Calibration Observations

This page for discussion of calibration image regiment to adopt during SV observation period

Daily Flats & Biases

The following sequence to be done each morning or afternoon of Science Verification (up to the observers to get it done at an appropriate time)(exposure times according to the DECal team (D. James):

Sequence Filter ADU target Exposures x time
Bias Any 0 9 x 0s
Y Dome Y 8K 10 x 10 s
z Dome z 8K 20 x 10 s
i Dome i 8K 10 x 22 s
r Dome r 8K 10 x 10 s
Linearity r 16K 5 x 20 s
g Dome g 8K 10 x 30 s
u Dome u 8K 10 x 15 s
Bias Any 0 9 x 0s

Scripts and Special Instructions

Standard pre-night biases and flats script: DECam/sv_prenight_zeros_flats.json
Standard post-night biases and flats script: DECam/sv_postnight_zeros_flats.json

22 Nov: Gary added following scripts which omit u band. These are preferred for (most) nights where no u-band data will be taken:
Standard pre-night biases and flats script: DECam/sv_prenight_bias_grizYdomes.json
Standard post-night biases and flats script: DECam/sv_postnight_bias_grizYdomes.json

The post-night script ends by turning off vsub and inserting the block filter.

You load these scripts from the Exposure Control section of the Observer Console GUI. Click on Load Exposure Script, click on Select DECam Script, select the script file and click Submit

Dark frames

Collect the following sequence of dark observations once per week:

Sequence Filter ADU target Exposures x time
Dark15 (dark) 0 1 x 15s
Dark100 (dark) 0 3 x 100s
Dark300 (dark) 0 1 x 300s

Scripts and Special Instructions

Standard dark sequence script: DECam/sv_dark_sequence.json

Twilight frames

Twilight frames have been forbidden as of 10 Nov to prevent damage to the array. Don't do any!

Attempt to obtain 5-10 twilight exposures in one filter per evening so that each filter is seen roughly once per week. The usual issues with making successful twilight flats:

Sequence Filter ADU target Exposures x time
Twilight (1 filter per night) 8k 5-10 x (Variable) s

Here is a Guide for Twilight Flats based on sky count rates for the first week of SV.

Scripts and Special Instructions

  • You will need to adjust exposures by hand to keep the counts per exposure near the target of 8k ADU per pixel.
  • Turn tracking on (so stars do not smear) but move the telescope by a few arcminutes between each exposure (so stars are shifted on the array).
  • Ideally, commence twilights when the desired counts are obtained in about 1 sec, then continue until it takes >20s to reach target counts. Then it is time to start taking photometric standards (see below).
  • Telescope can be anywhere, e.g. zenith. Of course further west = brighter sky at twilight.
  • A student project will be to analyze the twilight counts vs filter, sun altitude, and telescope pointing to obtain a model that SISPI can use to automatically select exposure times. But this will not be a requirement of SV.
  • Dara may have some twilight requests from the community SV projects, hopefully these can be obtained in the AM. DES SV will only request standards in the AM, not twilight exposures.

No standard script is available

DECal Flats

We hope to use DECal in two modes. One is "full scan" in which DECal scans the full passband for every filter at 1-2 nm bandwidth (John Marriner notes that SN requirement is anything <5 nm). The other is a "fast scan" mode in which DECal is set to wider bandwidth (10 nm) so that flats can be taken in daytime, and also only looks at red / center/ blue part of each filter passband. This is just to check consistency/stability of color response.

Full scans should be done 2 times total in Commissioning and SV observing. We await DECal week results to learn whether this can be done in daytime, the time it takes to complete, etc. During normal observation this would be repeated about once per month, using cloudy nights. Ideally we would also take full scans on the day that DECam CCDs are operated at warmer-than-usual temperatures.

Fast scans can be done at least once per week, and also on the night of warmer CCDs. Exposure times and practicality of this mode await results of DECal week (might want e.g. to take lamp off / lamps on / lamps off sequences to remove ambient light).

Sequence Filter ADU target Exposures x time DECal setting
DECal Fast u u 8k 5 x (TBD) s red/center/blue parts of filter band
DECal Fast g g 8k 5 x (TBD) s red/center/blue parts of filter band
DECal Fast r r 8k 5 x (TBD) s red/center/blue parts of filter band
DECal Fast i i 8k 5 x (TBD) s red/center/blue parts of filter band
DECal Fast z z 8k 5 x (TBD) s red/center/blue parts of filter band
DECal Fast Y Y 8k 5 x (TBD) s red/center/blue parts of filter band

Scripts and Special Instructions

For the time being a DECal expert is required to run a DECal scan. No standard script is available.

Photometric standard fields

Douglas has proposed an extensive list of fields with useful photometric standards. Gary proposes that we limit our observations to these five that he has listed among the top 2 priority levels:

# StdFieldName   RAHMS    DECDMS    Equinox  expT_u  expT_g  expT_r  expT_i  expT_z   expT_Y   CatName          
SDSSJ0000-0000  00:00:00 +00:00:00 2000.0    30.     15.     15.     15.     15.      20.     Stripe82Marriner
SDSSJ0200-0000  02:00:00 +00:00:00 2000.0    30.     15.     15.     15.     15.      20.     Stripe82Marriner
SDSSJ0320-0000  03:20:00 +00:00:00 2000.0    30.     15.     15.     15.     15.      20.     Stripe82Marriner
MaxVis          06:30:00 -58:45:00 2000.0    30.     15.     15.     15.     15.      20.     DESYear1MaxVis 
C26202/HST      03:32:30 -27:51:05 2000.0    30.     10.     10.     10.     10.      20.     HSTCalSpec   

Scripts and Special Instructions

There would be 3 standards sessions per photometric night, ideally. The goal of these observations is to assess the stability of the zeropoints and color & airmass terms across each evening and as a function of the conditions viewed on Rasicam. We therefore want to observe fields with a maximal range of airmass with 1<X<2 at the beginning, middle, and end of each night. If we take a single exposure in grizY at each field, this takes 4 minutes, so 12 minutes to do 3 fields x 1 exposure x 5 filters. I suggest that we limit u-band tests to just one or 2 nights, since DES is not in the market for u-band photometry. I believe a single exposure should be sufficient since there are 62 CCDs to choose from. We will have many other opportunities to test short-term reproducibility of photometry, we do not need to do this in standard fields.

Exposure scripts are available for each of these fields/sequences. The scripts use the coordinates given in the table above to slew the telescope for the first exposure in the sequence. No offset (dither) is used for the consecutive exposures. Given the current uncertainties in the telescope pointing repeat observations of the same star field can be offset by several arc min.

22 Nov: Gary made scripts to skip u-band, which are preferred unless u data is taken that night.

Field Script Script w/o u band
SDSSJ0000-0000 DES/sv_sdssj0000_0000.json DES/sv_sdssj0000_grizY.json
SDSSJ0200-0000 DES/sv_sdssj0200_0000.json DES/sv_sdssj0200_grizY.json
SDSSJ0320-0000 DES/sv_sdssj0320_0000.json DES/sv_sdssj0320_grizY.json
MaxVis DES/sv_maxvis.json DES/sv_maxvis_grizY.json
C26202/HST DES/sv_c26202hst.json DES/sv_c26202_grizY.json
  1. Evening twilight:
    Observe SDSS-0000 and C26202 field. Third field is SDSS-0320 if it is at X<2.1, otherwise SDSS-0200.
  2. Middle of night (01:30 local time):
    Observe SDSS-0000 and C26202 fields, then one of MaxVis, SDSS-0320, or SDSS-0200. Vary choice of third field each evening so that we can tie all these fields together. Take an extra i-band exposure of 100s duration at one of the SDSS fields, to produce a deeper estimate of the star flat.
  3. Morning twilight:
    Observe MaxVis, C26202, and SDSS-0320

Here is an airmass plot for these fields for the midpoint of SV: