[Mesa-users] A Query regarding swithcing off magnetic braking

Tue Jul 18 15:38:20 UTC 2023

Hello,

> But in other cases, it stopped automatically with both mb on and off. For
example, I am sharing a screenshot of one of the runs with mb = on.

In the screenshot you shared, your model is crashing because of a timestep
dive, since the solver cannot find an acceptable solution. This is likely
for the same reason I previously stated.
You have a cold ~1d-2 Msun model which is sampling opacities from outside
the tabulated opacity table regime, and the solver is trying its best to
find a solution.

If your goal is to run the model further,  it was suggested to me by Giulia
Cinquegrana that  "you can make custom molecular opacities with AESOPUS
that extend further than logT=2.7". However do keep in mind that you might
encounter other unforeseen issues as your model descends lower in mass. My
recommendation is to establish a universal stopping condition for your runs
that satisfies your project goal.

> If I manually stop a seemingly unending evolution in the terminal after a
huge number of retries, can that be considered a final result for that
evolution?

I guess, but this is not a consistent approach. I would recommend finding a
good stopping condition that you can use for all your runs. e.g.
star_mass_min_limit = 2d-2.

-EbF

On Mon, Jul 17, 2023 at 11:58 PM ABHIJNAN KAR 20064 <
abhijnank20 at iiserbpr.ac.in> wrote:

> Also another trivial question: If I manually stop a seemingly unending
> evolution in the terminal after a huge number of retries, can that be
> considered a final result for that evolution?
>
> On 18-Jul-2023, at 11:16 AM, ABHIJNAN KAR 20064 <
> abhijnank20 at iiserbpr.ac.in> wrote:
>
> Hi Ebraheem,
> Thanks for the help! However, in my case, this unending MESA evolution
> occurred only in some cases(that too with all of the cases being mb on).
> But in other cases, it stopped automatically with both mb on and off. For
> example, i am sharing a screenshot of one of the runs with mb = on.
> <Screenshot 2023-07-18 at 11.15.37 AM.png>
>
> On 18-Jul-2023, at 2:23 AM, Ebraheem Farag <ekfarag at asu.edu> wrote:
>
> Hello Ojha,
>
> Both models, with and without magnetic braking, reach a point where the
> doner's mass is sufficiently small ~ 1e-2 Msun. At this point, both models
> are simultaneously quite cold and low in density.  A large portion of these
> very low mass models live in the "extended" section of opacity tables (at
> logR > 1, where logR = logrho - 3*logT + 18). In this regime the MESA
> opacity tables are extrapolated from the logR = 1 edge. The Fa05_gs98 lowT
> opacity tables being extended in this regime only extend down to logT =
> 2.7. It doesn't look like there is a condition in MESA to generate opacity
> values for temperatures lower than logT = 2.7, so it appears the code is
> returning a NaN for the opacity here, for both models.
>
>  rho =     2.5146799359618161D-06
>  lnrho =    -1.2893365024554088D+01
>  T =     4.9966225551057875D+02
>  lnT =     6.2139323811978784D+00
>  logKap =                        NaN
>  opacity =    -9.9999999999999997D+98
>  dlnkap_dlnd =                        NaN
>  dlnkap_dlnT =                        NaN
>  d_opacity_dlnd =    -2.4822215369416772D-08
>  d_opacity_dlnT =     7.2510282819076687D-03
>
> for:
>  logrho ~ log10(2.5d-6) = -5.602059,
>  logT ~ logT(500) = 2.69897 < 2.7d0,
>  logR = 4.3010 >> 1d0
>
> One model timestep dives and crashes, while the other rides the edge of
> the table finding hotter solutions with logT>= 2.7 which still satisfy the
> structure equations, though both models effectively stall for the same
> reason. With or without magnetic braking, this result is inevitable.
> This appears to be a bug that could be fixed, and I will open up a bug
> report on the topic, but for the time being, I would recommend setting a
> lower limit on your model to step the evolution when the doner's mass
> decreases beyond 0.02 Msun?
>
> star_mass_min_limit = 2d-2
>
>
>
> Hope this helps.
>
> -EbF
>
>
> On Sat, Jul 15, 2023 at 10:41 AM ABHIJNAN KAR 20064 <
> abhijnank20 at iiserbpr.ac.in> wrote:
>
>> A gentle reminder!
>>
>> On 12-Jul-2023, at 12:45 AM, Pulkit Ojha <pulkit.ojha at niser.ac.in> wrote:
>>
>> Hello,
>> Thank you for your reply.
>>
>> I'm using MESA version r22.05.01. I do not have any desired stopping
>> condition. In my runs, the evolution stops when the donor mass becomes very
>> small and software isn't able to find any consistent solution, stating "dt
>> < min_timestep_limit" or "no. of redo exceeding redo_limit".
>>
>> In the case of "keep_mb_on = .true." , it is true that the timestep
>> becomes so small after first 5k+ steps. I'm attaching binary_history files
>> for both the cases and the same set of initial parameters. I'm also
>> attaching a snippet of terminal output of MESA run with "keep_mb_on =
>> .true." .
>>
>> Truly
>> Pulkit
>>
>> On Tue, Jul 11, 2023 at 12:06 AM Ebraheem Farag <ekfarag at asu.edu> wrote:
>>
>>> Hello Ojha,
>>>
>>> Would you mind sharing which version of MESA you are using?
>>>
>>> What is your desired stopping condition?
>>>
>>> > but if case we do not switch it off i.e. "keep_mb_on = .true."
>>> , evolution isn't stopping at all. After running for 12-13k steps, the mass
>>> transfer rate as well as the masses of both the stars become constant but
>>> the steps are just repeating themselves, even till 40k steps.
>>>
>>> It sounds like your timesteps have become very small. Can you comment on
>>> the difference in time step size between both runs with and without  "keep_mb_on
>>> = .true." ? Are the size of timesteps different between runs? What is
>>> controlling the timesteps in your terminal output in both runs? A short
>>> print out attachment of a snippet from your terminal output would be useful.
>>>
>>> Thanks,
>>> EbF
>>>
>>>
>>> On Sun, Jul 9, 2023 at 12:16 AM Pulkit Ojha via Mesa-users <
>>> mesa-users at lists.mesastar.org> wrote:
>>>
>>>> Dear All,
>>>>
>>>> I'm evolving LMXB models with NS accretors with the specifications
>>>> described in the attached inlist_project file. There is this option of
>>>> "keep_mb_on" for calculating the angular momentum loss due to magnetic
>>>> braking. It decides whether to switch off magnetic braking when the star
>>>> becomes fully convective. If we choose to switch it off, models are running
>>>> fine (stopping after 5-8k timesteps) but if case we do not switch it off
>>>> i.e. "keep_mb_on = .true." , evolution isn't stopping at all. After running
>>>> for 12-13k steps, the mass transfer rate as well as the masses of both the
>>>> stars become constant but the steps are just repeating themselves, even
>>>> till 40k steps.
>>>>
>>>> Can anyone help me to understand what's going on there? Is the software
>>>> getting stuck in some loop or the values of mass transfer become so low
>>>> that they do not appear changing?
>>>> Please find attached the inlist and extras files.
>>>>
>>>> --
>>>> Pulkit Ojha
>>>> 5th Year Integrated MSc.,
>>>> School of Physical Sciences,
>>>> National Institute of Science Education and Research
>>>> <https://urldefense.com/v3/__https://www.niser.ac.in/__;!!IKRxdwAv5BmarQ!dPs0I_NtKoQRnpfmatdrcvp3atkIEBt9Cn2kPKqPIy-lnf3EwwHGazY74Dp19ztpDcIcxkwWPxnEGTAAm7DH0uOT$>
>>>> Bhubaneswar, INDIA
>>>>
>>>>
>>>> _______________________________________________
>>>> mesa-users at lists.mesastar.org
>>>>
>>>> https://urldefense.com/v3/__https://lists.mesastar.org/mailman/listinfo/mesa-users__;!!IKRxdwAv5BmarQ!dPs0I_NtKoQRnpfmatdrcvp3atkIEBt9Cn2kPKqPIy-lnf3EwwHGazY74Dp19ztpDcIcxkwWPxnEGTAAm0IMGyGA$
>>>>
>>>>
>>>>
>>
>> --
>> Pulkit Ojha
>> 5th Year Integrated MSc.,
>> School of Physical Sciences,
>> National Institute of Science Education and Research
>> <https://urldefense.com/v3/__https://www.niser.ac.in/__;!!IKRxdwAv5BmarQ!axq4MEK3mDgQQ8m_6A2KCQLaV38v1CbzecDpFCIx6dqG_MieRMpbp4w1S8wrqNwM7R7SdMByZ-FHfLQNagHMZ72LjA$>
>> Bhubaneswar, INDIA
>>
>> <snippet_MESA__run.png><data_compare_mb_on_and_off.zip>
>>
>>
>> Regards,
>> Abhijnan Kar(He/his/him)
>> BS-MS,IISER Berhampur
>>
>
> Regards,
> Abhijnan Kar(He/his/him)
> BS-MS,IISER Berhampur
>
>
> Regards,
> Abhijnan Kar(He/his/him)
> BS-MS,IISER Berhampur
>
>
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20230718/414e0dab/attachment.htm>