Autocorrelation spectra plots

At the end of every scan, the postob procedure will run a script to extract some Mark5 data and display the autocorrelations of the 16 channels. Good data should contain quite flat bandpasses and zero phase. For example:

If the spectra look like this, then everything is OK, dismiss the gv window. The DBBC will sometimes drop channels, resulting in zero amplitude autocorrelations, or may show strange spectra with a step in the middle, like this:

In these cases, it is necessary to reprogram the DBBC. Instructions are here

Please make a note in the log when you reprogram the DBBC. Note that previous spectra plots are kept on pcfsyg in /home/oper/auto_cor_test so it's possible to view these and see when problems occurred. To view these files, use gv, e.g.

gv 2010_11_22_024630.epi

For an AUST experiment the filters make the channel spectra look like this example from hb 12m:

Sometimes, a spectrum will look like this:

This appears to be a bug in the extraction/plotting software (it happens when the software extracts no data) and can be safely ignored if it's a one-off error. Wait for the end of the next scan to see the next plot - if it's still corrupted, lease check to make sure data is being recorded.