Analysis of startup, shutdown and middle sequences, Campaign 6 Abstract: All SRSRN data are within limits. The SL SURN before startup was very high (25.1 DN vs. the 8.8 DN limit) and was over the limit in three other measurements as well. Thus, SL exceeded the 8.8 DN limit in four of six cases. The SRSRN results have a slight larger variation than the previous campaign, but the variations are still within 2%. The variation in SL SURN remains high (as in Campaign 5), and in this case the mid-campaign to shutdown variation exceeds the 20% yellow limit (21%). The change is 12% from startup to mid-campaign, and 14% over the full campaign. The other modules behave well. Only SH behaved well in the photoresponse stability check, with no pixel exceeding the yellow limit. SL, LL and LH all have cases where pixels exceed the red limit. SL and LL have many pixels exceeding the yellow limit, and this problem appears to be attributable to a change in the background. Analysts: Lei Hao & Greg Sloan Procedure version: 2.7 Campaign 6, Startup, Shutdown and Mid-campaign anneal YEAR 2004 DOY 104 CE1 used for all data collection AORKEY = 9537024 for startup AORKEY = 9536512 for the mid-campaign anneal sequence AORKEY = 9536768 for shutdown Number of files = 64 for each No known errors. Steps 2 through 5 run OK. Unruly.pro results for startup (IRS-001): Noise characteristics of the modules (in DN) SUR data SRS data noise pre post noise pre post Module limit anneal anneal limit anneal anneal SL 8.8 25.1 9.1 15.0 12.2 12.5 SH 8.8 9.4 7.0 15.0 13.1 12.8 LL 14.5 14.0 11.5 15.0 12.9 12.8 LH 17.6 17.3 17.4 15.0 12.9 13.1 Unruly.pro results for middle (IRS-002): Noise characteristics of the modules (in DN) SUR data SRS data noise pre post noise pre post Module limit anneal anneal limit anneal anneal SL 8.8 8.0 9.9 15.0 12.3 12.2 SH 8.8 6.5 6.7 15.0 12.9 12.7 LL 14.5 11.1 12.7 15.0 13.0 12.9 LH 17.6 16.9 17.3 15.0 13.1 13.0 Unruly.pro results for shutdown (IRS-003): Noise characteristics of the modules (in DN) SUR data SRS data noise pre post noise pre post Module limit anneal anneal limit anneal anneal SL 8.8 7.9 8.9 15.0 12.2 12.2 SH 8.8 6.6 7.0 15.0 12.7 12.8 LL 14.5 12.1 9.0 15.0 13.0 13.0 LH 17.6 17.5 16.8 15.0 13.2 13.0 The analysis applied the PIRAT procedure unruly.pro to FITS files generated by MIPL. For SUR data, the quoted values are the standard deviation of the read noise map. For SRS data, these are the median of the standard deviation map. Comments: All SRSRN data are within limits. The SL SURN before the initial anneal is very high (25.1 DN vs. the 8.8 DN limit), but this is a common occurrence before the initial anneal. Of more concern is the fact that SL is over the limit in three more cases: after the start-up anneal (9.1 DN), after the shutdown anneal (8.9 DN), and significantly over the limit after mid-campaign anneal (9.9 DN). As in Campaign 5, SL is over the limit four out of six times. Besides SL, SH has one case exceeding the limit of 8.8 DN: before the startup anneal (9.4 DN). The rest of the modules are well-behaved. Changes in read noise between anneals and from startup to shutdown: SUR data SRS data 002pre/ 003pre/ 003pre/ 002pre/ 003pre/ 003pre/ Module 001post 002post 001post 001post 002post 001post SL 0.88 0.79 0.86 0.98 1.00 0.98 SH 0.93 0.98 0.94 1.00 1.00 0.99 LL 0.97 0.96 1.06 1.02 1.01 1.02 LH 0.97 1.01 1.01 1.00 1.02 1.01 Comments: SRSRN has a slight larger variation than the previous campaign, but the variations are still within 2%. The SL SURN keeps varying a lot, the same as the previous campaign, and the mid-campaign to shutdown variation exceeds the 20% yellow limit (21%). The change is 12% from startup to mid-campaign, and 14% over the full campaign. Besides SL, other modules behave well, with variations no more than 7%. Photoresponse stability: Module 002 pre to 001 post 003 pre to 002 post 003 pre to 001 post yellow red yellow red yellow red SL 98 1 1990 0 1992 0 SH 0 0 0 0 0 0 LL 650 1 54 0 667 1 LH 3 0 3 1 5 3 SH behaves well, with no pixel exceeding the yellow limit. SL, LL and LH all have cases where pixels exceed the red limit. For SL, the bad pixel in the 002/001 ratio is 98,91. In LL, pixel 23,119 exceeds the red limit in both the 002/001 and 003/001 ratios. In LH, one pixel in the 003/002 ratio (5,122) and three pixels in the 003/001 ratio (57,15, 1,120, 5,122) exceed the red limit. Pixel 5,122 has been a problem in the past. SL has almost 2000 pixels exceeding the yellow limit in the 003/002 and 003/001 ratios, and LL also has over 600 pixels exceeding the yellow limit in the 002/001 and 003/001 ratios. The PU fields in SL are clearly visible, and the orders are easily distinguished in LL. These features are typical of a changing background. A channel-by-channel analysis of the SURN in SL. The SURN in SL has continued to frequently exceed its limit through IOC, SV, and normal operations. One hypothesis is that the presence of the peak-up arrays on SL make this module more susceptible to jailbarring, which in turn raises the spatially calculated standard deviation of the read noise map. To test this hypothesis, D. Devost has modified unruly.pro to print the SURN results for the individual output channels. Here, we give the results for the four cases where the SL SURN exceeded the 8.8 DN limit. activity measurement total SURN ch. 0 ch. 1 ch. 2 ch. 3 IRS-001 pre-anneal 25.08 25.89 16.29 17.46 24.57 IRS-001 post-anneal 9.14 9.46 8.67 8.95 9.24 IRS-002 post-anneal 9.93 10.24 9.81 9.19 10.46 IRS-003 post-anneal 8.91 8.95 9.27 8.48 8.88 If the jailbarring were responsible for the elevated noise levels, then the individual channels examined separately would show smaller standard deviations than the combined sample. The larger standard deviation of the combined sample would arise from different means in the read noise map for the individual channels. The table above shows that the in each case, two channels have a larger standard deviation than the combines dample, and two have a smaller standard deviation. This result eliminates jailbarring as the cause of the high SURN in SL. In three of the four cases, channel 0 has the highest noise. This fact may be significant, and we will repeat this test in the next campaign to follow it.