[Mesa-users] runaway mass transfer in WD-WD binary

[Kareem El-Badry]

Thanks Jim, I will play with this.

Just clarify, is your donor here initialized as a main sequence model, or
did you load a white dwarf model? I'm guessing the latter, because I think
a normal 0.2 Msun star would be too large to fit into a ~1 hour orbit. If
so, any chance you have the model handy for testing.

I have had success evolving "normal" CVs through the period minimum (at ~1
hour periods) with this setup -- I only run into problems when the donor is
a WD / proto-WD.

Thanks,
Kareem

On Wed, Sep 2, 2020 at 5:12 PM Jim Fuller <jfuller at caltech.edu> wrote:

> Here is an inlist from a binary model I made that evolved stably through
> the period minimum. I'm not sure which combination of different parameters
> allowed MESA to get through it, but maybe it will help you out.
> -Jim
>
> On Wed, Sep 2, 2020 at 4:47 PM Kareem El-Badry <kelbadry at berkeley.edu>
> wrote:
>
>> Hi Jim,
>>
>> Thanks for the suggestion. You're right that the papers from which I got
>> my assumption that mass transfer should be stable assumed it was
>> conservative. My reasoning for using fully non-conservative mass transfer
>> was that novae should periodically blow the accreted material off the
>> surface of the WD. I'm not sure whether this occurs smoothly enough in
>> practice for non-conservative mass transfer to be a good approximation.
>>
>> In any case, I did try running with mass_transfer_beta = 0, but I get
>> basically the same behavior, so I don't think this is the root of this
>> particular issue (at least, not the only one).
>>
>>
>> Best,
>> Kareem
>>
>> On Wed, Sep 2, 2020 at 9:28 AM Jim Fuller <jfuller at caltech.edu> wrote:
>>
>>> Hi Kareem,
>>>
>>> It sounds like the mass transfer is becoming unstable. It should be
>>> stable for conservative mass transfer, but you're using non-conservative
>>> mass transfer, and I think that usually is unstable for white dwarfs. Have
>>> you tried with conservative mass transfer, i.e., mass transfer beta = 0 ?
>>>
>>> -Jim
>>>
>>> On Wed, Sep 2, 2020 at 7:00 AM Kareem El-Badry <kelbadry at berkeley.edu>
>>> wrote:
>>>
>>>> Hello MESA users,
>>>>
>>>> I'm trying to simulate mass transfer from a low-mass He WD (0.13 Msun)
>>>> onto a point mass that represents a massive WD (1.1 Msun). Mass transfer
>>>> should be stable and the period should increase once mass transfer begins
>>>> [I think] because (a) the orbit expands when the He WD loses mass, and (b)
>>>> the radius of the He WD increases when it loses mass, reestablishing
>>>> contact with the Roche lobe at a longer period.
>>>>
>>>> However, in the MESA calculations, Mdot gets very large as soon as the
>>>> He WD overflows its Roche lobe, and the calculation grinds to a halt. I've
>>>> attached inlists and a model.
>>>>
>>>> I am using version 12778.
>>>>
>>>> Some things I have tried:
>>>> - decreasing varcontrol_target and mesh_delta_coeff
>>>> - decreasing fm and fr
>>>> - setting set_to_this_tau_factor = 30
>>>> - turning off diffusion
>>>>
>>>>
>>>> But none of these had an appreciable effect. Does anyone have relevant
>>>> experience or an idea what the problem might be?
>>>>
>>>> Thanks,
>>>> Kareem
>>>>
>>>> _______________________________________________
>>>> mesa-users at lists.mesastar.org
>>>> https://lists.mesastar.org/mailman/listinfo/mesa-users
>>>>
>>>> _______________________________________________
>> mesa-users at lists.mesastar.org
>> https://lists.mesastar.org/mailman/listinfo/mesa-users
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20200902/7036ed19/attachment.htm>