The SDSS Filter Set


The filter system was defined in Fukigita et al, 1996.

The curves we exepected to get can be found at Princeton. There, Jim Gunn took measured CCD QE curves, values for the reflectivity of aluminum, and for an atmosphere, and came up with effective filter curves.

New curves now at filters_jeg_jun2001.dat: Jim Gunn has recently synthesized filter response curves for the 2.5m telescope, using Mamoru Doi's measurements and including the responses of the mirror surfaces and the transparency of the atmosphere. These should still be considered preliminary, but that they are the best available at the moment.


Mamoru Doi built a machine to measure the QE of the camera. He measured the actual QE of the camera and CCDs as put now existing on the telescope, and found (Doi's reports: 1 and 2 and 3 ) that the filters had changed over time, probably due to their being used in vacuum.

The new data is on Mamoru's web page (at

So far, Mamoru has performed the calibration only on CCD column 1.

Status of Laboratory Filter Testing

Testing continues in Japan: Fukugita's report

Construction of the New Curves

We take Mamoru's curves and divide by 0.000034, which is his normalization factor to make the filter+CCD+alum curves of JEG and his filter+CCD curve agree. The effect of the changing reflectivity of aluminum is mimimal, changing from 0.84 at 400nm to 0.89 at 900nm, less than 10% and very smooth. We will therefore take Mamoru's curve as including alum though the constant normalization factor.

We are left with applying the effect of the atmosphere.

I take the Palomar atmosphere and scale to 1.2 airmasses and to APO via

	mag = mag * 1.2 * e^((1700.-2788.)/7000.)
Applying this to Doi's measurements gives the plot below, where it is seen that the main effect is to shift the filters redward edge to the blue:

The green line shows JEG's numbers for the filters+atmosphere, the white line shows Doi's measurements, and the red line shows the resulting new filter curves.

James Annis
15 June 2000