0 votes


I'm trying to get photometry from archival DECam single-epoch image cutouts, in order to calibrate some spectra (the calibration is time-sensitive, otherwise I would just pull numbers from the coadd catalogues). I'm using proctype=Resampled and the images are specifically from the DES supernova survey -- object='DES supernova hex SN-X1 tiling 20'. The BUNIT in the cutouts is ADU, so I'm trying to convert flux values to magnitudes with the equation:

m = m0 - 2.5*log(flux)

where flux = gain * counts [ADU] / exptime.

The issue I'm having is with the gain -- each header lists two gain values, one for each of amplifier A and B, and I can't tell which spaxel belongs to which amplifier. How do I find out which gain to use? Here's an example of the relevant part of the header:

EXCLUDED= '' / DECam components not used for this frame                         
HEX     =                  339 / DES Hex number                                 
TILING  =                   20 / DES Tiling number
DETECTOR= 'S3-40_124750-19-1'  / Detector Identifier                            
CCDNUM  =                    4 / CCD number                                     
DETPOS  = 'S25     '           / detector position ID                           
GAINA   =                4.061 / [electrons/adu] Gain for amp A                 
RDNOISEA=              6.05089 / [electrons] Read noise for amp A               
SATURATA=              42059.0 / [adu] Saturation for amp A                     
GAINB   =                4.249 / [electrons/adu] Gain for amp B                 
RDNOISEB=             6.080319 / [electrons] Read noise for amp B               
SATURATB=              42605.0 / [adu] Saturation for amp B  



asked Mar 22 by ddunne (160 points) | 50 views

2 Answers

0 votes
Hello Delaney,

The basic answer about gain is that since the data is flat fielded with dome flats and star flats the individual amplifier gains are not relevant.  The effective gain is some average over all amplifiers because the flat fielding is done with an mean correction of 1.

However, it seems to me that the simpler answer is that the Community Pipeline (CP) computes a zero point for you and is in the header - MAGZERO or MAGZPT; respectively in ADU and per second.  This is based on a reference catalog which may be Pan-STARRS if dec>-30 or Gaia G or (for very early processing on an approximation using USNO-B1).

Let me know if you want more information.

Frank Valdes, pipeline scientist
answered Mar 22 by Frank Valdes
0 votes
I may not have been completely clear in my answer.  The equation you want is

m = MAGZERO - 2.5 log10 (ADU)

where MAGZERO is from the header and ADU is the instrumental counts over whatever aperture.

answered Mar 22 by Frank Valdes
Hi Frank,

That makes sense -- thank you!


251 questions

252 answers


1,708 users

Welcome to Data Lab Help Desk, where you can ask questions and receive answers from other members of the community.