Startup Transients, 70um and 160um Arrays

Principal: dmkelly
Deputy: alberto
Data Monkey(s): Round up all the Ge monkeys for this and MIPS-160
Priority: Necessary
Downlink Priority: Normal
Analysis Time: 48h
Last Updated:

Objective

To determine how soon after turn-on we can begin to collect high quality scientific data with the Ge arrays. The metrics are read noise, dark current, responsivity, and stim repeatability.

Description

This task should be run immediately after the end of the MIPS start-up task. It should follow the exact same routines as MIPS-160: 70 and 160 um Thermal Anneal Behavior. This consists of pointing the telescope at a region of sky that is appropriate for dark measurements and pointing the scan mirror to the 70um primary dark position (which also places the 160um array in the dark). Data are then collected by alternating between the read noise and dark current IERs. The test should run for about an hour. MIPS-160 should be run later during the same campaign. Comparing these two data sets will allow us to discriminate between anneal-related and start-up related transient effects.

Data Collected

mips_sur_C0F9N50 
mips_sur_C0F9N1 
mips_sur_C0F9N85 
mips_sur_C0F9N50 
mips_sur_C0F9N1 
mips_sur_C0F9N85 
mips_sur_C0F9N50 


mips_mobs_phot
CESCANPOS '102,"MIPS"'
CEGERSTCON '10,127,4,"MIPS"'
CESCANCON '"CHOP","FWD",0,2048,2048,0,511,"MIPS"'
CEGESTIM '"AUTO",63,"BOTHOFF","BOTHOFF","BOTHOFF",10,10,"MIPS"'
mips_rn_10s50.exp -
  CEMIPSUR '0x0,0x0,0,"NO_COADD",9,50,"MIPS"'
mips_gestimrep_10s86.exp -
  CEGESTIM '"AUTO",63,"BOTHOFF","BOTHOFF","BOTHOFF",10,10,"MIPS"'
  CEMIPSUR '0x0,0x0,0,"NO_COADD",9,1,"MIPS"'
  CEGESTIM '"AUTO",6,"AON","AON","BOTHOFF",10,10,"MIPS"'
  CEMIPSUR '0x0,0x0,1,"NO_COADD",9,85,"MIPS"'
  CEGESTIM '"AUTO",63,"BOTHOFF","BOTHOFF","BOTHOFF",10,10,"MIPS"'
Observing program:  rn,stim,rn,stim,rn
mips_backto_mobs

Data Reformatting Requirements

Array Data Desired:

All Arrays

Data Reformatting Option:

Special Instructions:
Each set of 50 DCEs should go in a separate file (three total). Each pair of 1 DCE and 85 DCE observations should be combined into a single file (2 total).

Task Dependencies

It is essential that we have working and stable stims for this test. If this condition is not met, this task should not be run. All four metrics (read noise, dark current, responsivity, and stim repeatability) are affected by the bias. It is not clear whether the settling time for the arrays is bias-dependent so it is probably not necessary to repeat this test if the Ge bias voltages are changed.

Calibration Dependencies

Output and Deliverable Products

From the read noise data set, we will calculate the 70um and 160um read noise and dark current. From the stim repeatability data set, we will calculate the 70um and 160um responsivity and stim repeatability. We will plot these values versus time and compare them with similar plots from task mips-160. From these two data sets, we will determine how soon after power-up we have stable enough array behavior to get good scientific data from the two Ge arrays.

Data Analysis

Step by step analysis:
1) Run all data through mips_sloper -l -a flight_ce# filename.  Do not run 
   the data through mips_caler or the SSC pipeline since stim flash 
   calibrations are not desired.
2) Process the mips_rn_10s50.exp data using the idl routine get_rdnoise.  See 
   the read noise values below for flags and typical values.  The criterion 
   for stability is not to reach the ground-based values (120e at 70um, 450e 
   at 160um) but to reach stability.  This requires plotting the read noise 
   values versus time and looking to see when the values stabilize (defined 
   by a decrease of less than 10%).  Read noise values should be calculated 
   on an array-wide basis.  It is likely that finer time gradations will be
   needed, which will require running get_rdnoise on subsets of the full 50
   DCEs.  Flags are available that make this easy to do.  From this analysis, 
   an estimate should be made of the amount of time required after power-on 
   before stability is reached.  For this experiment, an average time can 
   probably be assigned to each data set and associated with the read noise 
   values manually.  For later parts of this analysis it will become 
   necessary to extract timestamps from the headers.  SCLK-OBS is probably 
   the appropriate telemetry item.  Time zero will have to come from the 
   housekeeping telemetry.  I am not yet certain how to recognize the end 
   of the MIPS startup IER and the transition to the beginning of instrument 
   operations.
3) The responsivity and the dark current of the array will be changing
   with time during this test.  One measure of these changes is the
   response to the background light during the stim repeatability and
   read noise measurements.  For the read noise data, plot background
   brightness versus time.  The array average is calculated as a part
   of get_rdnoise, so all that is required is to hack the code to write
   these data out to file.  Since only half of the 70um array is in the 
   dark and scattered light is noticeable on the other half of the array,
   make calculations of dark levels on a module-by-module basis.  The
   IDL routine mean_slope_70modules.pro was written to do this for the
   70um dark position test, mips-080.  Start the analysis by plotting 
   the average brightness for each set of 50 DCEs versus time.  Then 
   repeat after breaking each data set in two.  If there is considerable 
   change between the two halves of the data set (greater than 20%), 
   calculate brightness in finer gradations and perhaps for individual
   DCEs and plot versus time.
4) For the stim repeatability data, stim transients will have a significant
   influence on the data.  This is especially true with the 1-minute stim
   interval used during this test.  Each stim interval consists of 6 DCEs.
   Ignore the stim flash DCE and then calculate module-by-module brightness
   levels for each of the 5 DCEs.  Plot versus time.  This will reveal the
   stim latency and should show whether the last DCE is clear of latency
   effects.  Plot the background light levels from the read noise and stim
   repeatability data sets together to see if they match up.  There is no
   criterion to judge whether the array behavior has settled at this point,
   but an assessment will become possible once the data from MIPS-160 have
   been analyzed.  It is unlikely to be either possible or necessary to
   separate scattered light contributions from dark current in this
   experiment.
5) The responsivity will also affect the light level measured during the
   stim flashes.  Since the background responsivity is being plotted on
   a module-by-module basis, calculate stim brightness in the same way.
   Plot brightness versus time for all stim flashes.  Normalize both stim
   brightnesses and background brightnesses to a starting value of unity
   and plot together to compare quick response and slow response responsivity
   behavior.
6) The stim response will certainly change with time, but in a well-behaved
   system the change will be smooth so that stim flashes provide good
   calibration of the changing responsivity.  Use odd stim flash brightnesses
   (measured on 2-minute intervals) to calibrate the even stim flash
   brightnesses.  Use idl routines generated by John Stansberry for this
   (insert names here).  Calculate mean and sigma values for the stim
   brightnesses on an array-wide, module-by-module, and pixel-by-pixel
   basis.  One sigma value should come from each stim repeatability
   measurement.  Plot the values from the two experiments versus time.
   Based on ground measurements, good behavior is characterized by a
   repeatability of xxx% (get number from John).  Compare the data with
   this value to estimate the amount of time required to achieve stability.
7) Identical data will be collected as part of MIPS-160, 70um and 160um
   Thermal Anneal Behavior.  An identical analysis should be performed on
   the data from that test, and the results should be compared to these
   data.  As with this test, a thermal anneal will be performed shortly
   before the start of data taking.  That test will be made later during
   this same campaign, long after CE start-up transients have settled out.
   Many of the trends that we see in the data from this task will be due
   to the thermal anneal and to cosmic ray exposure.  By comparing with
   the data from MIPS-160, the effects of the CE power-up can be isolated.
   Again, the goal is to estimate stabilization times for the read noise
   and stim repeatability and also to look for differences in how the
   responsivity grows with time.  The stability values found during MIPS-160
   should supercede the ground measurements as criteria for assessing the
   stabilization time required for good performance following CE power-up.

Software Requirements

Actions Following Analysis

Determine the warmup time required for good Ge array performance. This time should be determined by comparing the array performance with the requirements and by looking at the rate of change in array performance with time. Look at the activities scheduled for the start of each MIPS campaign and see which tasks are affected. The first task is likely to be the Vrst optimization, which lab testing has shown to be unaffected by warmup transients. If we are starting on dark current, flatfield, illumination correction, or flux calibration activities during this warmup time, we will have to consider putting a pause in the timeline to avoid introducing systematic errors in our calibration. If the settling time is much longer than an hour, a review committee should be formed to decide how to proceed. We will likely follow the usual schedule, but we might have to reschedule the calibration activities for later in the campaign.

Failure Modes and Responses

The current baseline is to allow 30 minutes for CE warmup. In the absence of evidence to the contrary, we should assume that the half-hour settling time is sufficient and we should proceed with our IOC campaigns as written. If the data from this test are rendered invalid, we should calculate the read noise as best as possible from the dark IERs and we should determine the stim repeatability from the dark current, flatfield, illumination correction, and flux calibration tasks. If these parameters meet requirements, the settling time is sufficient and does not need to be changed.

Additional Notes

Summary of read noise data from ground testing, 10s DCEs:
These tests were run using the following flags:
mips_sloper -l -a flight_ce# filename   (replace # with 1 or 2)
  (or use the IDL routine do_sloper, 'filenames', '-l -a flight_ce#')
get_rdnoise, n_dce_rej=10, 'filebase'
The numbers below are the get_rdnoise slopes and read noise for 95% w/ zeros.

                                     70um                     160um
                 Date   CE    RdNoise(e) Bkgd(e/s)    RdNoise(e)  Bkgd(e/s)
ATLO testing    02-108   2    105+/-50   280+/-106    404+/- 375   2844+/- 1812
Brutus testing  01-302   1    123+/-67   212+/- 62    280+/- 275   1315+/- 1188
		01-302   2    132+/-76   204+/- 58    358+/- 287   1629+/- 1398 
		01-286   1     43+/-48    84+/- 91    705+/- 664   5540+/- 3663
		01-228   2     97+/-41   236+/- 68    410+/- 248   1846+/- 1390
		01-227   1     96+/-44   241+/- 69    426+/- 415   1964+/- 1377
		01-226   1    104+/-46   233+/- 67   2105+/-1536  49070+/-33506
		01-222   2     90+/-45   315+/-107 10840+/-11270 275600+/-196200
		01-220   1    104+/-39   185+/- 61   1514+/- 870  28158+/-20203
		01-121ov 1             (n_dce_rej=15) 813+/- 609   2438+/- 1855
		01-117   1    242+/-58   452+/-150 (glitch in 70um data)
		01-116   2    119+/-60   307+/-105   1074+/- 656   5368+/- 2853
Data hereafter were not rereduced because of a missing keyword in the 
headers, so they may not be entirely consistent with the preceeding data.
		00-356   2                           1192+/- 683   4454+/- 3510
		00-353   1                           1684+/-2569   2776+/- 4257
		00-351   1    130+/-50   356+/-114   7581+/-14787 25125+/-42239
LBTC testing    00-124   1                            457+/- 424    967+/-  830
		00-124   2                            445+/- 343   1360+/- 1284
		00-124   2                            376+/- 255   1344+/- 1312
		       FLIGHT 160um ARRAY INSTALLED
		00-038   1    152+/-69   293+/- 86    322+/- 240    740+/-  705
		00-031   1    151+/-68   246+/- 99    317+/- 186    866+/-  925
		00-031   1    147+/-68   259+/-101    563+/- 477    812+/-  694
		00-024   1    261+/-120  242+/-142    254+/- 175    665+/-  642
		00-024   1    261+/-123  246+/-142    294+/- 246    702+/-  798