Re: [mesa-users] negative delta nu!?!

[Warrick B. <wb...@bi...>]

Hi,

So I looked into the issue with the "failed to find required l=0 modes" 
message to try to make sense of what's going on.  I think the segment of 
code that causes the error message to be printed is unchanged from my 
original report, and I've again concluded that it isn't a cause for 
concern, except that the message doesn't correspond to what has happened 
in the code.

Actually, looking at the code, I'd say the line that prints the message 
"failed to find required l=0 modes" is almost unnecessary.  It relies on 
the ouptut of get_radial, which, in almost each case, will print its own 
message about why the code didn't proceed to compute the higher-degree 
modes.  The one case that currently doesn't print an error message is if 
the oscillation code itself fails.  Here are lines 637--646, inside 
get_radial:

     call get_one_el_info(s, 0, &
        nu_lower_factor*l0_obs(1), &
        nu_upper_factor*l0_obs(nl0), &
        iscan_factor_l0*nl0, 1, nl0, store_model, &
        code, ierr)
     if (ierr /= 0) then
        !write(*,'(a65,i6)') 'failed in oscillation code', s% model_number
        !stop
        return
     end if

Any such early return means the function will return .false. and MESA 
won't compute the non-radial mode frequencies.  I'm not sure why the error 
message "failed in oscillation code" is commented out: I'd find that quite 
useful!

So, for extras_support.f in r9575, I suggest:
- commenting line 374 and
- uncommenting line 643 (maybe with mention that it's in the radial mode
   calculation).

For other angular degrees, MESA just prints an error message if the 
oscillation code returns with ierr /= 0, so I don't think any changes are 
warranted.

Cheers,
Warrick



------------
Warrick Ball
Postdoc, School of Physics and Astronomy
University of Birmingham, Edgbaston, Birmingham B15 2TT
wb...@bi...
+44 (0)121 414 4552

On Wed, 10 May 2017, Warrick Ball wrote:

> Hi Jean,
>
> On the first issue, I started your run a few minutes ago but it looks like 
> it'll take a while to finish.  Out of curiosity, which revision are you 
> using?  Your GYRE input gave me a read failure in r9575 (no namelist object 
> "diff_scheme") so I've just used the default for now.
>
> Regarding the "failed to find required l=0 modes", I looked into this over 
> two years ago:
>
> https://sourceforge.net/p/mesa/mailman/message/33028229/
>
> Ignore my proposed fix: that wasn't the solution.  It seems Rich started 
> looking into where the code changed but I couldn't find anything further on 
> the mailing lists.  I've cc'ed Rich, in case he can recall if he made any 
> progress.
>
> As far as I know, it's nothing to worry about.  But now that you've brought 
> it up again, it's inspired me to have another close look at why the message 
> is issued and what should be happening instead.  Hopefully more as the day 
> progresses here.
>
> Cheers,
> Warrick
>
> PS:  I noticed your inlist has the line
>
>      l0_n_obs(:) = 6 ! the observed radial orders (ignored if < 0)
>
> AFAIK, this should be the observed radial orders of the radial modes, which 
> presumably aren't all 6!  (You can choose to let MESA infer the radial orders 
> of the modes: see lines 660--686 in 
> $MESA_DIR/star/astero/work/astero_support.f.)
>
>
>
> ------------
> Warrick Ball
> Postdoc, School of Physics and Astronomy
> University of Birmingham, Edgbaston, Birmingham B15 2TT
> wb...@bi...
> +44 (0)121 414 4552
>
> On Tue, 9 May 2017, Jean McKeever wrote:
>
>>  Here are the inlists I used.
>>
>>  2017-05-08 17:22 GMT-06:00 Jean McKeever <je...@nm...>:
>>        Hello,
>>
>>  I am running some models using astero and have ran into an interesting
>>  conundrum. As soon as astero starts computing individual frequencies, the
>>  delta_nu value, both printed on the graph display and
>>  written to the history file, becomes negative.  I am using gyre to compute
>>  only radial modes. Has anyone else encountered this or something similar?
>>
>>  On another note I am constantly getting  a "failed to find required l=0
>>  modes" message, however modes appear on the echelle diagram. I have double
>>  checked my gyre input files are in the right units and
>>  cover the range of observed modes. Is this something to worry about?
>>
>>
>>  Jean McKeever
>>
>>
>> 
>> 
>