The CIMAFITS Log

Modifications to the Arecibo FITS-format (CIMAFITS) and related back-ends

Note that modifications reported in the CIMA user's log or the CIMA engineering log are NOT repeated on this page.
Complete examples of the CIMAFITS headers are available on the CIMAFITS Headers page.
You can return to the main CIMA page by clicking here.

5 January 2009: Correction of old UT midnight roll-over bug

Since the beginning of the new year, an old UT midnight roll-over bug has reappeared. There has been a problem with coordinate interpolations at UT midnight roll-over (20:00 AST). During 20 seconds after UT midnight roll-over, coordinate parameters have been calculated incorrectly. This bug was originally fixed on 4 April 2006, but the fix stopped working and was removed on 7 March 2007 since it caused another problem. The original fix has now been restored. The version number of the CIMAFITS-files was not modified after the fix, but the CIMA WAPP software called 'Dual' was renumbered to 1.4.17 after the modification.

1 April 2008: Correction of horizontal coordinate bug

A bug has been found that puts in the wrong commanded coordinate system used as well as incorrect coordinate offsets from the commanded center position. The problem only affects data taken with the telescope commanded in horizontal coordinates (used by for example 'fixed azimuth drifting'). The problem is different depending on whether CIMA 2.3 has been used or whether data has been taken with 2.2 or an earlier version of CIMA.

This bug has now been fixed BUT ONLY for CIMA version 2.3. The version number of the CIMAFITS-files has been updated from '1.09' to '1.10' to reflect this correction.

To see the full announcement, click here.

5 January 2008: Correction of new CRVAL1, CDELT1 and UPPERSB bug

A bug has been found that might have caused the CRVAL1 header array parameter to contain the wrong center frequency for data taken immediately after a Doppler correction. In extremely rare cases, also the CDELT1 and UPPERSB header array parameters could be wrong indicating the wrong sideband. The data itself has been taken at the right frequency and with the right sideband - it is just the information in the header that can be wrong. The problem is that CRVAL1 has not been updated but contains the previous frequency value that was used before the Doppler correction. The error is thus typically small.

This bug has now been fixed. The version number of the CIMAFITS-files has been updated from '1.08' to '1.09' to reflect this correction.

To see the full announcement, click here.

7 December 2007: Upgrade and bug fixes of CIMAFITS and WAPP pulsar headers

Several bugs in the CIMAFITS headers have been fixed. DETAILS TO COME ...

The WAPP pulsar headers have also been modified with some bug corrections and the adaptation to handle dual board observations.

The version number of the CIMAFITS-files was changed from '1.07' to '1.08' after the bug was corrected. The header version of the WAPP pulsar files was incresaed from '8' to '9'. The version number of the CIMA WAPP software itself was incremented to 1.4.00.

7 March 2007: Correction of UT midnight roll-over bug

A bug has been found that caused the last telescope positions to be repeated during the first second after UT midnight roll-over (20:00 AST). The solution was to remove the bug fix added to the code on 4 April 2006. It seems that the change may have occurred with the recompilation of the CIMA WAPP software on 6 February 2007.

The version number of the CIMAFITS-files was changed from '1.06' to '1.07' after the bug was corrected. The version number of the CIMA WAPP software itself was incremented to 1.3.06.

15 February 2007: Correction of CRVAL1, CDELT1 and UPPERSB bug

A bug has been found that caused the CRVAL1 header array parameter to contain the wrong center frequency for certain combinations of frequency set-ups. In some cases, also the CDELT1 and UPPERSB header array parameters could be wrong indicating the wrong sideband. The data itself has been taken at the right frequency and with the right sideband - it is just the infromation in the header that can be wrong. CRVAL1 is typically off by several hundred MHz when it is wrong.

Pulsar observations could also be affected since the sideband indicated by IFLO_FLIP has been affected by the same bug.

This bug has now been fixed. The version number of the CIMAFITS-files has been updated from '1.05' to '1.06' to reflect this correction.

To see the full announcement, click here.

8 February 2007: Change of WAPP Doppler calculations

The formulae used for Doppler calculations when observing with the WAPPs have been changed to include the second-order term that so far has been omitted (for a PostScript page with details, click here). New code was also added to make sure that OFF_RA, OFF_DEC, OFF_AZ and OFF_ZA are calculated correctly when observing in galactic coordinates. Also, a bug affecting calculations of the galactic coordinates (CRVAL2G and CRVAL3G) that was introduced on 6 February 2007 was fixed. The version number of the CIMAFITS-files has been updated from '1.04' to '1.05'

29 September 2006: New cal tables for L-Band Wide and C-Band installed

The new calibration tables for L-Band Wide and C-Band have been installed. These values go into the FITS-tables TCAL_FRQ and TCAL_VAL. The new L-Band Wide table is valid from 16 July 2006 while the new C-Band table is valid since 25 May 2006. For more details see Phil's pages for the L-Band Wide and the C-band.

10 May 2006: Correction of PARA_ANG header array parameter

The PARA_ANG header array parameter containing the parallactic angle has up until now been calculated wrongly. It has been calculated using J2000 coordinates instead of using coordinates for the current epoch. This has now been corrected and the version number of the CIMAFITS-files has been updated from '1.03' to '1.04' to reflect this correction.

4 April 2006: Correction of UT midnight roll-over bug

There has been a problem with coordinate interpolations at UT midnight roll-over (20:00 AST). During 20 seconds after UT midnight roll-over, coordinate parameters have been calculated incorrectly. This bug has now been fixed. The version number of the CIMAFITS-files has been updated from '1.02' to '1.03' to reflect this correction.

25 May 2005: Correction of BEAM_OFFRAJ header array parameter

The BEAM_OFFRAJ header array parameter has up until now been calculated wrongly. It has been given as decimal hours corrected for the declination ( *cos(Dec) ). This has been changed to be decimal degrees not correcting for the declination. The version number of the CIMAFITS-files has been updated from '1.01' to '1.02' to reflect this correction. Several comments in the FITS-files have also been updated.

20 May 2005: Correction of OBSERVER header parameter

The OBSERVER header parameter does no longer contain the string 'unknown'. It is is supposed to provide the name(s) of the observer entered when a CIMA observing sessions starts, and the bug that has prevented this from happening has now been fixed in the 1.2.02 version of CIMA. Earlier versions of CIMA have not been corrected.

29 April 2005: Correction of PATTERN_ID header array parameter

The PATTERN_ID header array parameter that is supposed to hold a unique ID for each observation pattern (i.e. a complete observation cycle like the pair of ON and OFF scans in an ON-OFF observation) has usually been set up to a constant value at the beginning of an observation sequence and then keeping the same value throughout all loops in the observation sequence. This has been corrected in the 1.2.02 version of CIMA so that PATTERN_ID now gets a unique number for each loop. Earlier versions of CIMA have not been corrected.

22 February 2005: Correction of CRVAL5 data timestamp bug

A bug affecting the timestamps of the data has been found and corrected. Synchronisation errors between 1PPS pulses and NTP-controlled system clocks have frequently caused the timestamps indicating the start of each subscan (CRVAL5) to be off by one second. A typical indication of this is that the parameters ENC_TIME and OFF_TIME are one second off from the time given by CRVAL5 and that positional parameters differ between the WAPPs for the same CRVAL5 timestamp. This bug has randomly affected one or more WAPPs. The timestamps given by CRVAL5 up until now should thus not be trusted. The correct time can instead be obtained from OFF_TIME or ENC_TIME (after rounding to nearest integer).

This bug has now been fixed. The version number of the CIMAFITS-files has been updated from '1.00' to '1.01' to reflect this correction.

To see the full announcement, click here.

3 February 2005: CIMAFITS version 1.00 installed

A new version of CIMAFITS known as version 1.00 was installed on the WAPPs, replacing the old version which didn't have a version number. The new version does not contain any changes to the data format or to the main FITS-header (except for changes of some comments). However, a number of important changes have been made to the binary table header.

The modifications also correct some problems with the FITS-files:

Example FITS-files of CIMAFITS version 1.00: The CIMAFITS 1.00 version incorporates the following modifications in the binary table header:

30 November 2004: Correction of wrong sign for tracking errors

The tracking errors from the telescope have been added with the wrong sign to the commanded position when calculating the actual telescope position. Since tracking error usually are small, the errors introduced usually have been small. All the keywords listed in the note for 15 November 2004 have been affected (CRVAL2A, CRVAL3A, CRVAL2B, CRVAL3B, CRVAL2G, CRVAL3G, CROFF2, CROFF3, CROFF2B, CROFF3B, BEAM_OFFRAJ, BEAM_OFFDECJ and PARA_ANG). This problem was present in the software version used before 15 November 2004, so it has affected all WAPP FITS-files back to 16 September 2004 (and possibly earlier).

To see the full announcement, click here.

15 November 2004: Correction of bad telescope positions

A series of errors have been discovered in the routine used to interpolate the telescope position which was introduced on 16 September 2004. As a result, a number of FITS-header keywords dealing with calculated telescope position information contains erroneous values. The magnitude of the error is usually of the order of 10 arcseconds or less, however occasionally, the error could be as large as several degrees. This problem affects all FITS-files generated between 16 September and 9 November 2004. The following keywords have been affected by the error: To see the full announcement, click here.


This page is administered by Prakash Atreya ( patreya (a) naic . edu ) and was last updated on 9 January 2009.