Checks during the observations¶
$ : commands to insert in a shell
> : commands to insert in the operatorInput panel
: check the execution on the monitor
Warning
If LEAPCONTROLLER is not working properly, please check the manual procedure for the ROACH1 at L or P band.
Data quality¶
In the leap0 VNC, a window with 8 tabs, each for one node, should be open (W3). Go to W3 to check on data quality. Each tab is connected to a cluster node, i.e. a particular 16 MHz sub-band. In each node, go to the data directory:
$ cd /data-001
for node1$ cd /data-002
for node2, etc…In each node, check that the right folders (source and timestamp) are being created and that “dada” files are being written every 10 seconds. If no dada files are being written, there could be something wrong with the IP table. For example, if one node is malfunctioning (e.g. is turned off) but is listed in the IP table as one would expect, it will cause packet loss or possibly even not write any dada files in the other nodes. In that case, check that all nodes are functioning correctly (see daemons in W2). If the node is definitely not working, you will need to ask a qualified person to look into it [this will involve taking the node out in the IP table before re-starting data acquisition].
Check the data quality with
digistat
. Type:$ digistat *dada
At the prompt, type
/xs
or/xw
thenENTER
The curve on the right should be a nice Gaussian. If not, something went wrong with the receiver or with the cluster. Check that we are indeed observing L-band.
To exit the plotting, just press
CONTROL-C
.Check the bandpass using
passband
. Type:$ passband *dada
(then/xs
or/xw
)Check for any strange source of RFI. Check in particular the bandpass of node 06. In between 1420 and 1421 MHz you should see the HI line. If it is, the cluster is working properly, giving you the right frequencies.
In addition, you should see familiar sources of RFI. In node 01, you should see the Total Power line at 1375 MHz.
At the end of the session, in each node, check for any packet loss with:
$ ~/roach/scripts/packetloss.sh
It counts how many times the number is not zero, meaning there has been packet loss.
Occasional packet loss is nothing to worry about. However if the numbers are completely random and the cluster is losing packets in all nodes, we’re in trouble [ask a qualified person to go and manually reset the LEAP switch while data acquisition is ongoing].
External client¶
If something goes wrong with the external client (e.g. failing to give you antenna parameters), you should kill (using
Control-C
) the control.csh code.Then go see what is wrong with the External Client. If you manage to fix it, re-start:
./control.csh
.Failure of the external client does not affect data acquisition with the ROACH while the antenna is tracking. The ROACH keeps recording data. The external client only matters when the antenna is going from TRACKING to SLEWING or from SLEWING to TRACKING.
If the external client is not working at all and you need to manually launch data acquisition (i.e. you cannot use the
./control.csh
script that is normally used during LEAP to automate data acquisition), you can do the following:- at the beginning of the session, set up the ROACH using:
./control_init.csh
(W1) - Use
./start_simple.csh
to launch data acquisition (instead of./start.csh
) (W1) - Use
./stop.csh
to stop data acquisition without closing the daemons (W1) - Use
./start_simple.csh
for next source etc. (W1) - at the end of the session, use
./end.csh
to stop data acquisition and close the daemons (W1)
- at the beginning of the session, set up the ROACH using:
The difference between ./start.csh
and ./start_simple.csh
is that ./start_simple.csh
does not call the external client to get information about the source. If the external client is working, use ./start.csh
and the right folders will be automatically created using the name of the source that is being tracked and the most recent timestamp.
On DISCOS-CONSOLE¶
ACS custom logging client
The logging display shows warning and error messages. Warning messages are indicated in yellow while error messages are in red Fig.8.
Check the possible error messages. Try to understand the origin of the problem and to solve it. In case of persistent/complex problem, contact and report the error messages and the associated UT to the responsible of the observation.
Meteo client
Check the wind speed on the Meteo client. If it exceeds 60 km/h, the antenna is automatically stowed. Fig.43
Scheduler
Check the status of the Scheduler is OK (green). During the tracking, @ is green while it is red during the slewing of the antenna. Fig.21
Check the update of the number of scan/subscan according to your schedule (.scd). Scans will be skipped if the target is not visible at the moment of the observations.
If you realize that the scan/subscan number is frozen or that the tracking is red while the antenna is tracking the source, stop the on-going schedule with:
> stopSchedule
(in the operatorInput)then start again the schedule:
> startSchedule=[projectID]/…scd,n
(with n the number of scan)MinorServo
Check the status and the tracking are green Fig.24.
ReceiverBoss
Check the status is OK (green). Fig.19
AntennaBoss
Check the status is OK (green), the coordinates of the source and the beam size value are correct (FWHM in degrees). The tracking @ is green when the source is correctly pointed at (within 1/10 of the beam), while a red circle appears when the telescope is slewing. Fig.12
Mount
Check the commanded azimuth and elevation of the antenna and its real position (in blue). The status must be green.
srt_mount_ok
Active surface
Check that the state of the active surface corresponds to your choice (shaped, shaped fixed, parabolic, parabolic fixed). Fig.4
“OK” should be green during the observations.
Sometimes, not all of the small squares of the active surface are green. Do not worry for that if they are spread randomly. Instead, it can be problematic if a large fraction (a whole sector) of the active surface becomes red, in particular in K-band (see Fig.42). Contact the responsible of the observation (observer friend).