Annual Report on CCD Imaging at the OAN-SPM 2007

Transcription

1 Annual Report on CCD Imaging at the OAN-SPM 2007 Michael Richer & Alan Watson November Introduction This is a report on the state of CCDs and small telescopes of the OAN-SPM. It is based on measurements performed between 26 and 31 October We have published similar reports every year since 2000 [1, 2, 3, 4, 5, 6, 7]. We are grateful to Benjamín García and Fernando Quirós for their help during this run, for their efficiency, and good humor in the face of so many instrument changes. 2 Measurements 2.1 CCD Characteristics We characterized the SITe1, SITe3, and Thomson 2k CCDs. We were unable to characterize the Marconi as it had to be sent to Ensenada for repairs. Table 1 shows the gain and read noise for each CCD. We report these figures for both gain modes for all of the CCDs. We measured read noises and gains using pairs of bias and flat field exposures and the findgain task in IRAF. The read noises are effective read noises, and as such include contributions from amplifier read noise, quantization noise, and spurious charge. Table 2 shows the linearity for each CCD determined from graduated flats. Fig. 1 presents the results graphically. As a measure of non-linearity, we fit a line to the ratio of observed to expected signals as a function of signal level for signals less than ADU. We then multiply the slope of this line by to obtain an estimate of the non-linearity of the detector over its full dynamic range. The non-linearities measured in this way are typically of order 1% or less, except for gain mode 1 with the Thomson 2k CCD, which consistently presents a slight nonlinearity. The single linearity test of the SITe1 at the 1.5-m telescope in gain mode 4 apparently resulted in a nearly 2% nonlinearity. However, it was linear in this gain mode at the 84-cm during at least three tests, so we suspect that the nonlinearity at the 1.5-m is spurious. 1

2 Compared to previous years, the linearity tests contain more scatter about unity. We found this effect last year for the Marconi CCD and attributed it to gain variations. This year, however, it appears in many of the tests. Fig. 2 compares linearity tests based upon long and short exposures. There is much less scatter in the test based upon long exposures. Perhaps, part of the scatter in the other tests is a result of lamp variations on short time scales that the monitor exposures are unable to remove. Another possibility is that shutter errors are contributing to some of this scatter (see Section 2.2). Table 3 shows the CCD operating temperatures and dark currents (per physical pixel). In the past, the temperature reading for the SITe1 controller was unreliable, but this has perhaps now been corrected. During our tests at both the 84-cm and 1.5-m telescopes (in that order), the temperature was stable at the first value indicated in Table 3. Subsequently, after the CCD was moved back to the 84-cm telescope for an observing run, the temperature was unstable and so was raised to the second value presented, resulting in a higher dark current. The difference in dark current at the two temperature settings is reasonable given the difference in the temperature settings. (Typically, the dark current doubles for a temperature change of 6-8 K.) Figure 3 shows the dark current patterns for the SITe3 CCD. This pattern is visible in images when the background is very faint. Our experience indicates that resetting the CCD controller or running the PMIS setup macro has no effect on this (as expected if it is dark current). Table 4 shows the read times (the time to read a full frame to the CCD control computer). For the SITe1, SITe3, and Thomson CCDs, two control interfaces are available, PMIS and the new Linux graphical interface. We were able to test the SITe1 with both interfaces at the 84-cm telescope, but only with the Linux graphical interface at the 1.5-m telescope. At the 84-cm telescope, the read times were shorter with the PMIS interface (these are the ones presented in Table 4). Oddly, the read times with the Linux graphical interface were significantly longer at the 1.5-m telescope than at the 84-cm telescope. Figure 4 presents the bit frequencies for the CCDs tested. These bit frequencies are derived from well-exposed flat field images (> 40,000 ADU). Given the number of counts, the lower bits should sample pure Poisson noise while the higher bits reflect the structure in the image. To illustrate, if the signal level exceeds 40,000 ADU, the Poisson noise will exceed 200/ gain ADU, a value that falls between 100 and 200 ADU for the gains available. Since these values correspond approximately to 2 7 = 128 and 2 8 = 256, equal numbers of bits 0 to 7 or 8 should have values of zero and one for a perfect analog-to-digital converter (ADC). In Fig. 4, the fraction of bits with a value of one is plotted as a function of the bit number. The frequency of 1 s for the lower bits is within 0.6% of the result expected for a perfect ADC for all of the CCD controllers. 2.2 Shutter Errors Figure 5 and Table 5 show the shutter error (the additional exposure time above that requested). These were calculated by comparing long and short flat-field exposures. The variations are as expected for five bladed iris shutters with travel times of ms. However, there are significant pedestals, which are presumably the result of timing 2

3 Table 1: CCD Electronic Characteristics CCD Telescope Mode Binning Gain Read noise Bias SITe1 84-cm e 29.5 e e e e e 8.7 e e e e 624 SITe1 1.5-m e 16.8 e e e e e 7.8 e e e e 660 SITe3 1.5-m e 10.2 e e e e e 12.4 e e e e 1462 SITe3 84-cm e 9.5 e e e e e 10.6 e e e e 1463 Thomson 84-cm e 6.0 e e e e e 5.5 e e e e 412 3

4 84-cm 1.5-m observed signal/expected signal observed signal/expected signal signal signal (ADU) 84-cm 1.5-m observed signal/expected signal observed signal/expected signal signal signal Thomson 84-cm 1.05 observed signal/expected signal signal Figure 1: A comparison of the observed/expected count rates for all of the CCDs. 4

5 Table 2: CCD Linearity CCD Mode Non-linearity SITe1 1 < 1% 4 < 1% SITe3 1 < 1% 4 < 1% Thomson 1 1.3% 4 < 1% Table 3: CCD Operating Temperatures and Dark Currents CCD Temperature Dark Current SITe1 a 99.1 C 1.9 e /pix/h SITe1 b 88.5 C 7.2 e /pix/h SITe C 4.3 e /pix/h Thomson 95.2 C 1.0 e /pix/h SITe 84-cm, observed signal/expected signal , long exposures, short exposures signal level (ADU) Figure 2: A comparison of linearity tests based upon long and short exposures. 5

6 Figure 3: The dark current pattern for the SITe3 CCD. This pattern is visible in images with very faint backgrounds (e.g., spectroscopy) or especially when used with 2 2 binning. The dark current for the SITe1 and Thomson 2k CCDs appears uniform. 6

7 84-cm 1.5-m frequency of 1's fraction of 1's bit 0 bit 1 bit 2 bit 3 bit 4 bit 5 bit 6 bit 7 bit 8 bit 9 bit 10 bit 11 bit 12 bit 13 bit 14 bit bit 0 bit 1 bit 2 bit 3 bit 4 bit 5 bit 6 bit 7 bit 8 bit 9 bit 10 bit 11 bit 12 bit 13 bit 14 bit 15 bit number bits 84-cm 1.5-m fraction of 1's fraction of 1's bit 0 bit 1 bit 2 bit 3 bit 4 bit 5 bit 6 bit 7 bit 8 bit 9 bit 10 bit 11 bit 12 bit 13 bit 14 bit 15 bit 0 bit 1 bit 2 bit 3 bit 4 bit 5 bit 6 bit 7 bit 8 bit 9 bit 10 bit 11 bit 12 bit 13 bit 14 bit 15 bit bits Thomson 84-cm fraction of 1's bit 0 bit 1 bit 2 bit 3 bit 4 bit 5 bit 6 bit 7 bit 8 bit 9 bit 10 bit 11 bit 12 bit 13 bit 14 bit 15 bit Figure 4: The frequency of 1 s as a function of the bit number. A perfect ADC would have a frequency of 0.5 for bits 0 to 7 or 8. All of the CCD controllers are within 0.6% this value. 7

8 Table 4: CCD Read Times CCD Gain mode Binning Read Time Read Time Linux PMIS SITe s 10 s s 5 s s 4 s s 4 s s 18 s s 7 s s 5 s s 5 s SITe s 31 s s 11 s s 8 s s 6 s s 58 s s 18 s s 10 s s 8 s Thomson s s s s s s s s Table 5: CCD Shutter Errors CCD Telescope Center Edge SITe1 84-cm +66 ms +40 ms SITe1 1.5-m +11 ms -5 ms SITe3 84-cm +61 ms +40 ms SITe3 1.5-m +40 ms +18 ms Thomson 84-cm +73 ms +38 ms 8

9 Figure 5: The shutter error (median-smoothed) for the SITe1 CCD at the 1.5-m telescope. The scale is from -10 ms to +40 ms. 9

10 errors on the part of the CCD control system. In the past, we have found that these pedestals appear to depend upon where the data are acquired for the Marconi, SITe1, and Thomson CCDs [7, 6]. This year, we have found variable pedestals for a given CCD at a given telescope on a given night. The values given in Table 5 should be interpreted as typical values, not absolute values. To minimize problems in photometry, exposures of at least 10 seconds are recommended. 2.3 Fringes In previous years we have investigated this issue [5, 9], and there is nothing new to report. Briefly, the Thomson suffers from severe fringing in I and in narrowband filters longward of 6000 Å and moderate fringing in R. The Marconi and SITe3 suffer from fringing in I as well as in narrowband filters longwards of 6000 Å, but of much lower amplitude than the Thomson, and with the SITe3 being the least affected of the two. We have never found evidence that the SITe1 suffers from fringing. 2.4 Zero Points Table 6 shows the zero points for various broadband filters. The zero point is the number of electrons per second expected from a star with a Vega-mag of 0 at 1 airmass and zero colour. 2.5 UBVRI filter sets See the 2005 report [6] for a comparison of the relative efficiencies of the three filter sets. The 2006 report [7] discusses colour terms for broad-band photometry. 3 Comments 3.1 Telescope Reflectivities Figures 6 and 7 show the efficiencies of the 1.5-m and 84-cm telescopes, respectively, for The reflectivity of the 1.5-m telescope is similar to that found last year (when it was last aluminized). The reflectivity of the 84-cm telescope, however, is lower than was measured last year by 20-25%. This telescope was also aluminized last year. Clearly, it would be profitable to re-aluminize the 84-cm telescope at the first opportunity. 3.2 SITe1 CCD The electronics associated with this CCD were damaged by a lightning strike at the beginning of September 2007 and only returned to the observatory just before our run. 10

11 In gain mode 1, the readnoise measured this year is substantially higher than that measured previously. In gain mode 4, however, the gain, readnoise, and spurious charge are similar to values previously found. The readnoise in both gain modes is lower at the 1.5-m than at the 84-cm telescope. The temperature reading is now stable and a colder temperature is maintained. The dark current is lower and its pattern is now uniform. We recommend using this detector in gain mode 4 only. 3.3 SITe3 CCD The dark current is higher than we measured last year [6], though this could easily arise as a result of being operated at a higher temperature. Otherwise, its electronic characteristics are similar to those measured last year. 3.4 Thomson 2k CCD It has been two years since we evaluated this CCD. Nothing seems to have changed. It has the best electronic characteristics of any of the CCDs, but is slow to read out, suffers strong fringing in the red, and is less sensitive than the other CCDs in U and I. 3.5 Shutter Errors These shutter errors have important implications for standard stars. If you place a star at the center of the CCD and ask for a 1 second exposure, the actual exposure time will exceed this by 3 7% (Table 5), meaning that your photometry will also be systematically wrong by this same amount. To get 1% photometry, we recommend taking standard star exposures that are no shorter than 10 seconds at the 84-cm and 6 seconds at the 1.5-m. However, it is probably worthwhile deriving the shutter shading on each run; simply compare the difference between a single 10-second dome flat and ten 1-second dome flats. 3.6 The Tcl/Tk graphical interface This year, the graphical user interface was available for all of the CCDs and was used with the SITe1 and SITe3 CCDs. As Table 4 indicates, this interface reads these CCDs more slowly than does the PMIS interface. We found that the read times with the graphical interface were longer at the 1.5-m than at the 84-cm. Also, when used with the SITe1 CCD, the first 10 physical columns are apparently discarded, since the overscan region begins in columns 1015, 508, 339, and 254 with binnings 1 1, 2 2, 3 3, and 4 4, respectively. This effect was seen at both the 84-cm and 1.5-m telescopes with the SITe1 CCD, but not with the SITe3 CCD at the 1.5-m. 11

12 m 1.20 efficiency/efficiency(2006) U2 B2 V2 R2 I2 filter Figure 6: Relative efficiency of the 1.5-m telescope for the past few years, based upon the zero points from the SITe1 calibrations ( ). All data are normalized to those from 2006, which is when the telescope was most recently aluminized. The mirrors were also aluminized in Table 6: Zero Points Telescope Filter SITe1 SITe3 Thomson 2k 84-cm U B V R I m U B V R I

13 84-cm 1.2 efficiency/efficiency(2006) U3 B3 V3 R3 I3 filter Figure 7: Relative efficiency of the 84-cm telescope for the past few years, based upon the zero points from the Marconi (2006, 2005) and SITe1 (2007, 2005, 2004) calibrations. All data are normalized to those from 2006, when he mirrors were last aluminized. The mirrors were also aluminized in

14 References [1] Reporte de la temporada de ingeniería del de septiembre de 2000, Michael Richer, Alan Watson & Henri Plana, 20 September [2] Las respuestas fotométricas de los CCDs, Michael Richer & Alan Watson, 11 February [3] Reporte de caracterización de los CCDs y de los telescopios de 1.5m y 84cm, Michael Richer, Alan Watson, Simon Kemp, & Sandra Ayala, 9 January [4] Reporte de caracterización de los CCDs y de los telescopios de 1.5m y 84cm, Sandra Ayala, Michael Richer, Alan Watson, & Simon Kemp, 6 January [5] Annual Report on CCD Imaging at the OAN, Alan Watson, Michael Richer, Arturo Godínez, & Teresa García, August [6] Annual Report on CCD Imaging at the OAN 2005, Michael Richer, Alan Watson, Gabriela Montes, & Georgina Bénitez de la Mora, August [7] Annual Report on CCD Imaging at the OAN 2006, Alan Watson & Michael Richer, September [8] CCD-OAN-E2V. Programa para la adquisición de las imágenes de un CCD, Enrique Colorado & Leonel Gutiérrez, December 2004 [9] Franjas de interferencia con el CCD Marconi, Michael Richer, March