NIRC2 Measured Sensitivities

TABLE 1. 2001-July/August
("First Light" commissioning)

Filter1	Camera2	Pupil3	Zero Point4 (magnitudes)	Sky5 (mag /arcsec^2)	PSdet6 (magnitudes)	SB Lim7 (magnitudes)	BL Time8 (seconds)
J	Wide	Circ	25.3	15.5	26.7	25.6	52
J	Wide	Insc	24,8	15.1	26.3	25.1	57
J	Medium	Circ	25.1	15.2	27.2	25.3	190
J	Medium	Insc
J	Narrow	Circ	25.1	14.9	27.8	25.2	574
J	Narrow	Insc
H	Wide	Circ	25.3	14.1	26.0	24.9	14
H	Wide	Insc	24.9	13.3	25.4	24.3	10
H	Medium	Circ	25.2	13.7	26.5	24.6	43
H	Medium	Insc
H	Narrow	Circ	25.2	13.6	27.2	24.6	158
H	Narrow	Insc
K	Wide	Circ	24.6	11.4	24.3	23.2	2
K	Wide	Insc	24.2	22.9	24.4	23.2	5
K	Medium	Circ	24.5	11.3	25.0	23.1	9
K	Medium	Insc	24.1	11.8	25.0	23.1	21
K	Narrow	Circ	24.5	11.3	25.7	23.1	36
K	Narrow	Insc	24.0	11.7	25.7	23.0	83
Ks	Wide	Circ	24.5	11.5	24.3	23.2	3
Ks	Wide	Insc
Ks	Narrow	Circ
Ks	Narrow	Insc
K'	Wide	Circ	24.7	11.7	24.5	23.4	3
K'	Wide	Insc
K'	Narrow	Circ
K'	Narrow	Insc
L'  9	Narrow	Circ	23.2	1.6	20.0	17.6	0.016
L'	Narrow	Insc	23.0	2.2	20.2	17.8	0.033
Ms  9	Narrow	Circ	21.2	-1.1	17.5	15.2	0.008
Ms	Narrow	Insc	20.8	-0.5	17.6	15.3	0.021

TABLE 2. 2004 March 31
(After realignment of the AO rotator. The telescope pupil may match the NIRC-2 pupil mask better than before.)

Notes:

#   Inconsistencies in the above table are due primarily to data being taken on different nights.

#  Mean Detector Gain = 4.0 e-/DN,  Readnoise = 37e-/pixel/read
 

1.  All broadband filters are correctly blocked to 5 microns and  are used with an open position on the second filter wheel.

2.    Camera      Wide =      0.04"/pixel
                            Medium = 0.02"/pixel
                            Narrow =  0.01"/pixel
3. Pupils:  NIRC-2 contains 6 pupils, 4 of which rotate in order to track the rotating pupil image.  Pupil rotation was not available during the first commissioning run, so only the circumscribed (circ) and inscribed (insc) circular pupils were used (note that the inscribed pupil is designed to rotate).

     The circumscribed pupil is a completely open aperture, blocking none of the thermal radiation from the secondary, spider vanes, or even the telescope/building seen around the edges of the mirror.This pupil has no mask for the central obstruction or spider vanes.

       The inscribed pupil is the largest circular aperture which lies completely within the zig-zag hexagonal edge of the pupil image.  It does have a central obstruction and spider vanes and normally would be used in a rotating mode.  It was used without rotation for the measurements below (the thermal contribution to the  background from the spider vanes should be minimal, it is dominated by the mirrors in the AO system).  It does block the outer points of the primary, effectively reducing the telescope to a ~9m size.

       Ideally, the sensitivities and backgrounds through these two pupils bracket the expected performance of NIRC-2 when full pupil rotation is implemented.

4.  Zero Point : Defined as magnitude = ZPt - 2.5log[counts/second].  There is  currently no explicit solution for the airmass correction. These numbers may improve a bit as the AO system mirrors are cleaned.

5. Sky    = Sky Surface Brightness in mag/sq. arcsec

6. PSdet  = Point source detection limit. 5-sigma, 1 hr., Aperture diameter = 1.22*lambda/D, min. 2 pixels.
Note that this assumes perfect correction.  If your Strehl ratio is 40% then your point source must be one magnitude brighter than the quoted limit.  If your Strehl is 10%, your point source must be 2.5 magnitudes brighter.

7. SBlim  = Surface Brightness Limit (Noise produced by the background)  1-sigma 1 hr. 1 sq.arcsec.

       The surface brightness limits assume the observed sky magnitudes.

       These numbers may improve, especially in the K-band and longer wavelengths, as the AO optics are cleaned and cooled relative to these initial measurements.

8.  BLTime = Time to reach background limit, in seconds.  Defined as the time when the DCS readnoise (52.3e-) is equal to the square root of the number of photoelectrons.  Note that it is better to be  safely beyond the point of equality, so you should actually go 2-3 times as long as this.  The readnoise on long integrations can effectively be reduced by using multiple endpoint reads (Fowler sampling), allowing one to reach the background limit in a shorter time.

 9. Thermal Imaging with the AO system:  For the L' band (and presumably Ms as well) there is a mottling of the background at a level of one part in 500-1000 when the AO system is on.

 #Grism efficiency: Extensive testing of the spectroscopic mode in NIRC-2 has not been done at the time of this writing.  To first order, you can assume a mean grism efficiency of ~50% over any photometric band filter relative to the filter without the grism.  Slit losses are not included.

NIRC2 Master
10 October 2001