Absurd Bayesian-like estimates

[MouadGF]

Hi,

I need help with this SED fit I'm trying to do. When using the pcigale.ini (file below) with the photometric data (also below) I get absurdly big and negative bayesian-like estimates of dust luminosity and star formation rates (see results below). The best fit values are normal. I tried changing the dust emission model (between casey2012 and dl2014), but the issue persists.

Thank you in advance

pcigale.txt
phototmetry data.txt
results.txt

[MouadGF]

After debugging, I found the parameter that breaks this: 'E_BV_lines' needs to stay low or the code starts coming up with negative star forming.

[mboquien]

Whe E_BV_lines is extremely high, you can indeed run into purely numerical artifacts. I would not recommend going much beyond 2 magnitudes, which is already very high.