AUTHORS : A.R. Martel, G. Hartig, and M. Sirianni
The main goal is to measure the global dark rate of the ACS Solar Blind Channel (SBC) and correlate it with the MAMA tube temperature and the Micro-Channel Plate (MCP) current. The warm and hot pixels are also identified and their count rates measured. A superdark frame is also presented. This report is an update of the TB/TV 1 dark frame analysis of Meurer (SBC Dark Frames, Jun 1999).
All the SBC dark frames discussed here were acquired in two sessions during the calibration phase of the Thermal Balance/Vacuum 3 campaign. The configuration of the instrument is detailed in these Web pages. On Jul 19, four 3600 sec darks (ID 32054-32057) were acquired consecutively at a constant tube temperature of 31.91 C and on Jul 21, eighteen 1000 sec dark frames (ID 32516-32527, 32532-32537) were obtained as the MAMA tube temperature warmed up from 20 C to 31 C. The two groups of Jul 21 darks were separated by a sequence of external 1000 sec F115LP flat fields (PtNe illumination) (ID 32528-32531) during which the tube temperature continued to increase.
A superdark frame was constructed simply by co-adding all the observed dark frames from the TB/TV 3 campaign (see Table 1) for a total integration time of 32400 sec (9 hours) (Note : The SBC is not affected by cosmic ray hits). It is shown in Fig. 1 and its associated histogram in Fig. 2. The counts vary from ~1 per pixel along the edges up to ~5 at the centre. The structure and features of the dark frame are detailed in Meurer (SBC Dark Frames, Jun 1999) and won't be repeated here. The global count rate of the superdark frame is 34.8 counts/sec or 3.31E-5 counts/sec/pixel for all 1024x1024 pixels.
2. Global Count Rates vs Tube Temperature and MCP Current
In Cols. 5 and 6 of Table 1, the global and "local" count rates of the SBC frames are listed. They were calculated by simply summing the counts of all 1024x1024 pixels, including the warm and hot pixels. The MAMA tube temperature and MCP current are tabulated in Cols. 3 and 4, respectively. From Fig. 1, we see that it took approximately ~5 hours for the temperature to reach a plateau near ~30 C. At 30.6 C (5.2 hours), reached in the first post-flat dark frame ID 32532, the SBC heat pipes turned on and abruptly cooled the tube temperature and leveled it off. In zero-gravity, the heat pipes would be continuously active and the warmup of the tube would be smooth with no discontinuity.
In Fig. 4, the MAMA tube temperature and the MCP current are strongly correlated. The thermistor that monitors the MAMA tube temperature is external to the rather complex tube assembly, so may not accurately indicate MCP temperature. The MCP is responsible for the dark current so its temperature is most directly correlated with dark current. Since the voltage between the input-face and output-face of the micro-channel plate is held constant at -2304 V or so, a change in the plate temperature (indirectly indicated by a variation in the tube temperature) will result in a change in the resistance and hence in the MCP current and dark rate. The bump near ~30 C corresponds to the darks acquired after the heat pipes produced the drop in tube temperature, as discussed above.
In Fig. 5, we find that the global count rate correlates well with the smoothly increasing tube temperature over 20-29 C. Over this regime, an exponential relation models well the observed trend. The dark global count rates are well below the global count rate linearity limit of ~300000 counts/sec (see ACS Instrument Handbook). Again, the bump near ~30 C corresponds to the last set of darks acquired after the heat pipes turned on. The last four frames at ~32 C have 3600 sec integrations (ID 32054-32057), hence a poorer resolution in temperature. A similar behavior is observed in Fig. 6, the global count rate vs MCP current and a best exponential fit is shown. We note that for consistency, only the data points at an MCP voltage of -2304.52 V were fitted in both Figs. 5 and 6 i.e. ID 32516 was ignored.
Table 1 : TB/TV 3 SBC Dark Frames
Note : The MCP current, IMCP, is extracted from the JMMCPI keyword in the engineering header. The MCP voltage is held constant at -2304.52 V for all the images except for ID 32516 which is at -2303.90 V.
3. Warm and Hot Pixels
The superdark frame can be used to identify warm and hot pixels. At the centre of the frame, where the counts are highest, the mean count rate is approximately 1.13E-4 counts/sec/pix with a standard deviation of 6.51E-5 counts/sec/pix. We define warm/hot pixels as those with count rates greater than 3 sigmas from the mean i.e. > 3.1E-4 counts/sec/pix or >10 counts/pix. Of course, this limit formally applies to the centre of the detector only and hot pixels may be missed near the edges, where the counts are lower. Practically then, we use the 'zap' procedure in Vista to scan the frame and locate the pixels that deviate by a specified sigma from the local mean calculated in a 40x40 box.
The resultant clean superdark image is shown in Fig. 7 and the 286 hot pixels are tabulated in Table 2. The global count rate on this clean frame is 34.1 counts/sec or 3.25E-5 counts/sec/pix for all 1024x1024 pixels. The hottest pixels are found in three clusters of 2 or 4 pixels denoted as Cluster 1, Cluster 2, and Cluster 3 in Table 2. None of these violate the local count rate linearity limit of ~22 counts/sec/pix. The clusters are visible in the 1000 sec frames and so their count rates can be correlated with the MAMA tube temperature. This is shown in Fig. 8 for the hottest pixel of each cluster : Cluster 1 : (56,282), Cluster 2 : (384,588), Cluster 3 : (843,82). Only the hot pixel of Cluster 2 behaves as expected : the count rate increases with temperature. The count rates of the Cluster 1 pixel are located on a high plateau at ~0.45 counts/sec/pix below ~26 C and on a lower plateau around ~0.2 counts/sec/pixel above this temperature. The count rates of the Cluster 3 pixel are relatively constant around ~0.15 counts/sec/pix for all temperatures. The Clusters 1 and 3 hot pixels are likely more sensitive to the initial SBC voltage (when turned on), followed by a degradation with time, than to the tube temperature.
It is worth mentioning that for STIS/NUV-MAMA, any pixel with a count rate of >2.0E-3 counts/sec/pix is flagged as "hot" and for the FUV-MAMA, this threshold is >1.0E-4 counts/sec/pixel (C. Proffitt, private communication). Thus, the majority of the hot pixels identified in Table 2 would satisfy the FUV-MAMA criterion but not the NUV-MAMA's.
CEI SPECIFICATION :
In Table 4-10 of the Contract End Item Specification (Part II, Nov 1995), the nominal SBC MAMA detector dark rate is constrained to <6.25E-5 counts/sec/pix. From Fig. 5, the observed dark rate is (0.5-5.2)E-5 counts/sec/pix over a temperature range of 20-30 C, within the CEI specification.
The global dark rate of the SBC MAMA detector was measured as a function of tube temperature. Hot pixels were identified and a superdark frame was constructed, with and without the hot pixels. The local count rate of the three hottest pixels was also measured as a function of temperature.