Table of contents
- 1. Prerequisits
- 2. System initialization
- 3.
- 4.
- 5.
Prerequisits
- Go through the system setup procedure for the DBBC3 and verify that the system is in a working condition.
System initialization
Additional information can be found also on the EHT-wiki
Module initialization
ONLY IF REQUIRED: Initialise the modules using the following command. It has to be repeated separately for each Mark6 that has modules that need initialising (example below references the Mark6 using hostname 'recorder1
').
This command will erase all existing data on the modules.
If you are unsure whether to initialise a set of modules, request guidance from AOC.
backendctl mark6 recorder1 modules 1,2,3,4 init-fresh
if the modules are still in "open" state they must be unmounted before
backendctl mark6 recorder1 group unmount
Repeat for all recorders
Initialize, configure & validate the DBBC3
Load the OCT_D firmware
On the DBBC3 desktop
- close any other running control software programs
- close the DBBC3 client program
- double click the icon labeled "DBBC3 Control OCT_D_v120.exe"
- answer first question with "y" in order to do a full reload of the firmware.
- wait until the control software has fully loaded and responds with "Waiting for connection on port 4000"
Setup the system
Verify that the setup for using 2GHz filters is activated:
- Inspect c:\DBBC_CONF\OCT_D_120\dbbc3_config_file_oct_D_120.txt
- Check that the 2GHz version of the fila10g files is being referenced, e.g. oct_D_2GHz_core3H_1.fila10g. If you find a reference to e.g. 1GHz setups you need to change the setup by following the instructions in the README file located in the c:\DBBC_CONF\OCT_D_120\ folder. In case
Note: the target setup for the DBBC3 is defined in /etc/backend.conf.
The default setup is valid for 230 and 86 GHz. For switching between 230 and 345 GHz the following changes need to be made to the setup.
- make sure the DBBC3 client is not running
- configure the DBBC3 using backendctl(on the EHT Control Computer cc-pico):
backendctl dbbc3 dbbc3 configure
- check for any errors
Validate the system
- make sure the DBBC3 client is not running
- validate the DBBC3 using backendctl(on the EHT Control Computer cc-pico):
backendctl dbbc3 dbbc3 check
- check for any errors
Check time synchronisation
Time syncronisation can be checked with the tick command via the serial interface.
Follow these steps below exactly. Omitting any step will lead to mal-functioning and will require to completely reload the firmware.
On the DBBC3 desktop:
- double-click the putty icon
- in putty open connection e.g. to DBBC3 Board A
- in the window hit enter to get to the command prompt and execute:
- tick
- compare the timestamps to a radio-controlled clock
- when done hit enter to stop the tick command
- close the putty window
Validate the VLBI System (Except DBBC3)
on the EHT control computer run:
backendctl whole check
This will check the setup of the control computer and the recorders. The check of the DBBC3 is not yet included in this procedure (see above).
Adjust power levels (DBBC3)
Basically low/high power levels should have been reported by setup script (see above).
In DBBC3 client e.g. on windows desktop or
on the control computer:
/home/oper/rottmann/dbbc3/utilities/dbbc3client.py dbbc3
check attenuators, e.g. for board A:
dbbcifa
attenuator settings should be within 20-40, agc should be on
if reported attenuator level is out of range 20-40 the IF power must be decreased/increased.
Do test recording
backendctl mark6 all run test-recording 20 30
Recording starts with a delay of 20 seconds. Visually check if all recorders are actually recording.
Record & plot
log into the recorder e.g. recorder1
ssh -Y recorder1
execute:
plotdbbc3_m6.sh
This will do a short test recording and plot the resulting spectrum in both polarizations
Load and execute the schedule
Scedules are located under /srv/vexstore
load the schedule that has been triggered by the AOC:
backendctl mark6 all schedule load trigger
Follow the schedule:
backendctl whole schedule follow trigger
Start the Mark6 monitoring client
copy the vex file (e.g. from /srv/vexstore/trigger) to /home/oper/shared/schedules
vex2xml.py -f {vexfile} -s Pv
check the contents of the generated {schedule}.xml if it contains scans
m6schedulemon.py recorder1 {schedule}.xml &
repeat for all recorders you want to monitor
Comments