[Mesa-users] YNT: Evolution of binaries, problem

[Josiah Schwab]

Hi Gökhan,

> Thanks for replying Josiah.
>
> I have tried the parameter you suggested but MESA still stucks on same
> point. I even had tried several values from 1d-7 to 1d-10.

Let me reiterate:

When the MESA timestep gets small, it is important to think a little
about what is happening physically (i.e. when is MESA getting stuck).
That helps in thinking through what might be going wrong.

You cannot defer until a model is "finished" to think about what is
physically happening.

Using MESA is like driving a car.  If you only start paying attention
when you reach the destination you may never reach the destination :)

Use the terminal output or plots to "watch" what is happening to the
binary.  For example, what are the star masses and the mdot?


Running your inlists, I see that with the adjusted version of
starting_mdot mass transfer successfully starts.  So then adjusting it
further unlikely to have any effect.

Now the stall appears to occur as the mass transfer rate begins to fall
and the system tries to come out of contact.  Again, one can look at the
controls associated with the roche_lobe scheme.  Since we see this is
happening when mdot is getting small, the controls setting the min mdot
(default 1d-16), may be of interest.

So I suggest you explore the effect of increasing this value, e.g.,

  roche_min_mdot = 1d-14


Let me also note that one doesn't usually have to fiddle in this way
when using the default Ritter scheme, since mass transfer turns on/off
smoothly as a function of RL filling factor.

> I don't think my error is related with mass transfer but a systematic
> error. Because when MESA hits the point where it stuck first time, I
> could be able to catch an error on terminal as follows

Indeed.  That simply indicates MESA is having trouble finding the next
model, and so is reducing the timestep.  It is not an error, per se.

You problems are definitely related to mass transfer.  When running
binary models, keep a close eye on quantities like mdot and the roche
lobe filling factor.

Josiah