Principal:
Deputy:
Analyst:
AORKEYS:
Last Updated:
The old approach to calculating 70um Vrst values is the by-eye method used throughout instrument testing. I used SIRTFSEE to calculate slope images for each of the DCEs. I always used the second of the two DCEs at each Vrst value. I ignored somewhere between 72 and 76 frames at the end of each DCE and calculated the slope based on the remaining 4-8 frames. I made column cuts through the third column of each module in the slope image. I then averaged the resulting values by-eye, determining the slope that best characterizes the ensemble of 32 pixels in that column. In cases of saturation, the number I got would be quite inaccurate. However, the value is quite good in the vicinity of the optimal Vrst value, so the method worked well.
For the 160um array, I again calculated a slope image after setting ignore frames to 72 or 76. I then determined the mean for the 5-10 good pixels in each module. These module means formed the basis for the Vrst optimization.
I developed IDL routines to determine the optimum Vrst values. When using these routines, the data are first reduced using mips_sloper with the flags -l -j CamI -s0. In CamI, I edited mips_param_flight_ce1 to set the A70 and A160 parameters for # of Std. Dev. for CR Negative Slope Rejection to 1000000. The advantage of using mips_sloper instead of SIRTFSEE is that I do not have to set the number of frames to ignore at the end of the DCE. mips_sloper will fit as many points as it can before the array hits the top or bottom rail of the A-D converter. A large number of points will be used in the fit when the slopes are close to zero, where the data are of greatest interest. Also, in the case of the B-side of the 70um array, the ramps are highly nonlinear in the Vbias data sets, and fitting to just the first 8 frames can give highly erroneous results. The .red.fits files are used as input into one of a pair of IDL routines:
vrst_70module_meanslope.pro vrst_70module_medianslope.pro vrst_160module_meanslope.pro
The former calculates the mean of the slope values for the 32 pixels
in a readout for each DCE, the latter calculates the median. The Vrst
IER collects 26 DCEs, two at each of 13 Vrst settings. The endpoints are
at nominal bias, while the middle 11 Vrst values run from +5 to -5 mV
bias in 1 mV steps. Two DCEs are collected at each bias. I use the
second DCE of each pair for the middle 11 Vrst values and perform a
linear regression fit to determine where the slope goes to zero. I
average the zero crossings for the four readouts in each module to
determine an optimum Vrst for each module. These numbers are summarized
below. The mean and median routines give very similar results, and there
is no clear preference between the two.
The individual slope values for each Vrst value can be found in the writeups
for MIPS-905 and MIPS-992. This writeup only captures the final Vrst values.
These values were determined using the by-eye method developed during
instrument testing.
The 70um B-side Vrsts (modules 5-8) made a noticeable jump in Campaign Q,
following the first real thermal anneals. These new Vrst values seem to be
holding in subsequent campaigns. The offset of the ramp slopes also changed
during this campaign. The Vrst change is 2.5-3.5 mV. This is large enough
that we need to replace the old Vrst block to get valid data for optimizing
the B-side of the 70um array. The new Vrst IER is expected to be used for
the first time in the first MIPS campaign.
The following is a summary of the 70um and 160um Vrst data determined
using the IDL routines. These values differ in value from the by-eye
method numbers, but the consistency from measurement to measurement is
pretty comparable. The jump in Vrst values in Campaign Q is confirmed
here. This jump, presumably due to the use of real 70um and 160um thermal
anneals, is present in all 12 modules.
Results
Task Camp 70_1 70_2 70_3 70_4 70_5 70_6 70_7 70_8 160_1 160_2 160_3 160_4
LaunchNom 90E 8F7 910 8E7 916 90B 920 905 97D 91C 91C 913
992 H 912 8FB 910 8E1 912: 903: 91C: 903: 97F 918 91D 915
992 I 910 8FC 90E 8DF 90C 8FF 912 8F3 97E 918 91C 915
NewFlight 911 8FB 910 8E0 90C 8FF 912 8F3 97E 918 91C 915
905 J 911 8FB 90C 8DF 908 8FB 910 8F1 97F 918 91C 915
905 K 910 8FA 90C 8DF 90C 900 910 8F3 980 918 91D 916
992 K 912 8FD 90F 8E1 90C 901 910 8F4 983 91A 920 918
905 O 912 8FB 90E 8E2 90A 8FF 912 8F5 982 919 91F 916
992 O 90C 8F8 90A 8DE 90C 901 914 8F3 983 91A 920 919
905 P 90E 8F9 90A 8DD 90A 8FF 911 8F3 980 918 91C 914
992 P 912 8FD 90E 8E1 90C 901 916 8F3 983 91A 920 918
905 Q 912 8FF 90F 8E1 8FE 8F7 908 8E9 981 91A 91E 917
992 Q 912 901 90E 8DF 8FE 8F4 904 8E9 980 91B 920 919
905 X1 913 8FF 90E 8DF 8FE 8F7 908 8E9 980 919 91E 916
vrst_70module_meanslope.pro
1 2 3 4 5 6 7 8 AORID Task
0.60 0.39 0.52 1.29 -0.34 0.42 -1.60 -1.13 7529984 992-H
-1.43 -1.57 -0.91 -0.91 -0.64 0.45 0.33 0.55 7529728 992-I
0.43 0.11 0.19 2.51 -0.77 -1.08 -1.18 -1.04 7642880 905-J
-0.05 -0.20 -0.69 -2.22 -1.67 -1.11 -2.39 -3.55 7675392 905-K
0.43 0.32 0.24 -1.13 -1.02 -0.81 -3.33 -2.36 7675648 992-K
0.64 0.27 0.40 0.07 -0.92 -0.83 -1.74 -2.27 7737600 905-O
-0.52 -0.22 -0.70 -1.74 -0.94 -0.68 -1.61 -1.68 7739136 992-O
-0.19 -0.15 -0.50 -1.27 -1.04 -1.26 -1.77 -1.95 7780352 905-P
0.67 0.55 0.39 0.15 -0.76 -0.59 -1.00 -1.58 7781376 992-P
1.99 1.88 0.89 -0.63 -4.44 -2.99 -5.14 -5.21 7853312 905-Q
2.37 2.34 1.16 -0.59 -3.40 -2.45 -4.90 -4.81 7784704 992-Q
2.06 1.89 0.74 -0.46 -2.94 -2.12 -4.37 -3.30 7985664 905-X1
2.51 2.50 1.14 -0.59 -3.41 -2.39 -4.71 -4.67 7988736 992-X1
2.07 1.86 0.68 -1.02 -4.37 -3.14 -6.16 -5.48 7859712 905-R
2.40 2.36 1.11 -0.45 -3.96 -2.23 -5.98 -5.00 7849216 992-R
vrst_70module_medianslope.pro
1 2 3 4 5 6 7 8 AORID Task
0.87 0.35 0.43 1.16 -0.31 -0.07 -1.32 -1.11 7529984 992-H
-1.27 -1.75 -1.12 -0.79 -0.80 0.34 0.26 0.79 7529728 992-I
0.63 0.13 0.13 2.07 -0.92 -1.02 -1.05 -1.36 7642880 905-J
0.10 -0.23 -0.69 -2.43 -1.71 -1.47 -2.89 -3.69 7675392 905-K
0.76 0.31 0.24 -1.36 -1.01 -0.74 -2.47 -2.29 7675648 992-K
0.74 0.19 0.18 -0.03 -0.75 -0.84 -1.74 -1.53 7737600 905-O
-0.43 -0.24 -0.69 -1.92 -0.90 -0.69 -2.21 -1.55 7739136 992-O
-0.12 -0.22 -0.51 -1.22 -0.81 -1.11 -2.00 -1.64 7780352 905-P
0.87 0.49 0.26 -0.04 -0.66 -0.71 -1.80 -1.24 7781376 992-P
2.14 1.69 0.74 -0.73 -4.20 -2.84 -5.38 -5.90 7853312 905-Q
2.52 2.15 1.11 -0.63 -3.40 -2.49 -5.17 -4.56 7784704 992-Q
2.26 1.75 0.67 -0.59 -3.19 -1.98 -4.28 -3.26 7985664 905-X1
2.59 2.31 1.10 -0.72 -3.74 -2.36 -4.16 -4.39 7988736 992-X1
2.15 1.71 0.55 -1.19 -4.40 -2.93 -5.32 -5.87 7859712 905-R
2.50 2.17 0.92 -0.47 -3.87 -2.57 -5.57 -5.33 7849216 992-R
vrst_160module_meanslope.pro
1 2 3 4 AORID Task
0.55 -1.28 0.20 0.71 7529984 992-H
0.29 -1.34 0.09 0.57 7529728 992-I
0.81 -1.23 0.31 0.73 7642880 905-J
0.60 -1.22 0.35 0.67 7675392 905-K
1.47 -0.82 0.91 1.39 7675648 992-K
0.83 -1.05 0.80 0.80 7737600 905-O
1.27 -0.87 0.71 1.31 7739136 992-O
0.01 -1.32 0.09 0.16 7780352 905-P
0.93 -0.82 0.96 1.36 7781376 992-P
0.73 -0.85 0.52 1.08 7853312 905-Q
1.37 -0.46 0.85 1.52 7784704 992-Q
0.64 -1.03 0.58 1.02 7985664 905-X1
1.37 -0.46 0.86 1.52 7988736 992-X1
0.71 -0.98 0.57 1.18 7859712 905-R
1.39 -0.30 0.95 1.52 7849216 992-R
Conclusions
In the limited data available so far, the Vrst values seem to be quite
stable for 70um side A and for the 160um array. The 70um side B Vrst
are less stable than the other Vrst voltages. Significant changes were
seen on both arrays in campaign Q, when we performed the first real Ge
thermal anneals. These changes seem to be holding in subsequent campaigns.
Output and Deliverable Products
None
Actions Following Analysis
None