===================================
 INGRID detector performance tests
 after the change to PCI

 Date: 26,27 June 2003
 A. Zurita (SJG run the dc script)
===================================

a/ Readout noise
   
   - 7.2+/-0.5 adu (in the post-pre)
   - pattern noise in the images (amplitude ~7-20 adu). There seem to
     be some pattern noise bands that change position and amplitude
     from image to image.
   
b/ Hot pixels and dark current (for array temp 82.882 K)
 
   - Aprox. 1% of hot pixels are hot in a 50s integration
     (same number obtained before change to PCI for the same 
      integration time and detector temperature
   - dark current also seems to be within reasonable 
     levels
   See plot 20030626.ps
  
c/ Pre-level in INGRID seems to be too high (~19000-21000 adu;
   depending on quadrants)

d/ Linearity and gain

  -non-linearity ~2% from 0-12000 ADU
  -full-well capacity ~28000-30000 ADU

  Linearity measured with the following configuration:

   o filters ks + Kcont
   o all dome lights on
   o GRACE door closed and lights off inside GRACE
   o This configuration gives ~350 adu/s
   o I used the script 'linearity_mod' with factor 4 (for kcont). 
   o I took 3s runs during the linearity test to monitor background 
     variations. I measured that the background variation vs. UT was 
     within 1%, with no trends (see plot back.ps). Therefore, the 
     test is reliable.
     
 Data in directory: 

 /scratch/whta/almu/ingrid/27jun03/

**Conclusion:

  All the parameters of INGRID are within normal values with
  exception of:
  
	- Bias level (PRE level) which is ~10000 ADU higher
	- There is pattern noise in the images

 It is necessary to confirm whether these are due to the PCI