relaxation bigger and bigger, i found that the divergence didn't happen ( however, nor did the convergence).

my question is, is the relaxation factor the bigger the better?