[Mesa-users] Low m high z stars - run in v10398 but not in v12778

[Samuel Baltz]

Hello,

when I attempt to run pre-main sequence stars with initial_mass from about
0.1 up to 0.5 solar masses, and varying initial_z from about 0.01 to 0.15,
I get the error "stopping because of problems dt < min_timestep_limit

terminated evolution: cannot find acceptable model

termination code: min_timestep_limit". This typically happens well before
the star reaches the main sequence. The strange thing is that this happens
in v12778, but I have no trouble running stars with identical settings all
the way through the main sequence in v10398.

Here are 3 troubleshooting steps I've tried. First, I replicated the v12778
error on two different machines with independent installations, using many
different inlists from the test suite. Second, previous mesa-users posts
have suggested that the which_atm_option (and the multiple controls it was
replaced with before v12778) might cause version-specific errors related to
metallicity, since some atm tables assume solar metallicity. The
controls.defaults file on my v10398 machine set which_atm_option to
"simple_photosphere" and that runs fine, so I used Rich Townsend's atm user
guide from September 2019 to enforce the equivalent settings on my v12778
machine (assuming that hasn't changed since v12115, but I couldn't find
anything about that in the v12778 release notes). However, I still got the
error, and trying many different atm_option and atm_off_table_option values
did not help either. Third, when I wasn't able to find a setting that would
make the star run longer, I tried copying over the controls.defaults from
my v10398 machine and updating the settings that were deprecated to their
equivalents in v12778; even running the same inlist with the same settings
(or their equivalents) didn't get rid of the error. Finally I should note
that lowering the min_timestep_limit can squeeze a few more iterations out
of the model but doesn't actually help it get to the main sequence (and I
believe many posts here have said that just changing the timestep limit is
rarely a solution to anything).

I've attached 2 files: my inlist from v12778 (some of the options that I've
tried varying are left as comments in the file, and nearly exactly the same
inlist works in v10398), and a csv that shows the number of iterations
before hitting the error for stars of 0.3 mass and varying metallicities. I
also included a few reference values from my v10398 machine to show how
much longer those runs are.

Thanks very much for any suggestions that anyone might have.

Samuel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20201002/9770dc5b/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: n_iterations.csv
Type: text/csv
Size: 265 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20201002/9770dc5b/attachment.csv>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: inlist_vlm
Type: application/octet-stream
Size: 2431 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20201002/9770dc5b/attachment.obj>