This is an old revision of the document!
The following assumes that a module (or modules) have already been loaded at the station.
From within the pcfs[hb|ke|yg] VNC session log in to Mark5 you want to set up (either mk5hb, mk5ke or mk5yg) as oper. Do this so that the Mark5 software will continue to run even if the network link to the telescope goes down. e.g.
ssh oper@mk5hb
First check if the dimino program is running:
ps -ef | grep dim
If it is, we want to re-start it so that output is logged to the correct place. Stop it with the command
Enddim
Then start it and give the experiment code and station ID as the argument. E.g. for r1456 at Yarragadee:
cd ~ ./dim r1456yg
Some messages will appear, ending with a message reporting dimino is ready:
201222802:40:54 dimino Ready. End with Enddim, please
The Mark5 gets it's system time from a local GPS receiver which runs an NTP server. You can check if it is synchronised with the command
ntpq -np
Look for an offset (reported in milli-sec) from the the first server listed of about 10 ms or less 1). If there is a large offset, the Mark5 system clock may need resetting. Contact the on-call person for assistance.
The following steps are done through e-remote Control or the Field System.
First set up the Mark5 recording mode. Use the experiment procedure file:
proc=<experiment_name><station_ID>
e.g.
proc=r1473hb
then type
ready_disk
Then one of the following depending on the experiment type.
setupsx
setup8f
This will configure the Mark5 for the experiment.
Now check that the times are all OK:
mk5=dot?
look for syncerr_eq_0
, and that the final number is less than ~5 ms.
If you see something like syncerr_gt_3
then fmset
may need to be run to synchronise the clocks. To do this, open a terminal on pcfs[hb|ke|yg], and type
fmset
When the interface appears, type
s
then confirm with a
y
Following this, the clocks should all be synchronised and agree with UT. Press the [Esc] key to exit fmset
. Then check the times again:
mk5=dot?
The fmset program can be used to confirm that the Field system time agrees with the station time. Open a terminal on pcfs[hb|ke|yg], and type
fmset
The Field System and Computer times should agree.
Now check the maser-GPS and delay through the DBBC and Mark5 (Formatter minus GPS) with the
maserdelay
and
clkoff
commands respectively which read the Agilent counters. Log these values in the checklist. They should bethe same to within 0.5 microsec. The Log Monitor should report the difference whuch should be less than +/- 0.5 us.
Issue the maserdelay
and clkoff
commands a few times to check that they aren't drifting. A drift in the formatter-GPS time is usually an indication of a DBBC problem, usually requiring a reconfig to fix.
If one or both counters are not responding to these commands, you will see an error message like this
ERROR sp -4 GPIB Device time-out on response c2
To fix this, try re-setting the counters. In the Operator Input window, type
counter
and try the maserdelay
and clkoff
commands again.
Now make a test recording using the Mark5. This is done through the Operator Input window.
disk_record=on mk5=dot? (Output should appear as before, but it should say FHG_on) disk_record=off
Now check the data are OK:
scan_check
Check the last numbers are the correct date at the start of the scan, duration, the data rate (should be 256 for the 8MHz mode, 128 for the 4 MHz mode) and 0 at the end for no problems. Then plot an autocorrelation from the data:
postob
The output should appear in the pcfs[hb|ke|yg] VNC session. Good data should contain quite flat bandpasses and zero phase. See autocorrelation spectra plots for an example and what problems to look out for. If there's a problem, the DBBC may need reconfiguring.
To clear the disk module, ONLY IF YOU'RE CERTAIN there is no data already on the disk!
mk5=protect=off; mk5=reset=erase
Check that the spectra are OK. If there are dropped or strange-looking bands, the DBBC may need reprogramming.