Regarding time step in simulation with sink particles.

[srihawk...@gmail.com]

Hello Everyone,

I am running simulation with sink particles for agn feedback. The simulation is running fine but after sometime the time step becomes very small (values of 1e7 - 1e9).  Whenever  I restart from last snapshot then it runs fine after few hours before time step getting small again.

For simulation with seed mass 1e6, after a day the time step become small while for seed mass 1e7 time step become small after a few hours.  

I would like some insights what I am missing in this problem and is there any way to improve the time step value. Thanks in advance.

Regards 

Sriram

[Arturo Nuñez-Castiñeyra]

hi, 
It might be that you have some cells with very high temperatures, which will make your timestep go wild. This is usual for the AGN runs. You can try putting a cap on the maximal temperature with T2max=1e10 for example.

Saludos

Arturo

[srihawk...@gmail.com]

Hi Arturo,

Thanks for your reply. Yes it seems like the problem is with AGN runs, also I  gave energy fraction for both thermal and kinetic mode feedback as 1. Would that be an issue? Because even after putting cutoff, the time steps are still going crazy. I am pasting my sink parameters for reference. 

smbh=.true.
agn=.true.
create_sinks=.true.
mass_sink_direct_force=0.0
mass_halo_AGN=3e9
mass_clump_AGN=1e9
mass_sink_seed=1.77e7
mass_smbh_seed=1.77e7
accretion_scheme='bondi'
eddington_limit=.true.
acc_sink_boost=1
verbose_AGN=.true.
AGN_fbk_frac_ener=0.5
AGN_fbk_frac_mom=0.5
T2_min=0
T2_max=1e8
sink_descent=.true.
fudge_graddescent=0.1