Processing CONTOBS output
Note
Check gixwrite.pl for an automated alternative as VAX aka blazar isn't part of the current observation procedure.
Getting the output files
- Type "ftp blazar".
- Enter username and password.
- Type "cd dub1:<kurpitsa>".
- Type "mget YYMMDD.GIX;*" (where YYMMDD is the date of the observations).
- Type "exit".
Creating .dat and .dr21 files
- Type "read_contobs_data.pl" (run the program on antimatter or some other
computer that has /data mounted).
- Enter input file, meaning the GIX-file, which was transfered from blazar.
- Enter output file, that is, the desired .dat file.
- Alternatively, you can start the program by typing:
"read_contobs_data.pl inputfile outputfile".
- Monitor output to see if errors occur (you can run this program under emacs
shell to make it easier; command "Esc-x shell").
- Check the output file if the results are OK. "Alert" means very short
integration times or high rms values, check them and don't accept any
observations that are of poor quality (also others than the ones marked
with the alert!), i.e. there should be no such data left in the file that
you would discard from the log.
- Insert tau and cal values by hand. The values of tau vary from 0.045
(clear and dry) to 0.05 (wet and cloudy) in winter and, respectively, from
0.050 to 0.060 in summer. Choose a suitable value according to the weather.
- Remove the header line, which begins with #.
- Copy & paste DR21 observations into a .dr21 file.
- Copy both files to /home/your_home_directory/GIX directory.
- Now the both files should have with the following columns in the following
order: date, time, source, azimuth, elevation, azimuth offset error, elevation
offset error, result, rms, Tant, Tout, RelH, tau, cal.
- In the case of one-point source, the value of azimuth offset error and
elevation offset error is "1p" (without quotation marks).
- You can, of course, write the files by hand, but make sure that you include
every (tab-separated) column mentioned above.
Notes about read_contobs_data.pl
- There are still some bugs in the program and some of them are listed here:
- if you observe the same source many times in a row at 1p (without doing
anything else in between), the program can't separate these observations
and reads the results from a wrong row. It has been fixed so, that there is
also an "intermediate result" in the stdout, and you can use that instead.
- If you stop the observation with ctrl y instead of 0, the results can't be
written to a file. So, always end the observation with 0 before exiting the
program.
- There is also problems with 5p observations, if contobs can't get a fit from
the last round. There is nothing to do about that, but maybe use the results
from the earlier round, or checking the original vax file for results.
- When using this program to write results from an autofile, please write the
results also to the log book.
Last modified: Fri Dec 19 16:14:12 EET 2008
/ maturun