[mesa-users] stellar-structural noise

[Bill Paxton]

Hi Alfred,

Okay. Thanks for the details.  The staggered mesh is probably irrelevant anyway since the wiggles aren't impacted by changes in mesh resolution.

Just for the record and for user's who may not have run into this issue, when interpolating cell center values such as density to cell boundaries, it is a good idea to take into account the different cell masses.  There are routines provided in star/public/star_lib to do this for you: 

      ! this routine returns an interpolated value at outer boundary of cell k
      real(dp) function star_interp_val_to_pt(v,k,sz,dq,debug_str)

and for mass fractions

      ! this routine returns an interpolated value of xa(j,:) at outer boundary of cell k
      real(dp) function star_interp_xa_to_pt(xa,j,k,sz,dq,debug_str)

Cheers,
Bill





On Oct 18, 2013, at 11:58 AM, Alfred Gautschy wrote:

> Bill,
> 
> thanks for your input!
> 
>> Just to follow up on Aaron's suggestion to check the EOS transitions,  the 1st to try is the temperature boundary between HELM and OPAL.  You can move it with these controls in &star_job
>> 
>>      ! transition temperature zone for OPAL to HELM at high T
>>      set_HELM_OPAL_lgTs = .false.
>>      logT_all_HELM = 7.7d0 ! HELM for lgT >= this
>>      logT_all_OPAL = 7.6d0 ! OPAL/SCVH for lgT <= this
>> 
> In the background my computer is working on your EoS transition
> suggestion. I will comment on the tests as soon I got some overview.
> 
>> What is the exact code you use to get U?  Perhaps something might show up there. 
>> 
> It is a *really* simple formula, and also performed well over the years:
> 
>   U = 4*pii*rsol3*10**(3*lgr+lgrho)/mss
> 
> rsol3: (solar radius)**3 since r is given in solar units
> lgrho: log10(rho) of rho in cgs
> mss  : mass in cgs
> 
> I account for the staggered mesh and interpolate rho to the interfaces; 
> however, whatever I do - right or wrong, so to say - the wiggles are
> unimpressed!
> 
> Just to make sure I did not miss a problem already in the dark past, I
> plotted also parts of the terms; the plots showed that the whole
> business is indeed well behaved ...except for some very slight changes
> in slopes in lgr and lgrho as a function of depth in the star (just the
> places I see the wiggles in the final quantity U).
> 
> Regards, Alfred
>