8.2 ACS Exposure Overheads

Exposure overheads are summarized in Table 8.1 and Table 8.2. All numbers given are approximate; they do not make detailed differentiations between overheads for different ACS modes and configurations. These overhead times are to be used (in conjunction with the actual exposure times and the instructions in the HST Primer on the HST Proposal Opportunities and Science Policies web page) to estimate the total number of orbits for your proposal, but they may not be exact. Please use the APT scheduling software to obtain the best estimates of overhead times for your Phase I proposal. After your HST proposal is accepted, you will be asked to submit a Phase II proposal to support scheduling of your approved observations. At that time you will be presented with actual, up-to-date overheads by the APT scheduling software. Allowing sufficient time for overhead in your Phase I proposal is important; additional time to cover unplanned overhead will not be granted later.

The following list presents important points for each type of overhead:

Generic (Observatory Level) Overheads

• The first time you acquire an object, you must include overhead for the guide star acquisition (6.5 minutes).
• In subsequent contiguous orbits you must include the overhead for the guide star reacquisition (6.5 minutes); if you are observing in the Continuous Viewing Zone (see HST Primer: Orbital Constraints), no guide star reacquisitions are required.
• Allocate time for each deliberate movement of the telescope; e.g., if you are performing a target acquisition exposure on a nearby star and then offsetting to your target, or if you are taking a series of exposures in which you move the target on the detector, you must allow time for the moves: 20 to 60 seconds, depending on the size of the slew, (see Tables 8.1 and 8.2).

Table 8.1: Science Exposure Overheads: General.

Table 8.2: ACS Science Exposure Overhead Times (Minutes).

Onboard Target Acquisition Overheads

Target acquisitions were used for the HRC coronographic finger and spot. HRC has been unavailable since January 2007. Information about the HRC is provided for archival purposes.

• Onboard target acquisitions were needed to place the target under one of the coronagraphic spots or coronographic finger.
• An onboard target acquisition was done once for a series of observations in contiguous orbits (i.e., once per visit).
• The drift rate induced by the observatory is less than 10 milliarcseconds per hour. Thermal drifts internal to ACS are even smaller. 
  

Scientific Exposures Overheads

• The overhead times are dominated by the time required to move the filter wheel, the CCD readout time, and any necessary serial buffer dumps. Again, it should be stressed that in Phase II, the overheads will frequently be lower, but it is important to plan the Phase I using the conservative overheads given in Table 8.2 to ensure adequate time for the proposal's scientific goals.

Spectroscopy Overheads

• Each CCD spectroscopic observation is preceded by an imaging exposure used for calibration, with exposure times of 3 and 6 minutes, respectively, for grism and prism observations. SBC prism exposures are not preceded by an automatic calibration exposure. Technically, this is an individual single exposure requiring all regular science exposure overheads. For the observer, however, it represents an additional overhead in the observation time budget, so it has been included in the table of instrument overhead times for science exposures. However, if the observing program is already taking an appropriate broadband image, the automatic imaging and associated overheads preceding the spectroscopic grism or prism observations can be avoided by invoking the Optional Parameter AUTOIMAGE = NO during the Phase II preparations. More details can be found in the Phase II Proposal Instructions.

Note that exposures with identical observing modes are automatically generated if the observer specifies:

• Phase II proposal optional parameter CR-SPLIT with a value greater than or equal to 2. (The default value is CR-SPLIT = 0. Single exposures are allowed, but are typically strongly discouraged. A CR-SPLIT value ranging from 2 through 8 may be specified if Number_of_Iterations = 1. However, it is recommended that such individual exposures be dithered rather than CR-SPLIT if possible, since dithering allows the observer to mitigate the effects of cosmic rays, unstable hot pixels, and chip defects, and can allow for the recovery of higher resolution than the native pixel scale of the camera.)
• Phase II exposure log sheet field Number_of_Iterations is greater than or equal to 2 (where CR-SPLIT must be set to "NO.")
• Phase II special requirement PATTERN is used to execute a dither pattern. In this instance, overheads will also include slew overheads.
  

The overhead time for serial buffer dumps arises, in certain cases, from the overheads associated with the onboard data management and switching over the cameras. The onboard buffer memory has the capacity equivalent to a single full-frame WFC image. If a commanded WFC image cannot fit into the available buffer space upon readout, the buffer must first be dumped. This process requires 349 seconds for an entirely filled buffer, or correspondingly less for a partially filled buffer.

Buffer dump overheads

• If a commanded exposure time is longer than about 340 seconds, for an exposure which cannot fit in the remaining buffer storage, an entirely filled buffer can be dumped during that exposure, and no overhead is imposed. However, if the next exposure time is shorter than about 340 seconds, then the dump may be required to occur between the two exposures, depending on the fraction of the buffer.
  
• Sequences of many short SBC exposures can also lead to serial dumps when the buffer becomes full. In this case the buffer dump time becomes an overhead to be included into the orbit time budget. This overhead can severely constrain the number of short exposures that can be squeezed into an orbit. Subarrays can be used to lower the data volume for some applications.

The APT scheduling software has an "Orbit Planner" module that shows the buffer-dump periods in relation to the exposure sequence. The prior discussions regarding parallel versus serial buffer-dumping only applies when no other HST instrument is taking data at the same time as ACS. Please consult the APT Orbit Planner for buffer dump management when using two instruments simultaneously.

8.2.1 Subarrays

The minimum exposure time for WFC is 0.5 seconds and for HRC was 0.1 seconds. The minimum time between successive identical full frame exposures is ~135 seconds for WFC and was 45 seconds for HRC. These times can be reduced to ~36 seconds using WFC subarray readout modes.

At the end of each exposure, data are read out into ACS's internal buffer memory where they are stored until they are dumped into HST’s solid state data recorder. The ACS internal buffer memory holds 34 MB or the equivalent of 1 full WFC frame, or 16 SBC frames. Thus, after observing a full WFC frame, the internal buffer memory must be dumped before the next exposure can be taken. The dump of a completely filled buffer takes 349 seconds and may not occur while ACS is being actively commanded. Of this time, about 340 seconds is spent dumping the image. Correspondingly, less time is required to dump the buffer in parallel, if the buffer is less than full when needing to be dumped to store the next exposure. The buffer dump cannot be executed in parallel with the next exposure if the latter is shorter than about 340 seconds. If the next exposure is less than about 340 seconds, the buffer dump will create an extra 5.8 minutes of overhead.

If your science program is such that a smaller FOV can be used, then one way of reducing the frequency of buffer dumps (and their associated overheads) is to use WFC subarrays. During subarray readouts, only one amplifier is used, and with potentially reduced number of rows from the full 2048. Many more subarray frames can be stored before requiring a buffer dump: four 2K-frames; eight 1K-frames; or sixteen 512-frames. Subarrays with fewer than 2048 rows also benefit from reduced overhead due to smaller readout times.