Guess for log(u) in the cooling Newton-Raphson scheme
I am looking at the EAGLE cooling code and at the top of cooling.c I can read:
static const double newton_log_u_guess_cgs = 1.414213562e6; /* log10(2e12) */
That does not seem right.
Either the comment is incorrect and should be sqrt(2e12)
or the comment is right but the value isn't. If it is the latter, could that explain why we have issues with the Newton scheme in some situations?