[Top] [Prev] [Next] [Bottom]

33.1 Calculating Exposure Times

The basic times most observers will be interested in are:

Timing Information for FOS

Type of Information



Observation start time


Modified Julian date

Group start time

FPKTTIME2 - group_elapsed_time

Modified Julian date




Group exposure time per pixel



Group end time


Modified Julian date

Observation end time

FPKTTIMEb (last group)

Modified Julian date

1 Header keyword

2 Group parameter

33.1.1 Exposure Time Details

The different times in the FOS headers are:

Please note that the start time for each individual group should be calculated from the FPKTTIME of the relevant group and not simply by a recursive addition of group elapsed time to the start time of a previous group.

The group parameters can be determined by running the task grlist to determine all the groups in the d0h file and then finding the value of the necessary keyword using hedit. Note that the FPKTTIME is accurate only to approximately 1/8 second.

33.1.2 Timing

Observation start and stop times are determined from the header information with a precision of +0.125 to -0.255 second, while exposure times are determined with a precision of 7.8125 microseconds. This has particular impact on certain RAPID mode timings as described in "RAPID Mode Observation Timing Uncertainties" on page 33-5.

Start and End Times

The start time of the observation (i.e., the time at which the integration was begun for the first group of data) is given in modified Julian date in the header keyword EXPSTART. The modified Julian date is the Julian date minus 2400000.5. Prior to January 1, 1995 EXPSTART was incorrectly populated with FPKTTIME. You should insure that your EXPSTART is approximately equal to the difference between FPKTTIME for the first group and group elapsed time.

The end time of the integration for each group of data is given in modified Julian days in the group parameter FPKTTIME. The end time of an observation is the FPKTTIME of the last group of the observation. Again, in many FOS datasets obtained before January 1, 1995, the EXPEND keyword is incorrectly populated with LPKTTIME rather than FPKTTIME of the last observation group. You should insure that the FPKTTIME of the last group has been used for EXPEND.

To calculate the approximate group start time, subtract the group elapsed time from FPKTTIME. The group elapsed time (which will be nearly the same for all groups in a given observation) can be calculated in two ways: 1) for datasets obtained after January 1, 1995: as FPKTTIME for the first group of data minus EXPSTART, or 2) for any dataset: from the formula given in the previous section. The start time of each subsequent group is then given as FPKTTIME for that group minus the group elapsed time (see Table 33.1).

As noted earlier, an additional uncertainty of approximately 0.13 seconds exists since the HST Science Data Formatter can be delayed in issuing the command to read out the FOS memory, which is recorded as FPKTTIME.

Exposure Times

The elapsed time for an observation differs from the exposure time (the actual integration time during which counts are accumulated) because the elapsed time includes the deadtime during which the FOS is doing housekeeping (e.g., reading out the diodes) and therefore not integrating. The total exposure time for the entire observation (all groups) is given by the keyword EXPTIME in the data header.

The exposure time per pixel, the integration time which contributed to the flux observed in a given pixel, differs from the exposure time for the spectrum whenever substepping is employed. The exposure time is divided among the NXSTEPS individual substepped spectra which together produce the single spectrum. The (typical) exposure time per pixel is therefore given by the exposure time divided by NXSTEPS. This quantity is contained in the group parameter EXPOSURE.1

[Top] [Prev] [Next] [Bottom]

1 Note, however, that the actual exposure time for a given pixel may be different from EXPOSURE, either because the pixel did not receive input from OVERSCAN diodes (because it was an edge pixel or because it was fed by input from one or more disabled diodes), or because a particular readout was rejected by the onboard burst noise rejection algorithm. Thus EXPOSURE actually reports the maximum possible exposure time per pixel. The calfos task correctly calculates the count rate of each pixel, taking into account the number of diode readouts which contributed to the counts observed in each pixel, using the information in the disabled diode reference table to compensate for disabled diodes and the noise rejections tallied in the reject array (the .d1h file) to compensate for times when the counts from a particular diode were not accumulated into memory due to noise rejection.

Copyright © 1997, Association of Universities for Research in Astronomy. All rights reserved. Last updated: 01/14/98 15:41:23