Owen's Notes

From EOVSA Wiki
Jump to navigation Jump to search

Routine Tasks

Daily Tasks

Check the following pages:

http://ovsa.njit.edu/status.php (the all-day plot from the previous day is at the bottom of the page)

http://www.ovsa.njit.edu/browser/ (this defaults to showing pipeline output from two days previous)

http://ovsa.njit.edu/pointing/ (the latest pointing plot(s) at the bottom)

http://ovsa.njit.edu/phasecal/ (the latest phasecal vs. frequency is somewhere in the middle, i.e. the bottom of the list of pcf*.png files).

http://ovsa.njit.edu/EOVSA/ant_status.php (Current antenna status).

Ensure that the schedule is running and that the stateframe is updating.

Check that the data files are being updated. Goto the user@dpp terminal and type:

ls /data1/IDB |tail


Perform a Phase Calibration

Check that the network monitoring function (fix_packets()) is working on the DPP. It can also be seen as dropouts on the spectrogram (http://ovsa.njit.edu/status.php).

If there are any problems with any of the antennas goto http://ovsa.njit.edu/EOVSA/ant_status/ant_status_form.php and fill out the appropriate form.

Weekly Tasks

Perform a Gain Calibration

Monthly Tasks

Perform a Pointing Calibration

Remote Login (23 March 2020)

Helios Login

The computers can be accessed remotely using SSH and VNC. Since I use Linux (Ubuntu) these instruction are for that OS.

To VNC into Helios, open a terminal and type the following command:

ssh -L 5902:helios.solar.pvt:20000 oweng@ovsa.njit.edu

I then open Remmina Remote Desktop Client. Make sure VNC is selected at the top left. Type in:

localhost:2

and enter the password when prompted.

The Helios display should now be visible.

Occasionally, the VNC connection will be lost. If this occurs the following procedure need to be performed.

From a new terminal window type:

ssh -L 8887:helios.solar.pvt:22 oweng@ovsa.njit.edu

From another terminal issue the command:

ssh -p 8887 sched@localhost

Now type in:

x11go

Now restart the VNC client.

Pipeline login

To log in to pipeline, from a terminal enter the command:

ssh -L 8888:pipeline.solar.pvt:22 oweng@ovsa.njit.edu

Then from a new terminal enter:

ssh -p 8888 user@localhost

Win1 login

To log in to the win1 machine, from a terminal enter the command:

ssh -L 8889:win1.solar.pvt:5900 oweng@ovsa.njit.edu

Then from a new VPN client enter the hostname:

localhost:8889

Phase Calibrations (23 March 2020)

Phase calibrations are to be performed every day when possible.

Reference Calibration

Before the phase calibration can be done, a reference calibration needs to e performed, which will be subsequently used as the reference for the phase calibration.

Access the Helios computer.

On the terminal window make sure that the ipython: /home/user tab is selected.

There should be a prompt as follows: pipeline:~>

At the prompt type the following command:

python /common/python/current/calwidget.py

The following window will open up:

The calwidget window.

In the date box near top right of the window type in or use the up and down arrows to select the previous days date and hit enter.

Click on the first available scan that does not suffer from a Windscram. Windscrams are marked in red. If a windscram occurred for less than 20% of the scan then it will be marked in yellow and may still be usable. The example below shows a windscram at 22:14:05.

The calwidget window with windscram shown at 22:14:05 highlighted in red.

Select the earliest calibration that does not have a windscram, and then click on 'Analyze as Refcal'. The Sigma map should now be populated as shown below:

The calwidget window with sigma map. It also shows power and phases for Antenna 1 and Frequency Band 12.

The above chart is for the low-band (channels 1 to 12). Yellow blocks indicate bad or unavailable data. Bands 3 and 4 are excluded due to intense RFI at these frequencies. Antenna 5 currently is not operating well, and will be excluded from any analysis. The green blocks indicate that the data may be alright if some points are excluded. Purple blocks indicate the data is good.

Also shown in the above plot is the power and phase for Antenna 1 and band 12. The blue points are the X-polarization and the orange points are the Y-polarization. These should be close to straight lines.

In this case we need to exclude all of Antenna 5 from the calibration. First click on any block for antenna 5 in the range of band 1 to 12. Next check the box 'Apply to all bands'. Move the mouse over to either the Power or Phase plots (either works). We need to exclude all points. Move the mouse slightly before the first point in the chosen plot and press the 'A' key. A green line will appear. next move the mouse slightly past the last point and press the 'B' key. A red line will appear. The screen should look like the figure below.

The calwidget window with antenna 5 having all points flagged for all bands.

Now click on 'Apply time flagging. All of antenna 5's bands should turn yellow.

Once complete, select the associated high band calibration and then click on 'Analyse as refcal'. In this case bands 7 to 52 will be shown on the sigma map as shown in the figure below.

The calwidget window with sigma map. It also shows power and phases for Antenna 8 and Frequency Band 34.

Antenna 5 once again needs to be excluded using the procedure described above. In addition antenna 13 has problems from band 13 to 52. We will remove these channels from the analysis. Ensure 'Apply to all bands is unchecked and select 'Apply to all bands above selected one'. Select band 13. As before select all of the points by pressing 'a' and 'b' at the appropriate points (start and end of the plot). Click on 'Apply Time Flagging'. All of the bands from the selected band up should turn yellow as shown in the figure below.

The calwidget window with Antenna 5 excluded and most of Antenna 13 excluded.

As can be seen, there are still 3 green squares. We can make these good by selecting each one and flagging bad samples. In this example I have selected Antenna 4, Band 23.

Ensure all checkboxes are unchecked. Slelect the box for the antenna and band the you wish to time flag. select (using mouse and a and b keys) up to two ranges of points to exclude. pressing the x key will remove the most resent line chosen. Once you have flagged the ranges, click on 'Apply Time Flagging'. Hopefully the box will turn purple. If not you may need to select different ranges.

The result of this procedure is shown below.

The result of time flagging an individual antenna and band. Antenna 4, Band 13 is now purple.

Repeat this procedure for the remaining bad antennas and bands.

Once completed, select one of the bands of the refcals from list on the left and then click on 'Set as Refcal'. The click on extend selection. NOw select the other refcal. Make sure both are highlighted. Click on 'Set as extended refcal'. Click on 'Save to SQL'.

Phase Calibration

After the reference calibration has been done, the phase calibration can be performed. Click on a good calibration for the high band (one that has not had a windscram and has a reasonable duration). Not that Phase Cals are not performed using the low band.

Then click on 'Analyze as Phasecal'. At the top of the screen click on the 'Sum Pha' tab. The important thing here is that the slopes look good. Below is an example display of the slopes. These slopes are with reference to the Refcal.

The calwidget window showing slopes for each antenna.

In the above display, we can see that all of the slopes look good except for Antenna 4. note that we have excluded Antenna 5 from the observations as these have no reference cal. By moving the mouse over points on antenna 4 the channel is displayed as the x value in the bottom right of the window. In this case, channel 24 and above will be excluded. Ensure 'Apply to all bands above selected one' is checked and click on the antenna 4 and channel 24 box. Select all points and click on 'Apply time flagging' Once done, click on the 'Sum Pha' tab again to ensure the slope now looks okay (See figure below). Once satisfied click on 'Save to SQL'.

The calwidget window showing slopes for each antenna with points excluded on Antenna 4.

System Gain Calibration (30 April 2020)

This procedure is designed to adjust the system gains. This should be performed weekly or whenever the calibrations become poor. The full details can be found at EOVSA System Gain Calibration. The steps below are a summary of that page.

The best time to perform this calibration is immediately after a Solar Point Calibration, as the antennas need to be taken off the sun.

The dBm values in the power and attenuation section of the Stateframe display should be between 1 and 4.

Initially click on the Stop button in the schedule window.

From the Schedule window issue the following commands:

$scan-stop

stow

femauto-off

femattn 0

$fem-init

Wait approximately 30s for the new settings to update. Then proceed with:

dcmauto-off ant1-14

$capture-1s dcm

Click on the Clear button and then the Go button.

From a pipeline terminal issue the following commands

ipython --pylab

import roachcal

tbl=roachcal.DCM_calnew('/dppdata/PRT/PRTyyyymmddhhmmssdcm.dat',dcmattn=10,missing='ant15')

In the above command yyyymmddhhmmss is the year, month, day, hour minute and second of the dcm file. Pressing tab will complete the filename. for the missing field, antenna 15 must be included. Add any other antennas that are not to be included on the gain calibration separated by spaces.

Enter

tbl

to view the contents of the tbl variable that was just created.

Band 1 has recently had bad attenuations. To use older good values for this band issue the command:

tbl = roachcal.override(tbl, bandlist=[1])

Other bands can be excluded by adding the bands to the bandlist (space separated).

It is useful to compare the new table with an existing table to check for errors/glitches. To do this issue the command:

roachcal.compare_tbl(tbl, t=Time('2020-05-08'))

Replace the time with the previous days gain calibration. All of the differences should be close to zero.

Finally:

import cal_header

cal_header.dcm_master_tablesql(tbl)

From the Schedule window click the Stop button and then the Go button.

In the schedule command box enter the command

$scan-start

Update Antenna Pointing (17 March 2020)

Access the Pipeline Computer.

From the terminal, start ipython:

ipython --pylab

Import the calibration and Time libraries:

import calibration as cal

from util import Time

Load up the solar pointing scans for a given date and time.

x,y,qual=cal.solpntanal(Time('2020-03-17 18:30'))

The above command will load the scans from the 17 March 2020 18:30 UT.

To view the beam widths as a function of frequency for each antenna and polarisation:

cal.sp_bsize(x,y)

An example of the X-beam widths for each antenna
An example of the Y-beam widths for each antenna

To obtain the pointing offsets:

xoff,yoff,dx,dy=cal.sp_offsets(x,y)

This will display the offsets for each antenna and feed. Ideally the offsets should be at zero. If particular antennas significantly deviate from zero then a correction can be made.

For example, the following command will adjust the pointing offsets for antennas 6 to 8, 10 and 13:

cal.offsets2ants(Time('2020-03-17 18:30'),xoff,yoff,ant_str='ant6 ant7 ant8 ant10 ant13')

The newer antennas with the alt-az drives that have had tracking updated will need to be rebooted. From the Schedule Window issue the commands:

reboot 1 ant6-8

tracktable sun_tab.radec 1 ant6-8

track ant6-8

The antenna pointing adjustment is typically performed once per month.

Helios Displays

The Schedule window shows the currently running schedule as shown below.

The Schedule Window

The left side of the window shows each of the scripts that is to be run. The orange bar shows the task that is currently being run.

The right side of the window shows each command that is to be run in the current script.

At the bottom of the window are a series of buttons. The STOP button will stop at the current script. Once stopped, the remaining buttons will become active.

Antenna Trouble Shooting (2 April 2020)

If an antenna in the State-frame Antenna Tracking section is highlighted red, it will need to be restarted.

In the state-frame display, click on the antenna display and then click on the tab for the appropriate antenna.

One or both of the Permit boxes will probably be red. The left column is for azimuth and the center column is for elevation. The procedure for Az and El is slightly different.

Access the windows computer.

Click on then Quick launch icon at bottom right and then click on the antenna that is problematic.

Click the Enable checkbox. The Drives On, Operate, Energized and Break off should all be green.

If the Permit lights are red the click Reboot System.

The current time should be updating about every second (there may be brief pauses).

Click Exit.

If the Elevation Permit fails to become green then access the following page (make sure to replace vik5 with the appropriate antenna):

http://vik5.solar.pvt

Turn the Antenna Relay on and wait for 10 seconds.

Turn the Relay off.


If the antenna still will not start goto the following webpage (make sure to replace ant5 with the appropriate antenna):

http://ant5.solar.pvt

Click Login and then enter the username and password.

Click Parameters and the change Option 10.00 to 1070 and Option 10.38 to 100. Click on change.

Then click on logout.

After the above procedures are completed you will need to reload the track table. This is done from the State Frame display using the following commands (again ensure that ant5 is replaced with the appropriate antenna):

tracktable pcal_tab.radec ant5

track ant5

Antenna Communications Failure

If an antenna is highlighted in the Communications section of the State-Frame display then the following procedure can be used to correct it.

Access the win1 computer.


Click on Start then select NI Max. Goto Remote Settings and then select the CRIO that is not communicating.

Click the System Settings tab at the bottom of the window. Status should be running.

Click on Refresh to reconnect.

The CRIO will have an IP address in the range of 192.168.24.41 to 192.168.24.53. If it is not working this IP address will be strange. Click restart.

If it is not already open, Click Programs then NI Labview 2015.

Computer Restart

Helios

To restart Helios, at a terminal enter the command

sudo reboot now

and wait about 5 minutes for it to restart.

Open a terminal and start the dropbox server:

python /home/sched/Downloads/dropbox.py start

Check to see if the time service is running. Enter the command:

chronyc sourcestats

If nothing is displayed, then the time service will need to be started by typing:

sudo systemctl restart chrony.service

Start the x11vnc server by typing:

x11go

Start the schedule by issuing the following commands:

cd ~/Dropbox/PythonCode/Current

screen

python /common/python/current/schedule.py &

Press <CRTL>AD to exit from screen.

Start the Stateframe display as follows:

screen

python /common/python/current/sf_display.py &

Press <CRTL>AD to exit from screen.

Other Issues

Temperature Controller (30 Apr 2020)

Occassionally a front end temperature will show an error. This is usually because the temperature controller that reports the temperature has locked up. To fix this type the following command in the Schedule window:

tec$bc ant5

Replace ant5 with the antenna(s) that need the temperature controller rebooted.

To clear the temperature history (that updates every minute) issue the command:

tec$sc ant5

Packet Loss on DPP

If the packet monitoring program on the DPP computer shows lines similar to:

2020-05-28 15:43:42.001 0.0 0.0

then data packets are being lost. To fix this problem follow the following steps:

Terminate the fix_packets() program by pressing [CTRL] C.
Type !nano </home/user/test_svn/shell_scripts/SMP_AFFINITY.sh>
Comment out the current cpu's and uncomment the following ones.
Restart the fix_packets() program.
Exit nano and save the file.

Occasionally the glitch will occur after the above procedure has been performed. In this case the DPP will need to be rebooted. The procedure to do that is as follows:

Terminate the fix_packets() program and exit ipython by typing exit
Enter the command sudo reboot now and wait about 5 minutes.
Log back into the DPP by typing ssh -X user@dpp
Enter the command /home/user/test_svn/shell_scripts/AMP_AFFINITY.sh
Enter the command rmlock. This will remove the lock file.
Start ipython by typing ipython --pylab
Restart the fix_packets() program.

Restarting the Schedule

If the schedule needs to be restarted then follow the steps below:

Click Stop and then Clear.
Click Today and then Go.

Pipeline Problems

If there is no data showing at http://www.ovsa.njit.edu/browser/ then something has gone wrong with the pipeline processing. Below is a image of what the webpage looks like when an error has occurred.

The 'No Data' errors on the EOVA browser page.

Clicking on the Newest button will refresh the page and upload the latest data. If this does not correct the problem take a note of the date in the web address of the page. Then click on 'Today's Data'. A list of fits files should appear as below:

Example list of files if pipeline processing was successful.

The log files will need to be inspected in order to dermine what went wrong.

Note: I am in the process of updating scripts, so the diagnostic procedure is not complete - OG

Warning on Power and Attenuation

On the state frame display, occasionally an antenna will turn yellow (warning). This will usually be because the 1st attenuator value is not zero. This will occur during a gaincaltest or a burst. The value should be even. If the value is 9, then the CRIO file will most likely need to be be reloaded.

DO THIS WITH EXTREME CAUTION. BE ABSOLUTELY CERTAIN THAT RELOADING THE CRIO FILE NEEDS TO BE DONE.

From helios, open a terminal and type cd /Dropbox/PythonCode/Current/crio_inis
For the antenna that need to be reset, find the most recent '.ini' file. For example, for antenna 10 this may be 'crio10-2019-09-28.ini'. Make a note of this file.
Enter the command ftp crio10.solar.pvt. Replace crio10 with the apropriate CRIO number.
Enter the username and password when prompted.
Enter the following commands:
cd ni_rt/startup
put crio10-2019-09-28.ini crio.ini
bye
replacing 'crio10-2019-09-28.ini' with the appropriate file for the CRIO that needs to be reset.
The above commands have reloaded the crio.ini file. The CRIO now needs to be reset. For example, Antenna 10 will be reset by issuing the following commands from the schedule window:
sync ant10
tracktable sun_tab.radec ant10
track

The gains and pointing may need to also be updated.

Schedule Commands not Working

If schedule commands are not working then the ACC may need to be restarted.

To determine if the ACC need restarting issue a command from Schedule. At the top right of the Stateframe the 'Task' should be updated. if it is not then follow the following procedure:

Go to the Windows computer and from the Quicklaunch menu select 'EOVSA E&C'.
Once running click on 'Reset ACC' and wait about 10 seconds.

Queue Overflow Error

On the Stateframe display, a 'Queue Overflow' Error may occur under the Frequency Tuning bar. To correct this issue type the following command into the Raw Command box on the schedule:

lo1a-reboot

Stop and then restart the schedule.

Antenna 12 Not Tracking

Antenna 12 will occasionally seem to stop tracking. This is usually because it has lost where it is pointing. It will move incredibly slowly until it reaches its limit before returning to regular tracking. This process cant take up to about 20 minutes. If it has not resolved it tracking problems in this time, proceed with normal trouble shooting.

Useful Tools and Utilities

Screen Command

On linux system, the screen command allows a program to be run in the background. This is particularly useful for processes that take a long time to complete. In addition it will display output when the screen is activated.

To start a screen session, open up a terminal and issue the command:

screen

Text will appear and you will be prompted to press space or return. Just press return.

Type in the command that you want to be performed. The process will then start. You can exit the screen by pressing

<CRTL> A D

The command will still be running. To re-open the screen at any time simply type:

screen -r