A brief report on the noise characteristics of the IRS in the
startup anneal sequence for Campaign 18

6 Feb 05 (updated)

Lei Hao and Greg Sloan

Abstract:

As measured by the SURN tests, the startup sequences in
Campaign 18 show read noise which has risen by 16% in SL,
20% in LH, and 24% in LL compared to the startup seqeunces
in the previous two campaigns.  SH appears to be unaffected.
These preliminary results require confirmation once further
anneal sequences from Campaign 18 are available.

Analysis:

    Noise characteristics of the modules (in DN) in startup

           SUR data                SRS data
           noise   pre     post    noise   pre     post
   Module  limit   anneal  anneal  limit   anneal  anneal

    SL      8.8     9.5    10.6    15.0    12.7    12.3
    SH      8.8     6.7     7.5    15.0    12.8    13.0
    LL     14.5    15.7    15.2    15.0    13.3    13.1
    LH     17.6    22.5    20.2    15.0    13.7    13.6

For comparison, here are the baselines for Campaigns 16 and 17,
also from the startup anneal sequences:

   Campaign 16
           SUR data                SRS data
           noise   pre     post    noise   pre     post
   Module  limit   anneal  anneal  limit   anneal  anneal

    SL      8.8     8.5     9.0    15.0     12.3     12.3
    SH      8.8     7.3     7.2    15.0     12.8     12.9
    LL     14.5    12.4    12.6    15.0     12.9     13.0
    LH     17.6    17.9    17.6    15.0     13.1     13.2

   Campaign 17
           SUR data                SRS data
           noise   pre     post    noise   pre     post
   Module  limit   anneal  anneal  limit   anneal  anneal

    SL      8.8     8.5     8.7    15.0    11.9    12.4
    SH      8.8     6.9     6.8    15.0    12.9    12.8
    LL     14.5    13.0    11.8    15.0    12.8    13.0
    LH     17.6    17.8    18.2    15.0    13.1    13.3


  The SRSRN results are unchanged in Campaign 18, but in 
  three of the four modules, the SURN results are 
  significantly higher.  Only SH continues to be well 
  behaved.

Averaging the startup data in the tables and comparing all
anneal sequences in Campaigns 15 and 16 gives the following
results:

          Average for     Max for   
  Module  Camp. 16-17   Camp. 15-17  Camp. 18
  SL       8.7 +/- 0.2     10.1      10.1 +/- 0.8
  SH       7.5 +/- 0.2      8.0       7.1 +/- 0.6
  LL      12.5 +/- 0.5     14.1      15.5 +/- 0.4
  LH      17.9 +/- 0.3     19.5      21.4 +/- 1.6

  From these data, it appears that SL, LL, and LH are 
  degraded compared to the average results of the startup 
  anneal sequences in the previous two campaigns.  SL is 
  slightly worse than its recent maximum, while LL and LH 
  appear to be substantially worse, with noise levels 10% 
  over recent maximum noise values in both cases.  Compared 
  to the average of the recent pre-anneal sequences, these 
  SURN results have increased by 25% in LL and 20% in LH.  
  For SL, the rise is 16%.  Over the past several campaigns, 
  the SURN noise levels have varied substantially from one 
  anneal sequence to the next, so any judgements of the 
  overall health of the arrays without analysis of further 
  anneal data from Campaign 18 would be premature.