-------------------------------------------------------------- From: C. Jacob Wolfson Date: Wed Aug 20, 10:51am To: t_test@sag.space.lockheed.com Cc: Subject: Yesterday's Images I took a quick look at the "images" that ChrisH gathered yesterday and have a comment or two for lots of people. Many of these are easily seen in the summary listing. 1) For the first frame of sequence DPM.ift_cmmodes the Quad selector moved from 10 to 19 (should be 20) and FW#2 moved from 89 to 125 (should be 125). On the 5th frame the Quad went to 20 and stayed there the rest of the time. 2) The exposure times for tap_dark_vs_T all read 6 (24 msec) rather than the desired 2 and about 20 msec values. However, noting how the maximum value of the brightest pixel varies between the two exposures I think they are different times. I have no idea what times - the dark current seems to low for the long exposure to be 20 seconds. Interestingly, the exposure value of 6 is what DPM.ift_cmmodes set. 3) The summary listing shows 4 images of many_pix at the end. The first of these has a 19.484 sec; which is what the long one of tap_dark_vs_T was to have. It's Imax value is like the long ones so maybe they were all 19.484 seconds and I need to rethink my comment on dark current. Like it is not completely impossible since this calculates to .1 elect/sec assuming the B amplifier has 45 e/DN and we published a prelaunch value of .3 elect/sec for the cold CCD of MDI ==> can't think why this would be different but we are dealing with small numbers. However a) are the labels off by one? b) why were all the exposure value wrong 4) The next three of many_pix have an exposure of 0.128 seconds which I'd tend to bet is what ChrisH asked for. And the Imax pixel is more than "short exposure" value and less than the "long exposure" value. 5) The images have horizontal patterns in them but of a very low level. 6) Subtracting two of the "short" exposures from one another yields a nose (all read noise basically) of 0.84 DN or like 38 electrons. Past experience says that anything other than one DN is real good and I bet the value will be less than 38 electrons when we do the same game with the A amplifer. My preliminary conclusions .... The shutter/camera system is doing properly but the data is mislabled. The dark current and noise are nice and low. Will be interested in what a variety of software types think ......... -------------------------------------------------------------- From: C. Jacob Wolfson Date: Thu Aug 21, 10:13am To: t_test@sag.space.lockheed.com Cc: Subject: "poor" pixels The other day we collected what I believe to be a series of soft darks of alternating duration 2 msec and about 20 seconds on B amplifier with the CCD at about -62C. Nothing is alarming but a few items of interest are worth recording. I'd not call them bad pixels but they are "worse" than others For 2 msec software darks on amplifier B The average DN is like 150.79 and this is all offset, not dark current. Pixels 364,671 and 633,594 are "always" the brights; about 160 DN Pixel 315,16 is sometimes a weaker but not random chap. The darkest pixels (about 146 DN) are random. For 19.484 software darks on amplifier B The average DN is like 150.85; the increase from 150.79 being dark current. Pixels 60,76 and 707,210 and 364,671 are "always" the brightest with about 424DN & 265 DN & 200 DN. So, call them hot pixels. The darkest pixels (about 146 DN) are random. We'll see how things look with the A amplifier and how this varies with temperature and how things look with light on them and etc. It also seems to me that we have basically no cosmic ray hits ..... a feature of being inside that vacuum tank perhaps. PS - Mons, you might want to throw this into the Calibration Section although more complete information will come out of the next couple of weeks of testing. -------------------------------------------------------------- From: C. Jacob Wolfson Date: Sun Aug 24, 6:53am To: t_test@sag.space.lockheed.com Cc: Subject: Sequence Warming of CCD I'll look at the data for confirmation but I believe we were running the CCD operational heater to keep it warm during the time of these special sequence tests and quoting from the Thermal Documentation The requirement for excluding power transitions during readout can force powered or unpowered intervals to be extented or deferred. Some unfortunate combinations of the observing cadence and the thermal control interval could result in several extension or deferrals to happen consecutively. The CC flight software shall not guard against it. So we may have tested this feature and looking at the data carefully will confirm if there was a combination of too long of time with heat on and some times things not getting on at all (or exactly how that works, I've forgotten) and thus it all makes sense ..... -------------------------------------------------------------- From: C. Jacob Wolfson Date: Mon Sep 1, 8:35pm To: t_test@sag.space.lockheed.com Cc: Subject: cm_test2 runs We did a cm_test2 run at hot soak temperatures (970901_150731) and things look norminal after some more careful examination. The CCD was at like -40 C and the dark current calculates out to be about 5 electrons and is consistent for both the A and B amplifier. We'll get more data on dark current versus time but it was nice to see this consistency. The noise level is like 1.6 DN or 22 electrons for amplifier A and 0.7 DN or 33 electrons for amplifier B and I think believing in DN levels below one is dicy. Meanwhile we had done a cm_test two the prior day when cold - see 970831_070026 and the images right after that. For some reason the first 5 images of this set were in 2x2 summed mode when they should not have been. They seem to be the first images of the day and I suspect we may not be getting initiated right within this procedure or something??? Once we deliberately command a 2x2 mode within the procedure the rest of it makes sense. But we don't have the A amplifier data to do the noise or dark current calculation. For the B amplifier, one of the pair used for noise calculations is only a partial image (probably lost packets or something) so can't do it. The dark current is basically too low to measure with the 20 sec exposure on B ==> with the numbers saying .00012 DN/sec which would be .005 electrons so let's not believe it. We can get the right value with longer exposures, of course. -------------------------------------------------------------- From: C. Jacob Wolfson Date: Mon Sep 1, 9:03pm To: t_test@sag.space.lockheed.com Cc: Subject: dark current vs temperature A very quick/trial run of dark current versus temperature gives roughly the following. Better runs are anticipated. Oh, yes, the data is for the A amplifier and has various averagings crudely applied. It does show, however, that like MDI ==> the variation of pedestal/offset shows more of a change with temperature than does the dark current when one is in the cold domain Temperature Pedestal Dark Current -67 82.8 0.1 -56 79.3 1.0 -46 76.5 4.6 I realize this is a factor of 3 or so higher than I just typed for the results of cm_test2 data that was obtained during the hot soak condition. These are all small numbers and to sort it out will require more data and being more awake than I am at present but figured it was worth distributing the info for those who are curious as to how things are panning out even before solid numbers are obtained. -------------------------------------------------------------- From: C. Jacob Wolfson Date: Mon Sep 1, 9:58pm To: t_test@sag.space.lockheed.com Cc: Subject: More Dark Current Numbers Sure, I should have combined all these little messages .... but I didn't. Anyway, during the cold soak we ran ltc which has exposures as long as 262 seconds. There were no lights on but we also didn't make it extra dark with the al filters (rather it was fused silica). The CCD was at -67 C and amplifier A showed Pedestal of 82.5 DN Dark Current of .12 electrons/sec Since this is a much longer time period than the others I played with it may be the best value for now. -------------------------------------------------------------- From: C. Jacob Wolfson Date: Wed Sep 3, 11:57am To: t_test@sag.space.lockheed.com Cc: Subject: Looking at some of the images from the functional test beginning around 970903_001251 yields the following. tap_baseline takes a 256 msec image with amplifier A and the internal light at 1.3V is in saturation. We need to either drop to 1.1V or change to a shorter exposure time. cm_test2 has useless 2nd and 3rd images since the change in ADC level puts it under-range when the CCD is cold. The whole changing of ADC levels is rather silly anyway so I'd not bother to change the proc. cm_test2 was run with a light on. This, at least for me is the first opportunity to see image data in camera summed mode without the inconvenience of dark current being the dominate (and confusing) signal. If one simply uses the center extract as the signal plus pedestal+ dark current and the dark corner as pedestal+dark current, the "results" are, for 1.64 msec exposures 1x1 unsummed/normal 13.9 DN 2x2 summed in the camera 53.5/4 = 13.4 DN 4x4 summed in the camera 206.4/16 = 12.9 DN Better background removal would probably make even better answers but the above sure says all is pretty sensible. Dark Current estimate for B amplifier at -67 degrees during the ltc that was run without light yields a value of 0.0023 DN/sec (or about 0.1 electron/sec). The other/sleeping night I put out a bunch of mini messages about pedestal and dark current. I now need to go pull them together and compare this to what had been said. But not right now. --------------------------------------------------------------