Differences between EISPAC width and IDL width

[andyto1234]

Hi,

I am wondering if the width between EISPAC and IDL are defined differently? I was comparing the width obtained through IDL and EISPAC, and it seems that the IDL width is much higher and over a much wider range. Below is a screenshot of the width, using a 2007-Oct-17 file, vmin=0, vmax=0.045. Thanks!

[yjzhu-solar]

Hi Andy,

There could be several reasons, from my previous experience in working on EIS line widths, except for a real difference between the EISPAC and IDL widths:

1. Are all widths in sigma (standard deviation), 1/e, or FWHM? I remember EISPAC uses sigma widths as output, while eis_auto_fit and eis_get_fitdata in IDL return FWHM. To convert from sigma to FWHM, you have to multiply a factor of 2.355 ($\sqrt{8\ln2}$)
2. Did you remove the instrumental broadening? eis_get_fitdata will remove the instrumental width; however, I remember EISPAC does not do that, although the instrumental width (in FWHM!) is restored in the slit_width metadata.
3. Did you remove the thermal broadening?

I hope this will help!

[MJWeberg]

Thanks for the detailed comments and questions, @yjzhu-solar!

Sorry for not following up on this sooner, @andyto1234. I keep meaning to take another look that the various other EIS fitting routines in IDL / SSW. For what it is worth, we did an analysis a while back comparing the fit parameters from IDL and Python/EISPAC when using the fitting templates (which were originally designed for use in SSW). We found, in general, that the parameters matched to within 1-2%, with the largest differences in regions with poor signal-to-noise. However, we have not done a comparison with the eis_auto_fit routine in IDL, so I am less certain of the differences there.