IR Guider/Imager photometry (JTR Sep 2007)

The JHKL' throughputs of the guider/imager are, 0.18, 0.23, 0.22, and 0.14, respectively. The throughputs have not changed since first measured in 2000. This is just over half the throughput of the IRTF's facility IR imager, NSFCAM (since replaced by NSFCAM2. (The lower throughput results from losses due to the additional layer of optics needed image and re-image the slit field for IR guiding and the slightly lower QE of the engineering grade guider array.) The noise performance of the SpeX guider/imager and NSFCAM arrays are similar, and so when background-limited, the sensitivity of SpeX is about 0.4 mags brighter than NSFCAM at JHK, and 0.50 mags brighter than NSFCAM at L'. With this correction in mind, observers can use NSFCAM's calculators: Signal-To-Noise Calculator, Calculator, and Exposure Time Calculator.

Note that the observing efficiency needs to be included separately. At thermal wavelengths the efficiency is typically ~60% due to the short on-chip integrations (~0.2sec) and so the exposure time calculated with the calculators needs to be increased by this factor. For on-chip integration times longer than one second or so the observing efficiency is close to 100%.

Up to 1700 DN the array is linear with an RMS scatter of better than 1%. Therefore the data need not be corrected for linearity up to 1700 DN. (It is very difficult to get flat-fielding better than 1% across the array so 1% is a realistic limit for photometry when using stars across the field.) At 2000 DN the counts are about 2% below the 200-1700 DN straight-line fit. So for photometry not to be limited by linearity to the level of 1% keep the counts below 1700DN, for 2% keep the counts below 2000 DN. There might be some oddities at well depths <100 DN but for faint objects the sky should fill the well to above this level.

Dark sky flats work best at JHK. Typical integration times of 60s, 60s, and 120s respectively. Flat fielding does not work in the thermal (L'). In this case dither the object around the field to improve the RMS.

The gain for the SpeX Guider is estimated to be 14 e/DN. The plate scale has been measured to be 0.1185 arcsec/pixel.

Below are the results of measurements of background count rates for the Guider/Imager through the broad-band filters. Also given are the estimated Zero Point (ZP) values for the filters. For accurate photometry take a standard star measurement.

  Background = background count rate in electrons per square
      arcsecond per second (assuming a gain of 14 electrons/DN).

  Zero Point = value to be added to the instrumental magnitude
      to obtain t
               mag = -2.5 log (DN/s) + ZP
      These values correspond to an airmass of 1.0. Values for
      an airmass of 0.0 can be obtained by adding the extinction
      coefficient given below to the given ZP value.
 
 Filter  Background   [magnitude/sq arcsec]  Zero Point Mag    Extinction Coeff
       (DN/sec/pixel)
 ------  ----------------------------------  ---------------   -----------------
  J_MK     3.8             [ 15.5 ]              21.15              0.102 
  H_MK     1.6 x 10(1)     [ 14.2 ]              21.24              0.059 
  K_MK     7.7             [ 13.9 ]              20.57              0.088
  Long6    2.2 x 10(3)     [  5.8 ]              18.6                 -
  L'MK     7.5 x 10(3)     [  4.9 ]              19.12              0.093

See here for a list of available filters. Long6 (3.13-3.53um) is an L-like filter in the Order Sorter Filter wheel which can be used for imaging.

Questions about these values should be directed to John Rayner.

Back to the SpeX Home Page

Maintained by John Rayner.