Cerro Tololo CCD News

** last update: April 30 2000 **


1. Conversion to Arcon is complete! (Jan 95)

2. STIS 2048 CCD (May 95)

3. Loral 3K CCD (Jun 95, update Apr 97)

4. Loral 1K CCD (Jul 95)

5. Recent Lab Results (Aug 95)

6. Tek (SITe) 2048 #3 (Oct 95)

7. Flat SITe CCDs??? (Feb 96)

8. BTC (aka LACCD) comes to CTIO (Aug 96)

9. SITe 2048 #5 (Oct 96)

10. BAD NEWS!!!! (Nov 96)

11. THE REALLY BAD NEWS... (Mar 97)

12. EVEN WORSE NEWS, PLUS GOOD NEWS!!!! (Jun 97, update Aug 97)

13. BTC agreement renewed for 1999 semester 1 (Aug 98)

14. STIS CCD (ex BURRELL SCHMIDT) commissioned, but briefly (Aug 98)

14. MOSAIC II IMAGER -- eight SITE 2Kx4K CCDs (Aug 99)

15. SITe 2048 #6 Peculiar Trapping Behavior (April 2000)

1. Conversion to Arcon is now complete

As from 1993, CTIO has been changing its CCD control system. In the past, CCDs were controlled with a so-called "VEB" (Vidicon Electronics Box). We now use a modern, much more sophisticated controller called Arcon.

For the past few years, some applications have used Arcon and others have used VEBs. VEBs have now been completely and permanently phased out. Thus, any information which one may have or find relative to VEBs is completely obsolete and is for reference only. It does not refer to equipment currently in use.

As a result of the conversion to Arcon, TI, GEC and Reticon CCDs have been retired and are no longer available anywhere at Cerro Tololo. Only Tektronix and Loral detectors are currently being scheduled.

2. STIS 2K with Arcon 3.7 on the Curtis Schmidt Telescope

The STIS 2K CCD is a front-illuminated 2048x2048 device with 21 micron pixels, made by Tektronix for the Space Telescope Imaging Spectrograph Program, and kindly loaned to CTIO by Bruce Woodgate and Richard Green. There are four bright columns (which mostly subtract out) and a few blocked columns. All four amplifiers work. A coating of Metachrome has been applied by Thomson (France), in order to provide some blue and UV sensitivity below the normal 4700A cutoff.

This CCD was commissioned at the Schmidt telescope at CTIO during February 1995, reading through the Lower Right amplifier only, using an Arcon CCD Controller. Just recently the CCD has had the remaining three amplifiers tested, and is now working satisfactorily in "quad" mode.

For imaging at the Schmidt, the normal gain index = 2, corresponding to a DcsT = 7 microseconds. Gain is 2.3 e-/ADU with a readout noise of 3.8 e- rms (all four amplifiers are within 10% of these values) and a readout time of 40 seconds. NOTE -- the full well capacity of this CCD is around 150000 e- but linearity starts to roll off severely at only 90000 e-, corresponding to 40000 ADU. Up to this charge level linearity is excellent and differential non-linearity between amplifiers is < 0.3%. We plan to try and improve these figures even more, since we achieve < 0.1% on our two Tektronix (SITe) 2048's.

Quantum efficiency figures are approximately 18% from 3000-4600A, rising to nearly 50% at 6500A, then slowly falling to zero near 10500A. The flat response through the U and most of the B band means that transformations of U-B and B-V photometry to the standard UBV system are straightforward, while in the red the CCD is competitive with thinned CCDs, and has the advantage of not fringing. Some sensitivity figures are as follows (these are scaled from measurents made by Pat Seitzer with the Thonmson CCD -- we'll have the real figures soon).

Sensitivity for BVI filters (10th magnitude star):

Filter     countrate (e-/sec)

B           10900
V           24800
I           13800

In a 30 second exposure the stars of the following brightness will just reach 40000 ADU (dependent on seeing, tracking, focus, so use only as a guide):

Filter      magnitude

B           10.5 
V           11.2
I           10.7

Some typical dark sky brightness values are:

Filter      countrate (e-/sec/pixel)

B                0.7
V                3.0
I                7.9

Cosmic ray rate is 2.5 events/sq. cm./minute, fairly normal.

The defective columns are as follows (a few very small traps visible on backgrounds of a few adu are not listed):

columns         rows affected        type

546-547        905-1024             hot, bleeds slightly 1-904
584            391-1024             trap
587            391-1024             trap
590            391-1024             trap
612           1025-1421             blocked
748           1025-1773             trap
1334            831-1024             hot
1742           1025-1082             blocked
1788            150-1024             hot
1804-1806      1025-2048             very hot
2006           1025-1303             blocked

The pixel scale at the Schmidt telescope is 2.0 arcsec/pixel and the field size is 68 arcmin. The Newtonian secondary vignettes only slightly (correctable by flat fields) and the full field is useable. Image quality is excellent (typically 1.4-1.9 pixels fwhm depending on wavelength -- and filter quality) over the entire field.

The filter bolt holds five 4x4 inch filters. The most popular filters are UBVRI, for which the following focus values were determined recently:

Filter    Focus (microns)

 U            190
 B            250
 V            230
 R            265

These values were measured at 15C, the focus changes by approximately -5 microns per 2C temperature change.

It is anticipated that by July the focus and filter changing will be controllable from the Arcon user interface, bringing to a completion this phase of upgrades at the Schmidt telescope, a combined effort by the University of Michigan and CTIO.

Alistair Walker & Ricardo Schmidt. 26 May 1995

3. Loral 3K with Arcon3.5 on 4M R/C Spectrograph

Updated with new info, April 1 1997.

The new Loral 3K CCD plus Blue Air Schmidt combination was first tested with the 4.0-m RC spectrograph during an engineering run on 15-16 February 1995. This is a thinned 3K x 1K CCD with 15 micron pixels. The CCD has a two layer AR coating and is UV flooded to maximise its QE over a wide range of wavelengths. It is flat.

All indications are that it is superior in all respects to both the Blue Air Schmidt + Reticon and the Folded Schmidt plus Tek1024. In particular we believe it to be the best choice of CCD for all the CS CCD runs in the present block of observing time.

Gain, readout noise, etc.:

The Loral 3K has two working amplifiers (lower left LL and lower right LR). However only one can be used at a time. We have set up the two video channels to have almost a factor two difference in gain, to allow the user more freedom of choice. Looking at the table below, it can be seen that LL is better optimized given that the full-well capacity is only 78000 e-. (ie LL, gain 2, gives 1.99 e/adu, 7.7 e- RON, and full well will occur at 39000 ADU). UNFORTUNATELY, the high video gain needed for LL has meant that LL suffers from some stability problems (noise bands, bias drifts) and FOR THE MOMENT, we recommend using the LR amp, at gain 4. The only advantage of gains 1,2,3 with LR is readout speed, use these gain settings only if readout time is critical for your program.

  i                       ARCON 3.5  /  Loral 3K
  n                                                           Full CCD
  d DCS    ___Read_Noise___  ____1/Gain_____  __Read_Noise__ SingleRead
  e (us)        (ADU)             (e-/ADU)         (e-)      Time (s)
  x            LL    LR          LL    LR        LL    LR
  ----------------------        ----------      ---------    ----------
  1   5       2.54  1.48        4.33  7.82      11.0  11.6      88.2   
  2  10       3.88  2.17        1.99  3.96       7.7   8.6     120.5  
  3  15       5.68  2.97        1.39  2.59       7.5   7.7     152.4   
  4  20       7.08  3.87        1.03  1.94       7.3   7.5     184.4 

Dark current is extremely low, 0.48 e-/pixel/hour

QE and System Efficiency:

The QE of the CCD (measured at KPNO) is:

Wavelength	QE (%)
3200A		78.9
3650A		73.9
4050A		73.0
5000A		86.6
6000A		93.0
7000A		93.9
8000A		73.9
9000A		41.8

There has been some scepticism expressed that the QE figures below 3000A are very optimistic. We do not yet have any really definitive measurements of our own, but figures of around 30-40 % at 3500A may be nearer the truth. The below system efficiency figures assume that the KPNO QE measurements are correct.

The overall system efficiency (fraction of photons striking the primary mirror which are detected by the CCD) was measured using standard stars. Using grating KPGL1 (632 l/mm 4200A blaze) and a wide (10") spectrograph slit the measured efficiency was:

Wavelength	Loral3K		Reticon
3000A		 2.4%
3500A		10.6%		 8.0%
4000A		14.1%		 9.6%
5000A		18.6%		10.4%
6000A		14.3%		 8.1%

The third column gives values for the Reticon #2 CCD using the same grating.

Image quality:

With a very narrow (50 mu slit which projects to 0.6666 pix), and at best focus, the measured FWHM of comparison lines is 2.3 pix. For a slit width of 150 mu (2 pix, 1.0") the best FWHM grows to ~2.6 pix, while at 225 mu (3 pix, 1.5") it is ~3.0 pix. There is slight curvature of the focal plane which results in some variation of the FWHM with position on the CCD. With the 150 mu slit the images are 3.3 pix FWHM or better over most of the chip (~4 pix in the extreme corners), while with a 225 mu slit the images are 4.5 pix or better over most of the chip (~5 pix worst case). Even the worst case images are quite symmetrical, and do not show the very broad assymmetric wings seen in out of focus images obtained with the Reticon (B A/Sch) and Tek1k (F/Sch CCDs. In general the images obtained with the Loral are much more uniform than with these other CCDs.


The gollowing table lists the coverage and dispersion (A/pix) obtained with the various gratings available for the R-C spectrograph. It is also valid for the Argus multiple object spectrograph.

Grating	 l/mm	 Blaze %	Cover.	Dispn.		Notes
	 	 (A)	 	 (A)	(A/pix)
250   	 158   	 4000   	11431   3.75   
400   	 158   	 8000   	11431   3.75   		*
510   	 300   	10000    	 5999   2.01   		*#
181   	 316   	 7500   	 5708   1.91   
kpgl2    316   	 4400   	 5708   1.91   
kpgl3    527   	 5500   	 3417   1.16   
420   	 600   	 8000   	 2981   1.02   		#
kpgl1    632   	 4200  	 	 2872   0.95   
kpglf    632   	 8200   	 2872   0.95   
450   	 632   	11000   	 2872   0.95   
kpgld    790   	 8500   	 2290   0.75   
kpglf    860   	11000   	 2101   0.68   
380   	1200   	 8000   	 1563   0.48   		#

% Littrow value: for the actual RC spectrograph configuration the effective
  blaze wavelength is 0.92 of the Littrow value.
* This grating is silver coated and so does not reflect light below ~ 4000A
# This grating is not very efficient when used in second order.


This CCD fringes redward of about 7000A. The maximum fringe amplitude is +/-1% which occurs at a wavelength of ~8500A. The fringe spaceing is ~40 pixels. Note that the fringe amplitude for the Loral is less than that for the Reticon (+/- 3-4%). We do not yet know how well the fringes are corrected by flat fielding. The spectrograph flexes by less than 2.5 pixels or 0.06 of a fringe spacing from the zenith to +/- 5h HA. Thus dome flat fields obtained with the white spot will probably be adequate for fringe correction in many cases. Note that dome flats can be obtained at two telescope/dome positions: North 0H, +20d 40m, dome pa 218; and South 0H, -81d 00m, dome PA 039. It may help to use the flat field position according to the declination of your objects.

Nonethless, until more experience has been obtained, we recommend that users working redward of 7000A and requiring better than 1% flat fielding obtain quartz flats (using the same slit width as for the object) for each object.

Note that it is possible to switch between Ne and Quartz lamps under software control. Set the manual switch on the comparison lamp in the cage to the "Quartz position" and select Ne as the comparison lamp in setspec/instrpars. The Ne lamp will automaticaly be selected for exposures of type comp and the quartz lamp for pflats.


4. Loral 1K with Arcon 3.9 on the 1.5m spectrograph

NOTE: Preliminary Information: the following is based on the results of the first engineering night with this system plus laboratory testing.

Initial tests of the Loral 1200 x 800 CCD (hereinafter Loral 1K CCD) plus 1.5-m spectrograph combination, were made during an engineering night on May 17. The science grade chip had only been installed a few days prior to this engineering run so characterisation of its properties was far from complete. However, its performance was judged sufficient to permit an initial evaluation of the image quality, throughput, fringing characteristics and flexure of the system at the telescope. The engineering run was highly sucessful and the measured performance is very encouraging. We expect that the Loral 1K CCD will be made available for visitor use following a second engineering run early in the second Semester of 1995. Thereafter it will be the only CCD offered with the 1.5-m spectrograph. To help in the planning of observations scheduled during second Semester the following discusses what is currently known about the performance of the Loral 1K CCD and compares it to that of the present GEC CCD.

Gratings, Resolution & Coverage:

Gtg	l/mm	blaze	     GEC	    Loral
		    A/pix 	Cover	A/pix	Cover

13	150	 5000	8.4	4800	5.73	6820
11 *	158	 8000	8.0	4600	5.45	6530
09	300	 4000	4.21	2400	2.87	3410
32	300	 6750	4.21	2400	2.87	3410
22 *	300	10000	4.21	2400	2.87	3410
58	400	 8000	3.15	1800	2.15	2560
16	527	 5500	2.36	1350	1.61	1920
26	600	 4000	2.11	1200	1.44	1705
35	600	 6750	2.11	1200	1.44	1705
56	600	11000	2.11	1200	1.44	1705
47	831	 8000	1.5	 860	1.02	1220
36 &   1200	 7500	1.05	 600	0.72	 850	

% Blaze is first order Littrow blaze. Effective blaze 
wavelength when used in the 1.5-m spectrograph is 0.89 
of the Littrow value.

* silver coated does not reflect light below ~4000A 
& Cannot be tilted far enough to be used in II order

The GEC CCD has 22 micron pixels; a slit width of 210 microns (3.8 arcsec) projects to 2 pixels. With this device there is no evidence that the resolution of the spectrograph is limited by either the camera optics or the MTF of the detector. The measured FWHM of comparison lines corresponds very closely to the projected width of the spectrograph slit down to the Nyquist sampling limit, and 2 pix FWHM resolution is routinely achieved. There is little variation of image quality with position on the chip, or with wavelength.

The Loral CCD has 15 micron, pixels and is 1.4 times longer than the GEC; a slit width of 143 microns (2.6 arcsec) projects to 2 pixels. Because of the finer sampling and larger size of this CCD it is expected that the camera optics will somewhat limit the resolution, especially at the extreme edges of the field. In addition at KPNO they have been unable to get images better than ~3 pix FWHM with their Loral chips. This has been attributed to diffusion of photoelectrons within the CCD. This effect is greatest at blue wavelengths since higher energy photons are absorbed closer to the surface of the CCD.

The following table shows the measured FWHM of arc lines obtained for a single tilt of grating 32 for a slit width 110.5 microns (2 arcsec) showing the dependance on position on the chip (and wavelength). Values are given for the center of the slit (Y=200) and at the two extreme edges (Y=130, 270).

FWHM (pix) as a function of position
Line	  X		  Y
(A)	 (pix)		(pix)
	    130	200	270	
4471	 112 |  3.12    2.78 	2.82
4764	 215 |  2.96    2.86 	2.70
5015	 303 |  2.89    2.69 	2.70
5876	 602 |  2.65    2.42 	2.46
6678	 879 |  2.52    2.13 	2.25
6965	 978 |  2.47    2.04 	2.47
7384	1121 |  2.62    2.13 	2.46
7635	1207 |  2.97    2.13 	3.14
In general, although the lines are wider than the projected slit width, and there is some variation with position, the resolution with a slit width of 2-3 arcsec is better than or comparable to what would be obtained with the GEC CCD and the same grating.

The graph shows the FWHM as a function of slit width for spectral lines at the center of the CCD (in X and Y). Curves are shown for 3 wavelengths 3888A, 6678A and 9224A


    GEC     Loral 1K
3000    20      25   
3500    19      48   
4000    17      65   
5000    22      83   
6000    35      93   
7000    45      91   
8000    30      83   
9000    14      59   
10000    3      10   

System efficiency:

The following are the measured system efficiencies (percentage of photons striking the telescope primary mirror which are eventually detected by the CCD) for the 1.5-m spectrograph with the GEC CCD using gratings 11 and 13

       GEC			 Loral 1K
    11      13		 11	 13
3000    0.0     1.1		 0.0	 1.4
3500    0.2     1.6		 0.5	 4.0
4000    0.9     2.0		 3.4	 7.6
5000    2.2     2.7		 8.3	10.2
6000    7.0     4.9		18.6	13.0
7000    6.4     3.6		12.9	 7.3
8000    3.6     2.0		 9.9	 5.5
9000    1.5     0.7		 6.3	 2.9
10000   0.4     0.0		 1.3	 0.0

The numbers for the Loral were estimated by scaling by the ratio of the QE's given above.

Unfortunately the engineering night had heavy cirrus / thick cloud. Therefore we do not have any measurements of the absolute sensitivity. However, observations of standard stars confirm that the sensitivity peaks at approximately 6000A, where the QE curve peaks, and that there is significant sensitivity down to the atmospheric cutoff at 3000A.

RON & Dark Current:

So far only the lower left amplifier has been comissioned and the CCD is being read in single channel mode.

The following table shows the gain (e-/ADU), RON (e-), and readout time (s) for the currently available gain settings.

			 Arcon3.9 == Loral 1K (1200*800)
   n                                                  Bin1x1  Maximum
   d  DCS  Delay  Read_Noise    1/Gain -> Read_Noise   Read   Linear
   e time            (ADU)      (e-/ADU)     (e-)      Time   Signal 
   x (us)          LL    UR     LL   UR    LL     UR    (s)   (ADU)
     ----  -----  ----------   ---------  ----------   ----   -----
1:   1   5     3     2.5   0     4.11   0   10.3    0    25.1   21900
2:   2   7     3     2.6   0     2.87   0    7.6    0    29.1   31400
3:   3  10     3     3.6   0     2.05   0    7.2    0    35.3   43900
4:   4  14     3     4.9   0     1.42   0    6.9    0    43.4   63400
5:   5  20     3     6.7   0     0.96        6.5    0    55.6   65534
Full well (~90,000 e- is reached before the ADC saturates at the higher gains (more e-/ADU). The Non-linearity (peak-to-peak gain variation) is believed to be less than 2% for levels below full well / ADC saturation.

Currently the dark current is very high ~15e-/pixel/hour. However, it is expected that this will be reduced to a few e-/pixel/hour by opperating the CCD in MPP mode and by running it at a lower temperature. Fringing:

The Loral 1K CCD fringes with substantial amplitude at wavelengths redward of 7500A. Press here if you really want to be horrified . The fringes run approximately perpendicular to the dispersion. The peak-to-peak amplitude and fringe spacing are given in the following table and shown in the accompanying graph:

Wavelength	Amplitude	Spacing
(A)		(%)		(A)
7500		 2.5		38
7750		 2.8		40
8000		 4.5		40
8250		 8.4		46
8500		11.3		38
8750		14.6		42
9000		16.0		54
9300		20.6		54
9500		19.6		56
9750		17.0		66
10000		11.4		50
10500		 7.4		60
At least in flat field frames the fringe amplitude does not depend on slit width. Nor does it depend (to first order) on the position where the light of a particular wavelength falls.

We do not yet know how well fringing can be corrected by flat fielding techniques. Given the above amplitudes it seams likely that for wavelengths shortward of about 8000A fringing is unimportant or easily correctable. Redward of this it is likely that it will be necessary to obtain quartz flats for each object and take great care in flatfielding the data. Even then, observations requiring high S/N at wavelengths redward of 8000A may not be possible with this CCD.


The following table shows the displacements (pixels) due to flexure parallel and perpendicular to the dispersion as a function of Hour Angle and Declination.

			    Parallel to Dispersion

  |                               HOUR ANGLE
DEC  |              WEST                                     EAST
+30  |                                   0.0h
  |                                  +0.10
0  |               3.0h                0.0h                3.0h
  |              +0.35               -0.16               -0.23
-30  |     4.5h      3.0h      1.5h      0.0h      1.5h      3.0h      4.5h
  |    +0.75     +0.42     +0.14      0.00     -0.09     -0.10     -0.06
-60  |   5.0h                            0.0h                            5.0h
  |  +0.79                           +0.13                           -0.06
-90  |                                   0.0h
  |                                  +0.68

			  Perpendicular to Dispersion

  |                               HOUR ANGLE
DEC  |              WEST                                     EAST
+30  |                                   0.0h
  |                                  -0.02
0  |               3.0h                0.0h                3.0h
  |              -0.66               +0.08               +0.34
-30  |     4.5h      3.0h      1.5h      0.0h      1.5h      3.0h      4.5h
  |    -0.75     -0.61     -0.29      0.00     +0.26     +0.34     +0.37
-60  |   5.0h                            0.0h                            5.0h
  |  -0.81                           -0.10                           +0.40
-90  |                                   0.0h
  |                                  -0.49

sheathcote@noao.edu mphillips@noao.edu rschommer@noao.edu rsmith@noao.edu

5. Recent Lab Results.

Recent measurements made in the CTIO La Serena Detector Laboratory.

1. Loral 1K (1200x800) CCD dark measurements

Temperature          Dark rate
degK             e-/hour/pix

122                 4.0
130                 4.5
150                11.5

These measurements confirm the need to operate these CCDs at relatively low temperature (compared to SITe CCDs for instance) in order to reduce the dark rate to a satisfactory low level for spectroscopy. The rate achieved here at 120-130K is typical for Loral CCDs, but at least a factor of 8 higher than for our Loral 3K CCD, which has an exceptionally low dark rate.

Jorge Bravo & Alistair Walker, July 27 1995

2. Loral 3K (3072x1024) Gain Calibration

The Loral 3K CCD has two operative amplifiers on the lower serial register, but only one can be read at a time due to the way that the CCD bond wires are connected (actually, aren't connected...). We have been attempting to see whether or not it is possible to increase the gain for the LL amplifier video chain, without significant penalty of read noise. If so, then this would allow a shorter pixel integration time, and hence shorter read time

This work has been successful. In the table below are the gain/read noise/readtime figures for the two amplifiers, where those for LR represent the original measurements. Most users of this CCD (4m spectroscopy) would have chosen

gain index=4, LR amplifier, 2.0 e/adu, 7.7 e RON, 3 minute readtime

Now they can choose

gain index=2, LL amplifier, 2.0 e/adu, 7.7 e RON, 2 minute readtime

which is a substantial reduction in read time. As is well-known with these CCDs, the read noise is state-of-the-art 1985...

GAIN Index  SLOPE      1/G [e/ADU]      RON [e]      Read Time
		       LL    LR        LL    LR     

1         5 us      4.33  8.21      11.0  12.6     88.2 sec
2        10 us      1.99  4.17       7.7   9.3    120.5 sec
3        15 us      1.39  2.74       7.9   8.6    152.4 sec
4        20 us      1.03  2.03       7.3   7.7    184.4 sec

Ricardo Schmidt & Alistair Walker, August 8 1995


Here are gain/read noise figures for our Tek 2048 #4 CCD (Arcon 3.6), a thinned grade 0 device we have been using at CTIO since 1993. The measuremsnts were made recently in the La Serena detector Lab. No changes or adjustments were made to the system specifically for this test. The measured performance was found to be identical to that measured more than a year prior to this test.

DCS Time 1/G    RON     1/G     RON     1/G     RON     1/G     RON   Readtime
us     e/adu    e      e/adu    e      e/adu    e      e/adu    e      secs
3	6.45	5.70	6.73	5.99	6.39	5.68	6.26	6.31	22
5	3.93	5.30	4.03	4.73	3.81	4.49	3.83	4.57	26
7	2.80	3.86	2.86	3.56	2.72	3.82	2.71	3.80	30
10	2.00	3.15	2.00	3.09	1.92	3.20	1.89	3.28	37
15	1.29	2.77	1.36	2.89	1.27	2.75	1.27	2.79	47
20	0.97	2.49	1.01	2.54	0.96	2.54	0.95	2.63	58
40	0.49	2.26	0.50	2.19	0.48	2.37	0.47	2.27	100
This table shows for CCDs with (conventional) LDD on-chip amplifiers that slope times of many microseconds are needed in order to achieve the lowest read noise. This is, of course, a well-known result. The noise figures at long slope times are significantly contaminated by a small amount of charge injection. If instead the noise is measured in the overscan, values of around 2.0 e- rms are obtained! From the top row of the table, it is clear for short slope times that quantization noise is expected to contribute, since the gain is several e/adu. Thus a supplementary experiment was performed, as follows:

Video amplifier gain increased by a factor of 2.0.

1 Noise re-measured at 3uS DCS Time.


DSC Time  1/G     RON  Read Time
us      e/adu     e     secs
3	  3.2	  4.2	  22
This result shows that, as suspected, quantization noise was an important contribution at short slope times for the first series of measurements, and indeed it would be worthwhile to repeat the experiment with even higher video gain whereby the RON might be expected to reach ~3.5 e- rms, still at a slope time of 3 uS. We have not yet performed this experiment.

Roger Smith & Jorge Bravo, Aug 8 1995

6. Tek 2048 #3.

Tek 2048 #3 is a thinned, AR coated 2048 x 2048 x 24 micron pixels, Grade 1 CCD received by CTIO in 1992. For more than a year it has been dedicated to direct imaging at the 0.9-m telescope, where it produces a 13.5 arcmin square field with 0.4 arcsec pixels. The latter is well-matched to the telescope, since new radial supports for the primary plus the installation of extractor fans in the telescope tube mean that images are now often 0.9-1.3 arcsec compared to the previous median of 1.7 arcsec.

Ever since conversion to use with an Arcon CCD controller (Arcon 3.3) Tek 2048 #3 has been read out using only two (the upper pair) of its four operative amplifiers. This was because a bond wire was detached from its post early in the life of the CCD, during (successful) attempts to remove contamination from the CCD surface... The bond wire was expertly re-attached by Mike Lesser (U. of Arizona) a few months ago and all four amplifiers have been characterized in the CTIO detector Lab. A final check-out recently took place at the 0.9-m telescope, and the "quad" mode of operation has now officially become the default.

The gain table is appended below. The upper pair of amplifiers are slightly quieter than the lower pair, but are all very good in absolute terms. The full well is not large considering the pixel size, and is smaller for the lower half of the CCD than the upper. All three of our Tek 2048's show the same behaviour, explanations are solicited.

d dcsT  ____Read_Noise_____  _____1/Gain____   __Read_Noise___  Read
e             (ADU)              (e-/ADU)           (e-)        Time
x (us)  LL   LR   UL   UR     LL  LR  UL  UR   LL  LR  UL  UR   (s)
----   -------------------   ---------------   --------------- ----
1   5  1.09 1.2  1.02 1.09   5.0 4.8 4.8 4.8   5.4 5.4 5.0 5.4   30
2   7  1.26 1.34 1.21 1.22   3.3 3.2 3.2 3.2   4.1 4.1 3.9 3.9   34
3  10  1.62 1.75 1.46 1.47   2.5 2.4 2.4 2.4   3.8 3.9 3.6 3.6   40
4  15  2.13 2.16 1.69 1.71   1.6 1.6 1.6 1.6   3.4 3.4 2.7 2.7   50
5  20  2.57 2.83 2.10 2.08   1.2 1.1 1.1 1.1   2.8 3.0 2.3 2.4   60

Full well ~150Ke- (quad) = 32K ADU (gain 1) 50K ADU (gain 2)
Full well ~225Ke- (upper) = 45K ADU (gain 1), OK for gain 2  
Full well ~150Ke- (lower) 
Alistair Walker 31.10.95

7. Flat SITe 2048's???.

Well no, we don't have any flat SITe 2048 CCDs. As a product of the method of packaging, thinned SITe 2048's have a bow "upwards", so that the center is typically 200-250 microns higher than the edges. This is of little consequence in slow beams, but in fast beams such as the CTIO 4-m prime focus (f/2.9) the defocus is severe.

John Filhaber designed a "field de-flattener" lens to replace the plane dewar window; this is a plano-concave lens of UV grade fused silica, with 1.0 meter radius. This was installed and tested on the February PFCCD run, and it works superbly well. We determined the power on the lens needed from focus measurements made at PF. SITe now provide a tracing of the surface profile when you purchase a 2048 from them, although care is needed as David Vaughan at KPNO has found that the profile changes (worse...) as the CCD is cooled to operating temperature.

Alistair Walker 19.2.96

8. BTCAM (aka LACCD) comes to CTIO.

BTC (Big Throughput Camera), previously known as LACCD (Large Area CCD) is a Mosaic Imager camera built by Tony Tyson (Bell Labs) and Gary Bernstein (U. Michigan). The detector array consists of four thinned SITe 2048 CCDs. BTCAM has just had a very successful observing run at the CTIO 4-m Prime focus, and next year will become a user-instrument at CTIO. In terms of detected astronomical photons per unit time, or "throughput", it is over 1000 times faster than WFPC/HST and four times any existing ground-based system.
Go to the BTC page.

Alistair Walker 13.8.96

9. SITe 2048 #5

This thinned, Grade 1 CCD, specified to have two operative amplifiers, was purchased in SITe's end-of-the-year half-price sale. It is destined for the new camera being built for the 4-m mult-object fiber spectrographs (first ARGUS, and then in 1998, HYDRA). In the meantime we have put the CCD in a direct dewar and it will mostly be used for direct imaging programs at the 1.5-m telescope. Although three of the amplifiers work at very low noise, identical to our other two SITe 2048's, the fourth is noisy (approx 12 e- rms) and thus the CCD is normally read out through the upper pair of amplifiers only. A brief summary of the CCD specifications is as follows:

     GAIN          e-/adu                    RON (e-)       READ TIME (s)
    INDEX    LL	  LR   UL   UR         LL   LR   UL   UR    (dual mode)

      1     3.93 3.71 4.06 4.04        20  5.0  4.8  5.0        64
      2     2.78 2.70 2.89 2.90        18  4.4  4.1  4.3        74
      3     1.91 1.91 2.05 2.04        17  4.2  3.6  3.7        88
      4     1.31 1.26 1.35 1.33        16  3.7  3.2  3.2       110
      5     0.97 0.95 1.03 1.03        15  3.4  3.1  3.1       132

      Full Well is 230 Ke- (<0.6% linearity over this whole range)
      Dark rate less than 5 e-/hour
      Charge Injection less than 1 e-
      One adjacent pair of blocked columns

Alistair Walker & Ricardo Schmidt 17.10.96

10. BAD NEWS!!!!

First piece of bad news is that the metachrome (lumogen) coating on the STIS CCD at the Schmidt telesope is delaminating. What resembles a long "hair" on the CCD, looks under the microscope like the Andes mountain chain. The analogy isn't a bad one, as it is presumably caused by two "plates" of the coating moving with respect to each other. The coating does appear to still cover the whole surface of the CCD, and flat-fielding continues to work very well. The "Andes" are not high enough to cause de-focus.

Second piece of bad news is that our grade 0 SITe 2048 #4 has developed a noisy amplifier (upper left quadrant). Noisy means in this case 8-10 e- rms, instead of 3-4 e- rms. Apart from the noise, the amplifier is still linear, albeit with rather different bias voltages than previously. This CCD is used mostly at the 4-m. For broad band PF imaging, sky noise usually dominates the read noise so it may be advantageous to still read the CCD in quad mode, ratber than just with the lower pair.

Alistair Walker 4.11.96


SITe 2048 #4 died suddenly in late December. No images can be obtained on any of the four amplifiers. The video out signals are strange, and feed-through pulses are much stronger than normal. Swapping in an engineering-grade CCD showed that the controller was working perfectly. The CCD has been returned to SITe for evaluation. This CCD has mostly been used for direct imaging at the 4-m, some direct imaging at the 1-5-m, and with the Echelles at the 4-m and 1.5-m. We will continue to offer a SITe 2048 for these applications (either #3 or #5). Note that the
BTC Mosaic Imager is also available for use at the 4-m PF.

So what do we have left after this attrition???

SITe 2048#3..(mostly used at the 0.9-m for direct imaging)

SITe 2048#5..(1.5-m and 4-m imaging, echelles, Fabry-Perot)

STIS 2048.....(Schmidt)

SITe 1024#2..(Some direct imaging, also Fabry-Perot)

Loral 1k.......(1.5-m spectrograph)

Lorak 3K.....(4-m RC spectrograph and ARGUS)

Alistair Walker, Ricardo Schmidt & Manuel Lazo 10.3.97


At the end of April the STIS 2048 CCD suffered four broken bond wires when a screw in the dewar came detached and its associated washer ended up on the CCD surface. The CCD was sent to Mike Lesser at the U of Arizona CCD lab. where the wires were successfully replaced. Before it was sent off the opportunity was taken to remove the cracked lumogen coating. The coating has been replaced by an equivalent concoction of laser dyes by Kirk Gilmore of Lick Observatory. He also coated a second STIS CCD, so we will have a spare. Unfortunately (August) both these new coatings are giving trouble. One has delaminated, and the other has many fine cracks when viewed under the microscope. So we will not have one of these CCDs back in service for a while yet. Additionally, it looks like we will only have two working amplifiers on each of these CCDs. Our original CCD seesm to have suffered some damage in the accident mentioned above, while the second CCD has a large serial trap.

The really good news is that we have purchased a Grade 1 4-amplifier SITe 2048. This CCD replaces SITE 2048 #4, which died. The new CCD, called SITe 2048 #6, entered service during July, and will be used for 4-m and 1.5-m imaging, echelle spectroscopy, and with the Fabry-Perot. It is a lovely CCD, nearly perfect cosmetically, with four low noise amplifiers.

Alistair Walker 13.8.97

13. BTC agreement renewed for 1999 Semester 1

The BTC (Big Throughput Camera) will continue at CTIO for semesters I 1999. Recent improvements include replacement CCDs for #1 and #2, a fast data reduction machine, and miscellaneous software enhancements. For Semester II 1999 we plan to offer a clone of the NOAO 8kx8k (SITe CCDs) mosaic imager, presently in use at KPNO on the 0.9-m and 4-m telescopes.

Alistair Walker 27.8.98

14. STIS CCD (ex Burrell Schmidt) Commissioned

The STIS CCD (a 2048 front-illuminated CCD made by Tektronix several years ago) until recently used at the Burrell Schmidt on Kitt Peak, was installed as ARCON 3.7 replacing our own STIS that suffered some trauma last year. The new STIS has a metachrome coating to give some UV sensitivity, and has two operative amplifiers (LL and LR). This CCD was the first to benefit from a video chain upgrade that permitted much faster waveforms, so the read time is only 30 seconds using 2 amps, as fast as our other 2048's when using all 4 amps. We installed the STIS at the Schmidt since it was our plan to take the SITe 2K#5, used there in the meanwhile, and dedicate it to Hydra, the new 4-m MOS. Plans have changed. We'll commission Hydra with the Loral 3K and Air Schmidt Camera, and in the long term use a 2Kx4K CCD with 15 micron pixels. So SITe 2K #5 goes back the Schmidt, and the STIS CCD will be a spare.

Alistair Walker, Roger Smith, Ramon Galvez 27.8.98


The MOSAIC II Imager has been commissioned at the 4-m PF. It uses 8 Grade 2 SITe 2Kx4K CCDs. These CCDs are thinned and AR coated and have good QE throughout the optical. See the Mosaic WWW pages for more details. Both BTC and the PFCCD imagers have been retired, although thanks to Tony Tyson & Gary Bernstein for allowing the BTC to remain as backup during semester II 1999. Note that CFCCD (F/8) is also no longer offered at the 4-m, ie MOSAIC II is the only optical imager offered at the 4-m.

Alistair Walker 27.8.99

16. SITe 2048#6. Peculiar Trapping Phenomenon.

This is a warning when reading the CCD QUAD, or via the LOWER pair of amplifiers. There is a peculiar trapping phenomenon that looks like a serial trap in the lower register, but isn't. And it's intermittant. When it's ON, stars to the right of column 325 will look trailed. Safe solution, but slower readout, is to use the UPPER PAIR of amplifiers only. Here is a description of what happens: 1) Something turns the effect on. It then looks just like a serial trap in column 325 in the lower register. Column 325 itself is a little lower than sky, and stars at higher column numbers show poor serial cte. 2) Once it turns on, it seems to stay on for many hours. 3) If the background is lower than some level then there is no trapping at all and images look OK, but if you do an exposure with high background the trapping is seen. 4) The hypothesis that taking images with sky above some level (around 10 Ke-) turns the effect on (always) does not seem to be correct.

Alistair Walker 29.4.00