WRONG CENTER FREQUENCY AND WRONG SIDEBAND IN ARECIBO FITS-HEADERS
WRONG SIDEBAND IN ARECIBO WAPP PULSAR DATA HEADERS

Data affected: all previous FITS-files up to 2007-Feb-15
               all previous WAPP-files up to 2007-Feb-15
========================================================

A bug affecting the reported center frequency and sideband selection
has been found and corrected.

Note that this bug only affects the values of the header parameters -
the data itself has been taken at the correct frequency and sideband.

The effect of the bug has been that the reported center frequency
(FITS-header keyword CRVAL1) for some set-ups have been wrong. It is
also possible that the wrong sideband is indicated (FITS-header
keywords UPPERSB and CDELT1).

Since the same procedure is used to write the header of the WAPP
pulsar-files, they can also be affected and have the wrong sideband
reported (the IFLO_FLIP flag).

This problem affects FITS-files and WAPP-files generated up until 15
February 2007, although it does not affect all files since it only
appears for certain frequency set-ups.

A corrected algorithm was installed 15 February 2007. The version
number of the CIMAFITS-files has been updated from '1.05' to '1.06' to
reflect this correction.

A check of CRVAL1 can be made to see if a file is affected by the
bug. Incorrect files typically have a CRVAL1 value several hundred MHz
offset from the requested center frequency. Since CRVAL1 is used by
all data reduction packages to determine the frequency scale, just
plotting a spectrum from a file and checking whether the frequency
axis is correct or not will tell you if the file is affected. Note
that this check has to be made on spectra from each individual WAPP
for non-ALFA observations, since different frequency set-up for
different WAPPs may cause one WAPP to be affected while another one is
not.


DETAILED DESCRIPTION
--------------------

AFFECTED KEYWORDS:

The affected keywords are:

CRVAL1             (center frequency)
CDELT1             (frequency interval between pixels)
UPPERSB            (indication if spectrum is flipped)

The CRVAL1 value is typically off by several hundred MHz in files
affected by the bug. During tests offsets between 30 MHz and 600 MHz
from the true center frequency have been seen.

The bug is caused by a combination of selected center frequency,
selected center frequency for each WAPP and the Doppler correction
applied. The problem is not random in nature, so a given configuration
will either always be OK or always be wrong.

For single-pixel receiver observations with the WAPPs where the
different WAPPs have been set to different frequencies, the files may
be partly affected in the sense that one or more or the WAPPs may be
affected by the bug while one or more of the other WAPPs are not,
since the presence of the bug for each WAPP is depending on the
individual frequency set-up of that WAPP.

The bug seems to affect the reported sideband much more rarely. If the
wrong sideband is reported, then UPPERSB will have the opposite value
('0' instead of '1' or vice versa) and CDELT1 will have the wrong
sign. The absolute value of CDELT1 is always correct since it is the
sign of CDELT1 that carries the sideband information.

A problem with CDELT1 and UPPERSB may only occur if CRVAL1 has an
incorrect value. If CRVAL1 is OK, then CDELT1 and UPPERSB should also
be OK. If CRVAL1 is wrong, there is a possibility that CDELT1 and
UPPERSB may be wrong.

The same procedure used to determine the center frequency and the
sideband is also used to write the headers of WAPP pulsar-files. While
the center frequency is not used, the sideband flag is and it goes
into the IFLO_FLIP flag.


					Mikael Lerner

					Arecibo, 15 February 2007