Fitting

Q1:

I tried to fit but got an error message ``Fit did not converge'' even though it's a simple spectral region.

A1:

Basically, it wasn't simple enough. Please see the explanation in §3.4. The fitting algorithm currently used works especially poorly on low-count data or regions where the continuum is curved. Also, to avoid degeneracies, the maximum line width is 100 eV.

Q2:

I tried to fit but got an error message saying there were too many peaks? Why can't profit fit more than 20 peaks at a time?

A2:

This is an attempt to avoid profit crashing. Currently, on some machines, fitting large numbers of lines leads to core dumps. We have found that limiting the lines to be fit to 20 or less significantly reduces this possibility.

Q3:

I tried to fit but profit crashed!

A3:

As noted in A2 above and in §3.4, we are trying to address this problem but given the many different machines and operating systems that can run profit the issue is complex. To avoid this, we recommend that you don't try to fit large regions until we can address the issue definitivity. However, please let us know via the Feedback form on the profit website if you find a file and spectral region where you can force a core dump consistently; we are searching for examples to test fixes against.

Q4:

I'd like to do fits that include the detector response (.rmf) and effective area files (.arf).

A4:

May we recommend using either XSPEC (http://heasarc.gsfc.nasa.gov/docs/xanadu/xspec/) or Sherpa (http://asc.harvard.edu/sherpa/), both packages that do this and far more. We are in close contact with the developers of both packages, and may include more features like this in the future. At the moment, however, profit should be used an exploratory tool.