Recalibration

All right, so I have first day of the the re-calibrated 2004 data from Jason… and I find the results very encouraging so far…

Question for Jason:

Why does it appear that there are "bad pixels" or something going on with the re-calibrated data that didn't happen with the original data?

badpixels.gif

First,

the good news:

The two calibrations overlap almost perfectly

directcompare.gif

Here's a closer look at the two calibrations (they are so close that the mostly appear on top of each other, except of course where the flux goes to zero…

closedirectcompare.gif

Now,

the bad news:

I removed what I'm calling the "bad pixel" data points and ran our iodine calibration analysis.

When I run the data on the 2 iodine exposures that were taken on the first day on order 67 we get that our iodine calibration gives a really close match… except where it matters: right on the FeII 1608 and 1611 lines.

Here's a plot of the iodine calibration results: the "experimental_67.3-1.ascii" is the newly re-calibrated data, the "output_67.3-1.ascii" is the original data.

whew.gif

Finally,

The much better news:

Seeing how close the two files were, I decided to look into what was happening. Since removing the "bad pixel" data points could skew the iodine calibration, I decided to remove exactly the same pixels from the original data and re-run the analysis.

The result is the "experimental_67.3-9.ascii" graph above, and to give a closer view:

closeup.gif

They all agree almost perfectly far away from the bad pixels, and when the same pixels are removed, they agree well within error.

CR removal?

I have plotted 3 things in the following plot.

1. In RED the Keck data from the original reduction.
2. In GREEN the convolved and shifted FTS iodine lines
3. In BLUE are the flux regions that are completely removed from the new reduction.

cr-remove.gif

It looks as though the original reduction finds accurate data in the regions that are now being removed because it looks like the Keck data is correctly seeing the iodine lines that "ought" to appear there.

Also, added for completeness because it was mentioned in an email:

The region around 5308 angstroms:

cr-2306-2311.gif

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-Share Alike 2.5 License.