The
Basics of Hectochelle Reductions with HSRED
1. The basics
- Caveats:
Hectochelle reductions are only offered on a very trial basis - all
reductions should be examined carefully to check for reduction fidelity
- I have only tested this on a limited set of data, and it works to
check for photons, but I would not trust all aspects of the reductions
reobustly. The reduction procedure is similar to hectospec, but
with none of the bells and whistles.
- Known Problems:
- The flat field seems poorly defined on the edges of
the chips - this causes some problems in the end with the spectral
shape there
- The wavelength solution appears to do poorly
redward of ~ 2290 \AA for the RV31 grating. All users should
check for fidelity here
- I have not tested for the optimal extraction
parameters - these may be changed in the future
- Initial Preprocessing:
- The preprocessing (creation of master biases, flats, etc)
require input lists. You should start IDL in the directory
containing all of your raw data for a given night. Create the
initial lists with:
- This script creates the lists
directory and the needed lists to combine each of the
observations.
- bias.list
- Contains all of the bias observations to be used in
the reductions. It is vital
that you check this file by hand to be sure that no non-biases are
included. Often non-biases are coded as biases, and you
don't want to be include these. Also, you don't want to to
include biases that are contaminated by light in the enclosure (these
will have more counts than a typical bias).
- arc.list
- Lists all of the arc files in the directory. Again,
these should be check to ensure that none of the exposure are
contaminated by light in the enclosure and that none of the arcs are
under or over exposed.
- dflat.list
- Contains dome flat files. Be sure none are saturated
- sflat.list
- Contains twilight flat files. Again, check to be
sure that these flats are not under- or over- exposed.
- dark.list
- cal.list - SINCE BATCH REDUCTIONS ARE NOT
SUPOPORTED WITH CHELLE FOR NOW, THIS IS NOT IMPORTANT
- Ignore this file for now - it is used for reductions
in batch mode, and will be covered when we get there.
- Once the lists are properly prepared, you can complete the
initial processing of the calibration files. The proper task is hs_calibproc which has several
possible modes.
- If you want to process the bias, arcs, dome
flats, and sky flats, you could use the command
- IDL > hs_calibproc, /dobias, /doarc,
/dodome, /dosky,
rerun=100
- If, however, you do not have all of these files to
process (you may not have sky flats for example), you can process
everything else with:
- IDL >
hs_calibproc, /dobias, /doarc, /dodome, rerun=100
- This command makes the master version of each
calibration type. All of the
calibration data are stored in the calibration/XXXX/
directory where XXXX is the four digit rerun number.
- Do the wavelength solution.
- Say you want to fit the master arc
- IDL>file
= 'calibration/0100/wset.fits'
- IDL>hc_dowave,
file, rerun=100, /debug
- This is a multiphased process
- 1. The arc is extracted and the first fiber on the
first chip is plotted. You need to identify several lines to generate
an inital solution - click once to zoom on a line, click again to
centroid. When you are done fitting arc lines type "STOP'
- 2. A 4th order polynomial is used to fit your lines
and identify others. You will see a plot of residuals from the
fit versus pixel number, identify bad points by clicking and hitting
return. When you are done type STOP. A higher order function is
fit to the new data and the residuals are shown again - you can iterate
if you need to. when you get a new residual plot that looks good,
type DONE (you need to have typed STOP already and be prompt for
done). This is clunky - I know - but it was fast and I have very
limited time.
- 3. Cross correlations are used to refine the fit for
the first fiber and then finds the solution for the second these are
plotted for you (white is data, red is a model arcspectrum).
- 4. The solution is traced across the chip
- 5. Repeat for the second chip
- Extraction:
- You are now ready to run the main extraction
routine. A couple of bookkeeping notes: if you have a
sflat.fits file in the calibration/XXXX/
directory, both the skyflat and the domeflats will be used for the
calibration of your data. If sflat is not included, then only the
domeflat is used. The full documentation for hs_extract, the
workhorse for the extraction, can be found in the source code, but the
general operations are found here. First, you should decide which
observations to extract and coadd. These should all be the same
configuration, obviously.
- IDL>
files = findfile('bootes04061_3*fits')
- (NOTE - you can add
/docosmic to any of
these calls
to enable cosmic ray rejection on EACH exposure):
- Standard extraction:
- IDL>
chelle_extract, files, rerun=100, outroot='bootes3'
- All of these will create the reduction/XXXX/
directory (XXXX is again the four digit rerun number) and the output
will data products will be saved there. There are a number of
files, but those of most interest are the :
- spObs file - these are the extraction for each exposure
- not fluxed
<>spChelle file - these are the coadded spectra - fluxed if
requested
- To examine
the data quickly:
- IDL>
file = 'reduction/0100/spChelle-test.fits'
- IDL>chelle_page, file, nsmooth=2
- Data model:
- spObs:
- ext[0] - wavelength / Angstroms
- ext[1] - flux
- ext[2] - inverse variance
- ext[3] - mask
- ext[4] - or mask
- ext[5] - plugmap - contains targeting information
- spChelle:
- ext[0] - wavelength / Angstroms
- ext[1] - flux
- ext[2] - inverse variance
- ext[3] - mask
- ext[4] - or mask
- ext[5] - plugmap - contains targeting information