Previous: Time-resolved observations with ISIS
Up: Time-resolved observations with ISIS
Next: Reducing dead-time
Previous Page: Time-resolved observations with ISIS
Next Page: Reducing dead-time

Sources of dead-time

Dead-time can be defined as the interval between exposures during which the detector is not actually detecting photons from an object. There are numerous sources of dead-time, both instrumental and human. In the following discussion we shall restrict ourselves to the instrumental, but it should be emphasized that a great deal can be gained by simply having a carefully prepared observing programme. The following list details the most common sources of dead-time and the amount of time they each contribute to the total figure. The list is in chronological order, from typing the exposure command to receiving the data on disk.

Typing commands -- This is the most obvious form of dead-time and occurs due to the inevitable delay involved in typing a RUN command and its associated parameters. Depending on typing speed, this can take up to 10 seconds. This may sound pedantic, but the use of the MULTRUN command (see section 9.3.2) actually represents one of the largest and most straightforward reductions in dead-time that can be made.

The control system -- Once the command has been typed, it is passed over the Utility Network to the specified CCD controller and the chip is then cleared in preparation for the exposure. The travel time of the command from ICL to the CCD controller is of order a few seconds and, for the time being, represents an irreducible minimum in dead-time.

Clearing the chip -- This is a surprisingly large contributor to dead-time -- a full EEV chip takes approximately 15 seconds to clear. At present, other than windowing or binning the chip (and hence reducing the number of pixels) there is little that can be done to reduce this.

Reading out the chip -- When an exposure has completed, a process known as clocking is used to transfer the charge from the chip. Once this process is complete, each charge packet is detected as a voltage across a capacitance, the voltage is amplified by an on-chip amplifier and then digitised. The data is then passed from the CCD controller through optical fibres to the FOX card in the control room and onto the DMS via DICI. This process, which we will refer to as read-out, is the dominant source of dead-time when observing. It takes approximately 110 s to read-out a full EEV chip, although this can be reduced significantly by windowing, binning and changing the read-out speed (see section 9.3.2).

Archiving the data to disk -- Once on the DMS, the usual procedure is to transfer the data to disk on the VAX. This process takes approximately 45 s for a full EEV chip and also involves the writing of file headers (or packets). Clearly, by reducing the number of DMS-VAX transfers and the size of the transferred frames the dead-time can be reduced (see section 9.3.2).