[mesa-users] Hate to pester but still problems at low masses

[Bill Paxton]

On Mar 25, 2015, at 9:24 AM, Michal Simon wrote:

> MESA still hangs up at approx 0.5 MY when calculating  the evolution 
> of a 0.1 Msun star with alpha(MLT) = 1.00 or values near 1, and similarly
> for 0,2 Msun.  I inserted into my in list_project the lines you suggested (and
> attach the in list).  With or without those lines MESA still hangs up at about the
> same time step.
> 



Hi Mike,

Last time I noticed that we were on the boundary between OPAL and SCVH eos and blamed that for the problem.
I should have kept looking!   Since the pre-ms models are fully convective it didn't occur to me that opacities could be a problem.  But of course the surface BC's depend on opacity even for the fully convective case.  And at the point where your models hang up, we've hit a boundary in the selected opacity tables.    The logR = 1 diagonal line is the edge of the tables for OPAL and also for the Alexander-Ferguson low T tables.  

As your pre-ms model evolves, the logR at the surface increases and eventually hits the logR = 1 barrier, and the result is nasty.  By reducing the tolerances we can get the solver to get through the bad spot, but the results aren't pretty as seen in the resulting HR track (this is the track we'd have seen last time if we had looked!).

The Freedman lowT tables go to higher logR, so if you can use them for your application, they offer a solution to the problem.  In fact, by switching to the Freedman logT tables we no longer need to relax the solver convergence tolerances -- they can go back to the standard settings.  The HR track looks like this after adding the following lines to the &star_job controls and removing the changes to tol_correction_norm and tol_max_correction in &controls.

      kappa_file_prefix = 'gs98'
      kappa_lowT_prefix = 'lowT_Freedman11'



This leaves the question of why is this a problem when have MLT alpha = 1.0 but not for larger alpha such as 1.8.   Turns out that it in fact IS a problem for larger alpha, but no one has noticed it (or at least I haven't noticed it and no one has called it to my attention).   Instead of causing convergence to fail, for larger alpha we just get a glitch in the HR track like this (for alpha = 1.8).

Again we are running off the edge of the Ferguson lowT table, so switching to the Freedman table eliminates the glitch for higher alpha too.

NOTE: just to be perfectly clear for the record, I'm not saying the Freedman lowT opacities are "better" than the Ferguson ones.  In cases where both apply it is up to the user to make a careful choice based on the needs of the application.  


Thanks for running into this problem!   It's been educational for us all. ;D

Cheers,
Bill
















-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: inlist_project
Type: application/octet-stream
Size: 1377 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment.obj>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: plot_T_RHO_kap.pdf
Type: application/pdf
Size: 50218 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment.pdf>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment-0002.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: H_R_off_fa_table.pdf
Type: application/pdf
Size: 22793 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment-0001.pdf>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment-0003.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: H_R_freedman.pdf
Type: application/pdf
Size: 22745 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment-0002.pdf>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment-0004.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: H_R.pdf
Type: application/pdf
Size: 22798 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment-0003.pdf>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150325/841f1963/attachment-0005.html>