Current Status of Amplitude Calibration in the EVN.
---------------------------------------------------

The following table shows the median absolute amplitude error for EVN stations
in 2005. These results are derived from the pipeline amplitude self-calibration
results of all EVN experiments recorded in sessions 1 and 2 of 2005 which have
been pipelined to date. The number in brackets after each entry is the number
of experiments that were used to determine the median error for that entry. 


Station   18cm        21 cm         C band      K band
Cm        0.22 (3)                         
Ef        0.08 (10)   0.10 (2)      0.11 (7)    0.19 (3)
Hh        0.23 (2)                  0.10 (3)    
Jb        0.20 (10)   0.29 (2)      0.42 (7)    0.67 (3)
Mc        0.22 (10)   0.26 (2)      0.05 (6)    0.15 (3)
Mh                                              0.27 (3)
Nt        0.21 (3)    0.27 (1)                  0.21 (2)
On        0.10 (8)    0.09 (1)      0.09 (6)    0.90 (2)
Sh        0.18 (4)                              0.22 (1)
Tr        0.33 (9)    0.51 (1)      0.13 (5)    
Ur        0.24 (5)    2.20 (1)      0.18 (4)    0.41 (1)
Wb        0.10 (10)   0.18 (2)      0.13 (6)

The blank entries are where insufficient data were available. Note that the
numbers above are the median absolute error in the SEFD, and so are
significantly higher than the rms error. A value below 0.1 can be considered to
indicate good calibration, while values above 0.2 probably indicate a
significant error. For comparison, the VLBA stations participating in GB055A
(C-band, session 3 of 2005) had a median SEFD error of 0.097. 

Points of note:

Results are very similar to those reported previously. C band calibration is
generally reasonable, but Nt and Ur continue to have some problems. 
Jb has developed a severe problem in 2005 (figures are much worse than
previously). It appears that the Mk2 had a sensitivity problem in the
October session, which was not reflected in the Tsys, and the RXG file for the
Mk2 was not updated in the June session.

L-band is a continuing problem which may be largely ascribed to RFI. The recent
move to wider bandwidth observing (with the consequent use of frequencies in an
even less friendly RFI environment) has exacerbated this. Note however that the
calibration accuracy varies enormously even for experiments with similar
setups. For example, GD018 from session 1 2005 had a median error of 0.25,
while GB055B (session 3 2005) had a median error of 0.08. The frequency
coverage in the two experiments were identical and UT ranges were similar
though the RFI did appear to be worse in GD018.

K band calibration is also a problem. Opacity effects, poor sensitivity and
pointing errors are all possible contributing factors. An investigation of
K-band calibration in the February session showed that the gain errors are not
stable and so it is not possible to transfer error estimates derived from one
experiment (e.g. the NME) to other experiments in the same session. Onsala have
implemented a new high-frequency (> 15 GHz) calibration scheme which uses a hot
load and chopper wheel instead of the conventional calibration diode. This was
used for the first time in the October 2005 session, and results from EM059A
(the only user experiment at K band) indicate that the calibration has
improved somewhat.

ANTAB Scripts
--------------------------------------
Most stations routinely produce their own ANTAB files using the 'antabfs'
scripts. A recent release of the scripts offers new support for the Onsala
high-frequency calibration scheme using a hot load and chopper wheel. An
'extended ANTAB' format which is compatible with the current AIPS ANTAB, but
which will also ease the use of the ANTAB tables with EVN correlator and
geodetic software has been defined. 

UVFLG
--------------------------------------
Stations continue to provide off-source monitoring data which allow data taken
when the antenna was off-source to be flagged. Since the beginning of 2005 this
scheme has been implemented at Noto where it also works well. Westerbork have
also introduced the capability to monitor the off-source status of the array
and to write output to the FS logs. Efforts are continuing to integrate this
fully into the existing EVN flagging system. 

Cormac Reynolds.