[mesa-users] pre-MS central T

[Tuguldur Sukhbold]

Dear MESA users,

I have one question just out of curiosity. In the past whenever I experience convergence problems trying to create a preMS model, I have been changing the starting central temperature, and it has been working 'great'. Today I noticed that in one particular model (attached inlist) it works when I change the default central T only by 10kelvins. 

Why such a small change (0.003%) affects convergence? People often attribute this kind of effects to the fact that we're solving highly non-linear set of eqn.s and small changes often propagate to create large effects in the end etc… but here were looking at rather simple case of just initial convergence.

My other question is, why MESA doesn't tweak this (and other similar parameters) automatically when it fails to find convergence?

thanks,
Tuguldur



&star_job

      mesa_dir = ''
      create_pre_main_sequence_model = .true.

      change_net = .true. ! switch nuclear reaction network
      new_net_name = 'o18_and_ne22.net'

      eos_file_prefix = 'mesa'
      kappa_file_prefix = 'gs98'

      set_rate_c12ag = 'Buchmann'
      num_special_rate_factors = 1
      reaction_for_special_factor(1) = 'r_c12_ag_o16'
      special_rate_factor(1) = 1.2

      set_rate_1212 = 'CF88' 
      num_special_rate_factors = 2
      reaction_for_special_factor(2) = 'r1212_to_ne20'
      special_rate_factor(2) = 1.0

!      pre_ms_T_c = 3.0001e5

      pgstar_flag = .false.
/

&controls

      initial_mass = 16.0d0
      initial_Z = 0.02
      initial_Y = 0.25

      photostep = 100
      profile_interval = 50
      history_interval = 1
      terminal_interval = 10
      write_header_frequency = 10

      RGB_wind_scheme = 'Nieuwenhuijzen'
      AGB_wind_scheme = '' 
      RGB_to_AGB_wind_switch = 1d-4
      Nieuwenhuijzen_wind_eta = 1.0d0

      xa_central_lower_limit_species(1) = 'h1'
      xa_central_lower_limit(1) = 0.72

      rate_factor_for_fe56ec_to_cr56 = 1d-4
      min_timestep_limit = 1d-12 ! (seconds)

      xa_scale = 1d-5
      newton_itermin = 2

      mixing_length_alpha = 1.0
      use_Ledoux_criterion = .true.
      alpha_semiconvection = 0.1

      overshoot_f_above_burn_h = 0.01
      overshoot_f_above_burn_he = 0.01
      overshoot_f_above_nonburn = 0.01
      overshoot_f_below_nonburn = 0.01
      overshoot_f_above_burn_z = 0.01
      min_overshoot_q = 1d-3 

      screening_mode = 'extended'
/

&pgstar

/