[Mesa-users] [EXT] PPISN Model Stalling in MESA-r24.03.1

Fri Sep 19 19:49:40 UTC 2025

Hi Mathieu,
Thank you for your helpful suggestions! To answer your questions:
                • For the restart at x950, I changed only gold_tol_max_residual3 from 1d-3 (default value) to 5d-3 to test the tolerance value.
                • In my grid, the 95 and 115 Msun models produced pulses successfully, but the intermediate masses (100, 105, and 110 Msun) all showed the same stalling behavior as the 100 Msun run.
                • I also turned off the stopping condition at the first pulse to allow the runs to proceed through all PPI events, but the 100, 105, and 110 Msun models still crashed before the first pulse.

As an update, adjusting gold_tol_max_residual3 to 5d-3 and 5d-2 allowed the 100 Msun model to run, though I understand that the approach could be better. I followed your suggestion and worked through the MESA Summer School debugging task -- it was very helpful and gave me a clearer idea of where to focus. I'll need to dig further to identify the root cause.

Thanks again for your suggestions!

Best,
Trang N. Huynh

From: Mesa-users <mesa-users-bounces at lists.mesastar.org> on behalf of Mathieu Renzo via Mesa-users <mesa-users at lists.mesastar.org>
Date: Thursday, September 18, 2025 at 12:47 PM
To: mesa-users at lists.mesastar.org <mesa-users at lists.mesastar.org>
Subject: Re: [Mesa-users] [EXT] PPISN Model Stalling in MESA-r24.03.1

Hi Trang,

From your terminal output it looks like you're restarting from photo x950 and at model 20979 (19 steps later) the issue starts appearing. Are you changing anything in the inlist at this restart?
Have you tried restarting from a bit earlier? Have you tried slightly different masses? Do you have any models that worked to compare this crashing model to? (Maybe another mass in your grid?)

From the masses, I suspect this is at the first pulse? If so, one thing to try is to change a bit the thresholds at which you switch from hydrostatic equilibrium to the HLLC solver (beware that you don't want to do that too early and "freeze" your simulation, nor too late and miss the beginning of the dynamical phase). This can be done with x_ctrl(1) in the test case setup, for example.

As a last resort, you may need to figure out exactly what residual is exceeding the tolerances: as you found, just making the tolerances larger, although it can sometimes help bypass small issues, it doesn't really solve the issues. Maybe this exercise from this year MESA Summer school <https://mesa-leuven.4d-star.org/tutorials/thursday/morning-session/> can be of guidance: it's not on PPISN, but retries because of max residuals are common in massive star evolution. The tasks of that exercise down to Task 5 should guide you to find what is the most problematic residual. Then, how to address that is something the MESA summer school exercise will not necessarily help with, but you will know more about why your model crash and hopefully that helps solving the issue.

Bests,
Mathieu
On 9/18/25 06:35, Trang Huynh via Mesa-users wrote:

External Email

________________________________
Dear all,

I am currently simulating a pulsational pair-instability supernova (PPISN) model using MESA-r24.03.1 and mesasdk version 24.7.1. The setup is a non-rotating, 100 M⊙ with  Z = 0 . My inlist and run_star_extras.f90 files are based on the ppisn test-suite included with this MESA release.

The problem is my model struggles to enter the pair-instability regime and consistently stalls during O-burning phase. I attempted increasing gold_tol_max_residual3 (from 1d-3 → 5d-3 → 5d-2), but the issue persists. Below is a snippet from the terminal output of the last retries:

Setting energy conservation error to zero until hydro is turned on for the first time
check gamma integral 3.6349043014391437E-002
max_residual > tol_max_residual 26835 5.1008342853913410D-03 5.0000000000000001D-03
solver rejected trial model
s% model_number 26835
s% solver_call_number 33067
nz 2861
s% num_retries 6530
dt 4.1498956078510518D-01
log dt/secyer -7.8810667950730071D+00
retry: max resid -- give up in solver 26835
residual_norm > tol_residual_norm 26835 1.0776741982634678D-06 9.9999999999999995D-07
I would greatly appreciate any suggestions or best practices to help the model progress into the pair-instability phase. To facilitate feedback, I’ve attached the zip files with relevant inlist files, run_star_extras.f90, full terminal output, and mesa build.log.
Thank you in advance for your time and help!
Best,
Trang N. Huynh (she/her)
PhD Student
Department of Physics and Astronomy
Louisiana State University
thuyn27 at lsu.edu<mailto:thuyn27 at lsu.edu>

_______________________________________________

mesa-users at lists.mesastar.org<mailto:mesa-users at lists.mesastar.org>

https://lists.mesastar.org/mailman/listinfo/mesa-users

--
Mathieu Renzo<https://www.as.arizona.edu/~mrenzo/>
Assistant Professor
University of Arizona - Steward Observatory
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20250919/b918873b/attachment.htm>