The Observation Summary of South Galactic Cap U band Sky Survey (SGCUSS) Zhen-Yu Wu, Xu Zhou and Zhou Fan 1. process September: In this month, Xu, Zhou, Zhen-Yu, Mike, and ED went to Bok telescope and tested the camera. At that time, the camera only had 3 CCD chips. Only 2 CCD chips could work, but the readout noise >100 e. We just obtained some g band images to check the focus plate and to test the method of coordinate calibration. The lamps for u band flat-field image were not ready. October In this month, Zhou, Zhen-Yu, Mike, and ED went to Bok telescope, and spent much time to test the camera. At that time, the camera only has 3 CCD chips. 1). 9-30: Twilight flat images were taken at the beginning of sunset. BIAS images were also obtained. Other images were obtained for test or under bad whether conditions (cloudy after midnight). 2). 10-01: In the daytime, the dome flat, daylight flat (with 1/5 size of the window of the dome) were obtained. In the nighttime, the sky is clear, part of U band Survey fields were observed. At the beginning of morning twilight, dome flat and BIAS imaged were obtained. 3) 10-02: the domes flat and BIAS images were obtained at the beginning of the observation. The sky is clear and parts of U band Survey fields were observed. At the end of the observation, BIAS images were obtained. 4) 10-03: the dome flat and BIAS images were obtained at the beginning of the observation. But the weather became bad after 1.5 hours observation. 5)10-04: at the beginning of observation, the temperature of Dewar is +5.0 ºC, liquid nitrogen were added and the Dewar need about 10 hours to cool down. The weather was also bad at the whole night. 6) 10-05: bad weather. Only BIAS images were obtained. November In this month, Zhou, Zhen-Yu, Mike went to Bok telescope. Mike use a new computer to replace the old one for running observation scripts with Windows operation system. At that time, the camera only has 3 CCD chips. 1). 11-02: We took the BIAS and DOMEFLATS. We initialized the telescope. Wind was strong. Frequently we met the tail and jumping problems. Some images were lost. The wind speed given by weather station was not right. 2). 11-03: We took the BIAS and DOMEFLATS. Gale. Dome closed until 4:00am. 3) 11-04: We took the BIAS and DOMEFLATS. We initialized the telescope. And the telescope jump was frequently. 1
4) 11-05: We took the BIAS and DOMEFLATS. Telescope jump was very frequently. 5)11-06: We took the BIAS and DOMEFLATS. Jumped and bad images happened frequently. 6)11-07: We took the BIAS and DOMEFLATS. We initialized the telescope. Jumped and bad images happened frequently. 20:35 Readout stuck! Reboot software and then the two PCs: useless! Mike fixed that. December In this month, FAN Zhou went to Bok telescope. At that time, the camera still keep only 3 CCD chips. 1). 12-03: We took the BIAS and DOMEFLATS. Initialized the telescope. At the beginning, part of a CCD had the readout problem. Mike changed the CCD's parameters to the old configure and then the CCD becomes 3 and a quarter. After rebooting both BokCCD & BokCCD2, the readout problem was fixed. In the end, we took the BIAS and DOMEFLATS again when configure has been changed. 2). 12-04: We took the BIAS and DOMEFLATS. Bad weather, we took a set of DOMEFLATS with different time to test the low-level linearity. We also took a dark. 23:00 the weather turned clear. 3) 12-05: We took the BIAS and DOMEFLATS. Initialized the telescope. Weather turned bad after 3:00. 4) 12-06: We took the BIAS and DOMEFLATS. Initialized the telescope. Everything was normal. 5) 12-07: We took the BIAS and DOMEFLATS. Initialized the telescope. Everything was normal. 6) 12-08: We took the BIAS and DOMEFLATS. Initialized the telescope. 2. Problems ONLY 3 CCD!! 2
α. bad images caused by telescope jump In the following figures, we listed some sample images with the problems caused by telescope jump. small and large step jump at RA(d5505.0039.fits) small step jump in RA(d5505.0081.fits) jump in DEC (one star become three, d5506.0049.fits) 3
Butterfly shape (d5506.0104.fits) b. bad images caused by readout errors Maybe due to the errors in the readout process, some images have been damaged. A sample image is listed. bad image caused by readout error(d5507.0027.fits) 4
bad image caused by readout errors(d5506_0021.fits, #1 CCD) c. bad images statistics October UT2 UT3 UT4 total 103 107 17 bad 11 0 2 November UT3 UT4 UT5 UT6 UT7 UT8 total 152 35 133 130 138 130 bad 40 15 43 33 18 17 December UT4 UT5 UT6 UT7 UT8 UT9 total 58 55 95 46 67 74 bad 7 3 1 1 2 2 d. Weather statistics October UT1 UT2 UT3 UT4 UT5 UT6 cloudy clear clear bad bad bad November 5
UT3 UT4 UT5 UT6 UT7 UT8 windy gale windy windy windy windy December UT4 UT5 UT6 UT7 UT8 UT9 clear Cloudy - clear Cloudy - clear Cloudy - clear clear clear e. large readout noise In the laboratory, the measurements show the CCDs camera have strong read out noise. The readout noises of all the gates were about 6-8e. By using the flat-fields and bias frames taken during the observation, we found the read out noise are much higher than the result of laboratory. The readout noises are about 7-17e. test file: dome FLAT files: p5471.0198.fits, p5471.0199.fits Bias files: p5471.0204.fits, p5471.0205.fits test method: IRAF: findgain (default parameters) The test regions for each CCD are chosen because those regions are free from bad pixels. We also chose different regions for each CCD and get the same results. test result: CCD # 1 region at this CCD gain readout noise gate#1 [1200:1700, 1200:1700] 1.79 7.01 gate#2 [2300:2800, 1200:1700] 1.83 6.83 gate#3 [1200:1700, 2400:2900] 1.77 7.53 gate#4 [2300:2800, 2400:2900] 1.82 13.61 CCD # 2 gate#1 [1200:1700, 1200:1700] 2.25 8.17 gate#2 [2300:2800, 1200:1700] 2.21 10.18 gate#3 [1200:1700, 2400:2900] 2.22 9.04 gate#4 [2300:2800, 2400:2900] 2.20 9.48 CCD # 3 gate#1 [1200:1700, 1200:1700] 2.31 11.42 6
gate#2 [2300:2800, 1200:1700] 2.29 9.26 gate#3 [1200:1700, 2400:2900] 2.26 17.01 gate#4 [2300:2800, 2400:2900] 2.33 11.26 f. the noise of U-band sky background Files: p5471.0104.fits, p5471.0105.fits, p5471.0106.fits exposure time of each image: 2.5 minutes the mean sky background for each CCD were derived with the program written by Zhaoji JIANG. Number of CCD mean background (ADU) #1 135.1 1.80 243.2±15.6 #2 97.9 2.21 216.4±14.0 #3 105.2 2.30 241.9±15.6 mean gain mean sky background±noise(e) The very low sky background noise is close to the readout noise of the CCD. So, it is necessary to reduce the readout noise of each CCD. g. the problem of #2 CCD with low level signal. We used dome flat d5471.00006.fits with DN=2000ADU to divide the dome flat d5471.00014.fits with DN=33,000ADU. Before division, the overscan was subtracted from each dome flat. After overscan subtraction, the DN for d5471.00006.fits is 200 ADU and the DN for d5471.00014.fits is 30000ADU. Following figure is the divided image, obvious structures can be seen from the left regions of this image. d5417.00006.fits / d5417.00014.fits for #2 CCD after subtracted overscan We also do another test, at this time, the dome flat d5471.00013.fits with DN=26,000ADU divide the dome flat d5471.00014.fits with DN=33,000ADU. Overscan 7
was also subtracted for each flat image before division. The following divided image shows that no any structures can be seen from this figure. So, the obvious structure in previous figure shows that there are some problems in #2 CCD at the low level signals. d5417.00013.fits / d5417.00014.fits for #2 CCD after subtracted overscan h. the unstable bias images We used 10 bias images obtained at the beginning of the night ut3oct2010, and 10 bias images obtained at the end of the same night to check the variable bias images. The following figure listed the median values of each gate of #2 CCD for each bias image. Number 0-9 are bias images obtained at the beginning of the night, number 10-19 are bias images obtained at the end of the same night. This figure indicates that during 10 minutes, the variation of bias is about 5 ADU, for some gate, this variation can be higher than 10 ADU. The variation of bias for one night is different for different gates of this CCD. For gate #1 and 2, the variation is about 8 ADU, but for gate #3 and4, the variation is about 40 and 60 ADU. For U band survey, the mean sky background is about 100 ADU for #2 CCD, thus, the large variations of bias for the whole night will affect the determination of sky background. So, in the pipeline software written by Zhaoji Jiang, the overscan is used instead of bias image. 8
3. the effect for SGUSS a. The effect of 3 CCD We have done simulations for survey observation with 3 CCD and 4 CCD. Our simulations indicate that observe the same regions of sky, the used time of 3 CCD is about 1.5 times that used for 4 CCD. Thus, the efficiency of 3 CCD is only 67% of 4 CCD in our simulations. In fact, we got 1145 useful images and observed about 290 deg 2. If we use 4 CCD, we can observe about 670 deg 2. Thus, the efficiency of 3 CCD is about 43.3% of 4 CCD. b. the effect of bad image we get 1340 images, among those 195 are bad images caused by telescope and camera errors, the bad image ratio is about 14.6%. c. the effect of large readout noise, variable bias and low signal problem in #2 CCD Above problems affect the low signal data of U band survey, also affect the photometric precision of survey fields. Maybe all data obtained with present camera can not be used for scientific research!? 4. other problems need to be solved 1) The guiding star camera can not work with the U band survey camera. We have tried to use the guiding star camera during the survey observation, but 9
some bad images produced when the guiding star camera and survey camera work together. Thus, a new automatic guiding star system should prepare. 2) Initialize the telescope At present, we need about 20-30 minutes to initialize the telescope. A automatic method is need to initialize the telescope. 3) twilight flat images We need the observation assistant to take the twilight flat images. 4) Statistics of the beginning time of each survey night (Local time). October UT2 UT3 UT4 Time of begging survey 21:34 22:29 20:22 Evening twilight 19:32 19:31 19:30 November UT3 UT4 UT5 UT6 UT7 UT8 Time of begging survey 22:27 19:33 19:30 19:07 19:24 19:05 Evening twilight 18:59 18:58 18:57 18:56 18:56 18:55 December UT4 UT5 UT6 UT7 UT8 UT9 Time of begging survey 20:06 22:50 19:16 00:14 19:58 19:04 Evening twilight 18:48 18:48 18:49 18:49 18:49 18:49 The data storage location: China: uband(159.226.170.53):/line3/uband-data/raw/scuss USA: Bok(140.252.86.107):/data/primefocus/uband/ 10