Re: molecular cooling implemented, question about errors in test-models

[Daniel Price]

Hi Ward,

 I just merged a pull request that adds a temperature floor as a feature, exactly with this problem in mind, so if you just pull the latest code into your fork you should get this.

The second error is from getting NaNs in the arrays for whatever reason, again I have replaced this with a more intelligible error message in recent version, so also suggest to pull latest code to help track this one.

Best wishes,

Daniel

On 11 Jul 2021, at 10:47 pm, Ward Homan <ward....@kuleuven.be> wrote:



Hi Daniel, Lionel,

On Friday I finished implementing the molecular cooling routines that my master thesis student was working on, and I'm currently running some test models.

It seems I'm encountering two types of errors, which I would like to discuss:

1) the first is a pretty obvious one: FATAL ERROR! eos: utherm < 0 : u = -4.606E-03

This is probably due to the fact that under certain conditions the gas cools too much, so we probably want to implement a lower limit to the gas temperature. Where would be the best place to set such a lower limit?

2) a less obvious one:  FATAL ERROR! mtree: totmass_node==0  =  0.000E+00

This error occurred when the model was already quite advanced, and I'm not sure what it means. I remember occasionally encountering this problem in the past, but I somehow always managed to circumvent it. Maybe now is the time to have a better look.

Thanks for the input!

Cheerio,

Ward