LSST Exposure Calculations
The online exposure calculator is here. This page has useful formulae for normalization.
Sources
To get conversions from AB magnitude to total flux in electrons/sec, chose a point source with AB = 24.0 (this is the default). Change the exposure time from 15s to 1s. Change the "analysis method" to a 100" software aperture and set the "optimal filter" to not weighted. These are not default settings and should give 100% of the signal in the aperture.
The "Total signal per exposure" gives the flux normalization as a function of the selected band, spectral type, and redshift:
Band 
Type 
z 
Flux24/15s 
Flux24/1s 
r 
Elliptical 
0.0 
807.37 
53.82 


0.5 
821.66 
54.77 


1.0 
813.30 
54.55 

Spiral 
0.0 
813.35 
54.22 


0.5 
806.89 
53.79 


1.0 
809.30 
53.95 
i 
Elliptical 
0.0 
710.93 
47.39 


0.5 
710.62 
47.37 


1.0 
711.29 
47.41 

Spiral 
0.0 
711.50 
47.43 


0.5 
713.93 
47.59 


1.0 
714.35 
47.62 
Since these results are fairly insensitive to type and redshift, adopt fiducial values of 810 (r) and 711 (i). The scaling with AB is 10^{0.4(AB24)}.
Noise
Use an unweighted aperture with diameter 2*sqrt(pi/4) = 2.25676 so that the aperture size is exactly 100 pixels, then with default parameters in iband:
"Background noise" = 319.48 ADU in 100 pix which implies 319.48^{2}/100 = 1020 ADU/pix of sky. Compare with 711*10^{0.4(20.224)} = 942. Results do not agree exactly since the sky and galaxies different spectra.
"Dark current noise" = 1.5 ADU in 100 pix which implies 1.5^{2}/100 = 0.0225 ADU/pix of dark current in a 15s exposure = 0.0015 ADU/s/pix.
 "Readout noise" = 50 ADU in 100 pix which implies 5 ADU/pix.
Sky
Run with different bands selected to get the mag/arcsec^{2} and BG noise columns below, then calculate:
ADU/pix = (BG noise)^{2}/100 = sky level, use sqrt(sky/n) to get sky noise in average of n exposures
Flux24 = (ADU/pix) 10^{+0.4(mag24)}/(0.2 arcsec/pix)^{2}, use to calculate sky level for different assumed surface brightness
 Visit Noise = sqrt(sky/2)
 Stack Noise = sqrt(sky/(2*nvisits))
 Last 2 columns from LSST Book Tables 1.1 and 3.2
Band 
mag/arcsec^{2} 
BG Noise /100pix 
ADU/pix 
Flux24 
Visit Noise 
Stack Noise 
Visits/Stack 
mag/arcsec^{2}(nom) 
u 
22.30521 
92.04 
84.7136 
444.608 
6.50821 
0.77788 
70 
21.8 
g 
21.71681 
197.69 
390.813 
1192.98 
13.9788 
1.39788 
100 
22.0 
r 
21.20972 
215.86 
465.955 
891.605 
15.2636 
1.00645 
230 
21.3 
i 
20.21425 
319.48 
1020.67 
780.778 
22.5906 
1.48958 
230 
20.0 
z 
19.03285 
446.95 
1997.64 
514.752 
31.6041 
2.23475 
200 
19.1 
y 
17.92805 
329.74 
1087.28 
280.171 
23.3161 
1.64870 
200 
17.5 
Signal to Noise
Since the online ETC only supports Gaussian profiles specified by the fwhm, use r(half) = sqrt(log(4))*sigma, fwhm = 2*sqrt(2*log(2))*sigma, which gives fwhm = 2*r(half) (!). Assume elliptical galaxy spectrum.
Source 
i(AB) 
ADU/exp 
r(half) 
fwhm 
SN/exp 
SN/stack 
1 
26.0 
112.67 
0.75" 
1.5" 
0.66 
13.05 
2 
26.5 
71.09 
0.50" 
1.0" 
0.42 
8.24 
DES / LSST Cross Calibration
DES has measured their "zero points" using the C26202 standard star, which has spectral type G, in the science verification data (from an email from Klaus H):
Band 
Zpt(e) 
Zpt(e)/Expected 
g 
591 
1.00 
r 
635 
1.15 
i 
638 
1.11 
z 
477 
1.11 
Y 
206 
1.33 
The zeropoint is defined as the number of electrons/sec collected from an AB=20 mag source.