User Tools

Site Tools


chara:trouble_shooting

Trouble Shooting


This document describes how to solve common problems that sometimes arise at the array.

Can't find fringes

This is a list of simple things to check if you can't find fringes:

  • Were the clocks synced? Make sure the [SYNC CLOCKS] button on Cosmic Debris has been pushed to start the night. If the OPLE server does not display the correct CHARA time and the errors don't read (0) or (1), the clocks were not synced.
  • Did the Astrolib update on OPLE? If the job queue is stopped too soon after slewing on Cosmic Debris, the correct calculations for the carts will not be done by OPLE and you may be searching for fringes with the wrong star data. The star identifier will be displayed in the server window. If it is not correct, hit STAR ACQUIRED on CD to update OPLE. The proper star can also be entered manually by typing hd #### into the OPLE server and hitting ENTER.
  • Are the PoP's correct? After a PoP change, the PoP's are sometimes not updated in CD or ople. Compare the PoP's in the configuration tab of CD or OPLE with the PoP Overview or Popperi gtk.
  • Are the carts behaving or are there vibrations or jumps of 100 or more microns every 3-6 seconds? How are the metrology signals? Are they strong and staying white? Red signals mean one or more metrology signals may have gone too low and homing carts is necessary. The METDATA function on the ople gui can be used to sample the metrology signal and may show noise spikes that can be disruptive to the smooth cart motion. Push the METDATA button on the configure tab and then again about 6 seconds later. Plots will pop up to show the frequency and power of any noise.
  • Is the target a high proper motion star? Red dwarfs are close stars and can have high proper motions. Scan a wider range to see if it is outside of the usual calculated scan range. Binaries can also have very high offsets from the expected position due to mistakenly using astromod calculations from the companion star.
  • Do you have enough flux from each telescope or on each baseline? Is the telescope tiptilt struggling with low flux? Can the camera gain be raised or the exposure made longer to help hold the star? Passing clouds or contrails can lower flux unexpectedly.
  • Did you get the same star in each telescope? Sometimes a busy star field and poor pointing of the telescopes can lead to the wrong star being acquired and locked by tiptilt. View the stars in the finder window to see if all the stars match.
  • Check the CHARA time on the GPS server. The “Ext-CHARA,” “CHARA-Sys,” and “Ext-Sys” time offsets listed on the GPS server should be small (< 0.01 sec). If there are large time offsets, then a GSYNC might be needed.
  • Check the time on the OPLE server. If the time is off or there are any lost ticks/seconds “Lost T/S” listed, then MSYNC ople using the GPS GUI and type “syncople” into the ople server.
  • Are the [MAN] buttons pressed (gray) for the moving carts on the OPLE Control gui? (The reference cart will remain green.)
  • Check that the carts are within delay line range (-1.3 to 44.25 meters) and errors are small. Were the carts homed and checked before the first slew of the night? Did any carts go to the front switch after slewing? This can cause them to lose their position. Turn off the OL button to send it back to zero. If it does not track on the home switch when displaying a zero position (0.000000), it was lost. Is the E2 cart stuck at 27.97m? See section on homing carts in OPLE and Metrology section below.
  • Are the LDC's working?
    • Is the glass position within allowable range (-10mm to 49mm) on all beams?
    • Have the commands “useldc on” and “autoldc on” been typed into the ople server?
    • Look out for “-10” and “49” errors on ople server. If the LDC glass goes out of range, the LDC will need to be homed. To home the glass, stop astromod by clicking [STOP] on the OPLE gui. Make sure the LDC velocities are non-zero (~ 50… must click [VEL] button to update velocity) and click [HOME] on the LDCs in use. The LDC position number should go to 0.000 or very close to zero. After the LDCs have been homed, turn astromod on by clicking [START] on the ople gui. Then click [REF] on cosmic debris to send the LDCs to the correct positions.
    • If LDCs are working correctly, the LDC positions (in mm) should roughly match up with the reference cart positions (in meters).
  • Check the instrument alignment. Is flux getting through to the detector? How long has it been since the last NIRO camera alignment? Classic and CLIMB programs can run for about an hour before the light will drift from the central pixel. Use the Classic or CLIMB gui to view the light on the pixels by clicking the PICTURE tab and then the PIXEL AREA button. Turn the camera off with the STOP button. Is the right dither power turned on? CLIMB 1 and Classic use different dithers. If Classic or CLIMB fringes are found in a scan, but not when in recording mode, the dither powers are likely not on. Are the camera settings correct for the seeing conditions and flux levels?
  • There is a script that will display offsets for all scopes against Hour Angle to help find offsets when using MIRCX and MYSTIC. You'll need to log into the MIRCX spooler computer with the command ssh spooler@mircx, then use the command fringe_predictor. A window will pop up displaying the last five nights of offsets vs. HA.

The new OPLE system

With the implementation of the new ople system which replaced the VME in Fall of 2021, new troubleshooting issues have arisen. Since there are now 6 new ople computers to run the carts for each scope individually, many issues will be specific to one computer or telescope and not to the system as a whole. Restarting the VME is now a thing of the past.

The traditional ople server will still be used to communicate with each new ople computer, identified as OPLE 1 to OPLE 6. When the communications are good, each active cart will be displayed in the ople server or ople gui status tab as before. At times, an ople computer can lose communications or a server can crash and the server or comms needs to be restarted.

If a cart cannot be started, stopped or otherwise commanded, look to the OPLESystem gui to see if a green indicator has turned red. A message will also often pop up on the ople gui saying the command could not be sent. This may require a simple start command to restart the server or a reboot of the computer to get it back to yellow and then a start to get it back to green. Do either of these steps with a right click of the red button and then select start or reboot. After a reboot, select start to load the servers after the indicator has turned yellow. When this is done, the ople server will need to be connected to the newly restarted server. Type “oo” into the ople server to open ople comms with the new ople server. It should now reappear on the ople server display and say System Ready to indicate comms are restored.

Some times a cart is stopped and cannot be commanded. If the cart has gone to the front hard or back hard switch, it will not be usable until it is moved from the switch and the Ople Controller box is reset. There are 6 silver boxes for these controllers with two green LED's for the front and back switches and two red LED's for the back hard and front hard switches. If a red LED is lit, there will be an error displayed on the message window and the cart is disabled. The cart will need to be moved ofo the switch and then the box can be reset with the RESET button on the front. The error display will go away and the red LED will be off. The cart is now controllable.

Restarting Servers

Restarting Servers using the bootlaunch paradigm

If a server is not running or Socket Manager reports that a server is dead, then look at the socket manager list to find out what computer the server runs on (socket_manager.list). You can also look at the up-to-date file by opening a terminal window and typing “less /ctrscrut/chara/etc/socket_manager/socket_manager.list”.

To restart a server, log on to the machine that runs the server and type “bootlaunch_master”. This script will go through the list of executables and will check which servers are running. If a server isn't running it bootlaunch_master will remove it from socket manager, clear the lock file, and relaunch the server. The instructions below describe how to restart individual servers, but this should not be necessary anymore.

A number of servers use an interim bootlaunch paradigm to restart. This is confined to servers that run on ubuntu machines, namely the telescope bunker computers and gps. The basic syntax is “bootlaunch_<server>” where “<server>” is replaced by the server the script is designed to address. The scripts have a number of safeties built in, so it is safe to run them even if a server is already running – they just output the process ID of the running server. The scripts also take care of the entry in socket manager as well any serial port lock files. All the pertinent information is world writeable, so one should be able to run a bootlaunch script as observe.

One thing of note about the output of the bootlaunch scripts, they call a number of other programs which themselves have output that may be misleading in the context of bootlaunch. Chief among these is the output of “tsockman”. If a server stopped unexpectedly, it may leave behind an entry in the socket manager. In order to launch a new server, one needs to clean out the socket manager entry if it is there. To do that, “tsockman remove <entry>” is called to remove “<entry>” before the new server is launched. If there is no entry, tsockman will respond with “Process by that name does not exist”. THIS IS NORMAL and is not indicative of an error. The server in question launched (without fanfare) right after that output text.

Note: The bootlaunch scripts will not start a new server if there is an existing process running. Therefore, type “ps aux | grep server_name” where server_name is the name of the server. If there is a dead process running, look up the process identification number (PID) and type “kill -9 PID” to kill the process and then run the relevant bootlaunch script.

Restarting Servers using the rc.local file


This procedure is applicable to servers that have not switched over to the bootlaunch paradigm, which may not be any at this point.

If a server is not running or Socket Manager reports that a server is dead, then look at the socket manager list to find out what computer the server runs on (socket_manager.list). You can also look at the up-to-date file by opening a terminal window and typing “less /ctrscrut/chara/etc/socket_manager/socket_manager.list” Note that servers can be running fine, but if the Socket Manager drops the connection to them, they are as good as dead when it comes to functioning with other servers or as part of a larger sequence.

Log on to the relevant computer by typing the computer name (ctrscrut, ople, s1, …). If the shortcut doesn't work then type “ssh name” where name is the computer name.

Find out if the server is running by typing “ps aux | grep server_name” where server_name is the name of the server.
[ctrscrut:599] ps aux | grep pico_1
observe 9281 0.0 0.0 61156 692 pts/3 S+ 13:58 0:00 grep pico_1
observe 12578 0.0 0.0 24524 11212 ? S Jun16 33:14 /usr/local/bin/pico_server /dev/ttyC8 /ctrscrut/chara/etc/pico_1.cfg

If the entry for the dead server shows up in the process list, then identify the process identification number (12578 for the example above) and kill the server by typing “kill -9 PID” where PID is the process identification number.

Look up the commands to restart the server by typing “more /etc/rc.local” (this is relevant for servers that run in the background). Press the space bar to scroll through the contents of the rc.local file. Locate the commands relevant for the server that needs to be restarted and copy and paste into a terminal window:

#Start PICO server for PICO #1
/bin/rm -f /var/lock/LCK..ttyC8
/usr/local/bin/tsockman remove PICO_1
/usr/local/bin/pico_server /dev/ttyC8 /ctrscrut/chara/etc/pico_1.cfg &

The first command removes the lock to allow the server to restart. The second command removes the name from the socket manager listing. The last command restarts the server. Note that if you are restarting the servers as observe, you will need to remove the part of the command in the rc.local file that saves information in /var/log/server_name.log file (the actual command typed should resemble the last line above).

There are text files on the desktop with many of the restart commands. Use these files for quick access to the relevant commands. The commands are edited and can be copied exactly as written. Files include Dome servers and all servers running on ctrscrut. Many of these commands are also located on the Restarting Servers page.

Shutters Server


The Shutters server can become unresponsive or disconnected from the Socket Manager. This server must be restarted from the lab and not from the Control Room. Follow these instructions to restart it. Note that Shutters runs on ople, not ctrscrut.

To start the shutter server on ople:

Log into the ople computer and kill the process labeled shutters with the PID as described in Restarting Servers above.
Turn off the power to the Shutters with the switch on the computer rack which is to the left of the computer desk and marked “SHUTTERS”. Restart the Shutters server with the commands below. After restarting the server and testing the gui to see that it works, turn the SHUTTERS power back on with the switch. There is a printed sheet of directions in the lab to help you.

/usr/local/bin/tsockman rm shutters
ctrscrut/usr/local/bin/shutter_server /ctrscrut/chara/etc/shutter.cfg &

MIRC-X CredoneImAcq Server

The MIRC-X CredoneImAcq server and GUI can be opened using the “credone” icon on the wolverine desktop. Note that you can only have one of these open at a time. If you get an error message and can't bring up credone, then close the error message and follow the steps below to remove the dead processes first:

From a terminal window, log on to mircx (old observe password):
ssh -X spooler@mircx

Look up credone process:
ps aux | grep credone

Kill all credone processes (including the ones for the error messages) using “kill -9 PID” where PID is the process number. Note that credone can't be restarted if there are any error processes in the list. Check “ps aux | grep credone” one more time to make sure all processes are cleared (the only one that should show up is the grep command that was just issued).

Then restart credone using the desktop icon or by issuing the following command on spooler@mircx:
credoneImAcq –no-display

(the first “–” should be two dashes: - and -).

Telescopes and Dome Servers

The Telescope won't move or stopped moving

Have the powers been turned on to the drives? Are the scopes disabled? The usual state of the telescopes is disabled until enabled. This is due to the stall function of the scopes which eventually disables the scopes when they are stowed. Enable the scopes by hitting [ENABLE] on the dome gui, the telescope gui control tab, or the bottom of the obsgtk. If a scope disables itself during a slew, it may be just an overcautious stall function or the slew gains may be too high or low. Check the gains and if they are OK, Hit ENABLE and the scope should continue to slew. If it disables again without moving, there may be something wrong at the scope, ie. a hatch left open, a ladder not put away, a tool on the floor, or something else physically impeding the motion of the scope. You will need to go to the dome to see what it is. The computer in the dome will give you control of the scope to turn it away from the problem.

Sometimes the dome guis get hung up and can cause erratic motion or no motion of the scopes. Check them for current times and continuous updates of numbers. If they are not updating, try to REOPEN them first. If that does not work, close the gui and open a new one. If a new one does not open, the dome server may be dead or Sockman lost track of it. See Dome Server Restart below.

Azimuth Limit Switches

As of 11-'17, the azimuth limit switches are enabled and can stop the motion of the scopes if they try to go beyond -90º or 450º. The scopes will not be movable with normal inputs so follow these instructions to return them from the out of range condition.

1. On the domegui MANUAL tab, click STOP so pulses won't be sent to the drive by the control software.

2. Make sure you understand why the limit was hit which may require a trip to the telescope. If the azimuth positions on all telescope servers and dome guis match, it is likely the limit switch causing the stall and not that the scope is actually in a wrong position.
Make sure all the scopes’ demand positions agree – for example, sometimes bringing a scope to a configuration that’s already on sky and issuing a slew command will make the additional scope go around North the “wrong” way.

3. Click the OVERRIDE ON button in domegui MANUAL tab. After this, the hardware doesn't care about the limits switches and you're free to move the telescope.

4. Click ENABLE then you can move the telescope back to its normal range of operation.

5. After the telescope is back in it normal range, click OVERRIDE OFF which makes the hardware aware of the limits again and then hit AUTO on the AUTO tab to resume normal operation.

However, If you notice that a telescope will stop near AZ 90 or 270 with the scope still being ENABLED and refusing to move, this is often due to the new AZ limit switch being armed around AZ 0º at some point earlier. To get it moving again:

1. On the domegui MANUAL tab, click STOP so pulses won't be sent to the drive by the control software.

2.
Make sure all the scopes’ demand positions agree – for example, sometimes bringing a scope to a configuration that’s already on sky and issuing a slew command will make the additional scope go around North the “wrong” way.

3. Click the OVERRIDE ON button in domegui MANUAL tab. After this, the hardware doesn't care about the limits switches and you're free to move the telescope.

4. Move the scope a bit back toward the direction it was coming from – for example, if the scope stopped at AZ 268 while rotating clockwise, move it back to 265 or so using AZ DEC. Then press STOP.

5. Move the scope past AZ 270 by pressing AC INC. Usally 2-4 degrees will do. Go past the point that shows the limit is triggered.

6. Click OVERRIDE OFF, go to the AUTO tab and press AUTO, then NEXT (also in the obsgtk; this will restore the original star's demand position to the scope).

7. When you have time, go to the scope and reset the limit switch; otherwise, it will stop each time you pass AZ 270/90. The LED will show red when on the limit switch and is tripped, ie. limiting motion of the scope. The LED will be yellow if it has tripped and is in the caution range, but not on a limit switch. A fine Allen key can be used to push the internal reset button. It will turn the LED green when restored.

The Telescope won't track

If a bright star is manually found and not slewed to by number, (usually after the scope has gotten lost), the telescope will not track as it does not know it has gone to a star. When you init the EL and AZ position of the scope to the star's position, you must enable the tracking by going to the dome gui, selecting the RA/DEC tab and hitting the [GO RA/DEC] button. The scope will track, but the pointing model may still be off. Try slewing to the same star and initing on it after it is locked in Tiptilt. This can be done in the dome with the computer at the telescope.

The Telescope won't stop moving or moves beyond the commanded position

Disable the scope whenever its motion is outside of what it is supposed to do. Turn the powers off to the axes as well on the Power gui. The dome server is usually to blame and will need to be restarted.

Dome Server Restart

Dome servers are now started using the bootlaunch_master command. The manual process that has been superceded is archived.

To restart the dome server:

1. Make sure the power to the drives is OFF. Disable the scopes.

2. Login to the relevant computer as observe. For example, type “s1” to log on to S1.

3. Work out the process ID number (PID) by typing bootlaunch_master

4. Use kill -2 PID to kill the server. Entering the command twice will show “No such process” if it has been killed. if it is not killed, use kill -9.

5. Type bootlaunch_master again to restart the server.

6. Turn the power to the drives back on.

7. Hit REOPEN and ENABLE on the domegtk, and type “otcs” in the telescope server.

You may have to reinitialize the scope on a bright star. If the powers were turned off quickly when the problem was noticed, the position of the scope should be retained and slewing to a bright star will get it in the finder. If not, you may need to go out to the telescope to find the bright star to reacquire the scope's position.

Be aware that when restarting a dome server, the telescopes position may not be retained and the dome gui may display Az: 0.0. El: 0.0. The gui will scroll messages that say Az: 0.0. El: 0.0 with an error in scope position. If the scope is known to be at STOW position, you can manually enter the scope's AZ and EL using the manual tab on the dome gui and hitting the INIT button.

Also be aware that the gains do not always return to the normal values of Slew 4-7 and Tracking 7-10 when a dome server has crashed and is restarted. They may come up as Gain 22 for all axes and modes and they will need to be turned down or the scopes will oscillate badly.

Telescope is not receiving the commanded position for a target.

Sometimes it happens that a telescope receives the wrong position for a target or does not receive the commanded position at all. The commanded position is listed on the telescope server in the first column under TCS Az/El; the second column lists the actual position of the telescope. Try entering the star designation directly into the telescope server, ie. hd 123456. If it does not accept the number, try closing and restarting the telescope server and hitting repoen on Cosmic Debris and the telescope gui. Try entering the star into the server again. If that does not work, it is possible that something is wrong with the dome server. To restart the dome server follow these steps above.

Telescope is tracking poorly, overshooting in slew, oscillating.

This might mean that the gain for the tracking servo is wrong. Note that changing this gain can be dangerous, especially if you set it too high as that can cause the telescope to oscillated and damage the drives. Please only do this if you are very very sure that it is necessary. Symptoms of bad gain are: The scope over shoots the position while slewing. The star will be seen to move out of the window and may come back after a few seconds. This means the slewing gain is too low. The scope oscillates when tracking or after a slew. The star will be tracing an ellipse, figure eight or other looping shape. This means the tracking gain is too low. You can damp this out with the telescope or dome gui by disabling the scope, then re-enabling it. Adjust the gain upward and watch it on the next slew. In all cases if either gain is too high the scope will go into “Fog Horn” mode, which is bad. This can be seen during slews on the twfs or labao as vibrating spots, usually in one axis.

You always want to use the lowest gain that still allows the scope to work as best as possible. If the tiptilt tells you the scope is oscillating slowly, the gain may be too low. If it is oscillating quickly it may be too high.

The usual values for slewing gains are 4-7 and tracking gains are 7-10. Note that these values may change as the temperatures change rapidly. Gains are usually higher when cold and lower when warm. Be sure to set mode back to AUTO if changing the gains left it in Slewing or Tracking mode.

How to Adjust CPUMotor Gains


The GAIN controls the gain of the feedback between the encoder and the drive velocity. A high gain means a “stiffer” response, but can lead to oscillations or fog-horning if it's too high.

The Fn controls the maximum frequency of the servo response. A high Fn means higher frequencies are allowed through, which can mean correcting for faster problems but if too high can also lead to oscillations or fog-horning.

The software will not make changes to either of these quickly as that is a dangerous thing to do. There is NO POINT to clicking the up or down buttons more than once every few seconds. Indeed it is bad to do so as you will confuse the software. The change between slewing and tracking is also slow for similar reasons. This is why sometimes the “wrong” thing seems to change. It is a sign that you are trying to do things too quickly.

Slew and tracking mode work differently, mostly because the speeds are so different.

IN SLEWING MODE

- If the gain is too low you will overshoot the target.

- If the gain is too high it will fog-horn.

- If the Fn is too high it will also fog-horn, even at low gains.

You need to have the lowest possible gain and Fn in slew mode that doesn't overshoot the target. Fn in slew mode should almost never be higher than 4. If it is, please turn it back down to 4. If you think this is a problem please let Theo know, along with a detailed explanation of what happened.

IN TRACKING MODE

- If the gain is too low it will keep moving between slewing and tracking.

- If the gain is too high it will fog-horn.

- The same goes for Fn.

In tracking mode you want the highest gain and Fn that allows the telescope to track well without fog-horning. If it “oscillates”, which you will see in the green dots of tiptilt oscillating, try turning up the gain, and also try turning down the Fn.

Some final remarks:

- The gain is temperature dependent, so when the temperature changes these things will change, but more so for tracking.

- The tiptilt system almost never causes oscillations, it almost always shows you that the scope is oscillating. If the white dots are centered on tiptilt and the green dots are moving the tiptilt is doing it's job and correcting for scope motion.

- If a drive gets disabled at the end of a slew, the gain is too low. At low gain and low velocity the encoder signal changes very slowly or doesn't change at all. After 5 secs the software interprets this as a stall and disables the drive. The tricky part is that increasing the gain to avoid this situation might make the telescope to oscillate during the next slew. So the gain should be low (4 or even 1) during slewing but higher 7 or 10 when the telescope is basically at the target position.

E2 AOB Dichroic Recovery

E1 Hut and Cooler Communications Recovery

S2 Acquisition Server Restart

S2's acquisition server resides on “S2pi4” computer. Logging on to it (ssh s2pi4) and running bootlaunch will restart it

Dome issue

At times the domes do not rotate, open or close, or otherwise behave. Some problems are simple and others are more complex.

Dome does not rotate

Sometimes when observing, the dome will not follow the telescope during a slew. This can happen when the Autodome feature is not turned on. Click the ON button on the MAIN tab of the telescope gui to enable it. This may happen after a server restart so always check the dome position with the spycam during a slew after a server restart. Also make sure the target position of the dome matches the telescope's position. If not, it will insist on being in the wrong place. If it is not at the same AZ as the scope, manually move it until it is centered on the telescope in spycam 1. If the dome AZ does not read the same as the telescope AZ, enter the scope AZ in the position box of the DOME tab of appropriate dome server and hit the INIT POS button to tell it at what AZ it is.

If the dome does not turn at all, even with the manual controls on the telescope or dome guis, the control may be set to manual on the control box. This can happen if there was work done at the dome during the day. If the dome opens, but does not turn, check to see that control of the dome rotation is in the computer position and not manual on the dome rotation controller box just inside the door of the bunker. Sometimes the drive wheel jumps in the track and cannot turn the dome, even when the motor works. This will need to be fixed during the day.

Dome does not open

Is the SPYCAM on and updating? An old image may show the dome is closed, but it has actually started to open. If the dome won't open, try hitting the SLIT CLOSE button on the telescope gui first. It may be that there was a computer or server issue and the computer thinks the dome is open already and won't allow it to open again. Hit the SLIT OPEN button to see if it works now. If it still won't open, go to the bunker and look to see if the power switch is on to the dome. It will be up and red if on.

Also look to see if the power cable is connected to the computer box. If it's connected up high to the manual dome controls, the cable needs to be brought back to the computer box connector. Turn the power off first to the dome before disconnecting the power cable. Slide it onto the connector at the box and turn the collar to lock it in place. Turn the power back on and see if you can open the dome from the computer in the bunker.

Dome does not close

If a dome will not close after observing, it may need to be closed manually at the bunker. Turn the power to the dome off at the switch, disconnect the cable from the computer control box and connect it to the manual control switches above. Turn on the power to the dome after connecting the cable and use the two controls to close the dropout first, then the slit. The dropout must be fully closed and the slit closed over it to be shut properly. Inform the staff that the dome did not function properly so someone can look at it in the morning.

HUT servers

The HUT servers control functions such as beacon and dichroic movements, heater and dehumidifier usage, and various AO functions. An observer may find that the obsgtk is no longer controlling the beacon LED's, beacon flat or dichroic alignments. This happens on occasion with E2 and other scopes. The HUT server may be the cause if it has quit or lost connection or the AOB may be at fault. To see if it is the server, open the HUT gui for the desired telescope from the CHARA menu. If the alignments can be changed from the gui, then the HUT server is ok. You can use the hut gui to continue observing. If the hut gui gives move error messages, cycle the power on the AOB and open a new hut server to restore the connection to the obsgtk. On the POWER gui, turn off the power to the AOB for the offending telescope and turn it back on. Stop the hut server by logging into the appropriate telescope computer and identifying the PID with the bootlaunch_master command and killing the process with the kill -9 #### command. Start the new server via the bootlaunch_master command. Hit REOPEN on the obsgtk to reopen the connection to the HUT server and hit reopen on Cosmic Debris as well.

If the server won't restart, a reboot of the power supply in the telescope bunker might be necessary. The power supply that controls the acquisition and finder cameras as well as their controllers is located on top of the computer rack in each bunker. The power supply has green readouts of volts and current. After turning the power off for 10 seconds and back on, try restarting the server from the computer in the bunker to see if it starts cleanly. If so, then restart the telescope server, reopen the connection to the telescope gui, and hit REOPEN on Cosmic Debris. Part of the HUT server also controls the AO table. If the AOB part of the HUT server doesn't work, then the power supply on the back of the AO table in the telescope dome might need recycling. This power supply controls the actuators at M2 and the AO table. The power supply box is a 6×9 inch aluminum box on the back of the AO table, behind the keyboard and monitor. Turn it off with the power button on the bottom edge of the box, wait 5 seconds and turn it back on. The HUT server should now restart cleanly. Restart the telescope server as well to make the connections to the telescope gui. Hit REOPEN in Cosmic Debris if you are observing to make all needed connections.

Tiptilt Server. (These sections apply to the old lab tiptilt system retired in April 2021)

OPLE and Metrology

OPLE Server doesn't open - complains about not finding run_ople

This usually happens if someone runs a “make clean” in the cvs tree but doesn't follow up with a “make install” on ople. To generate the run_ople file, follow these steps:

  • Open a terminal window and type “ople” to log onto the ople machine.
  • Change to the appropriate directory under the CVS tree by typing “cd control/cliserv/ople”
  • Update the CVS tree by typing “cvs update”
  • Run the Makefile script by typing “make”
  • Log on as root by typing “su” and entering password.
  • Type “make install”
  • This will install run_ople. Try opening the ople server again.

Metrology and Homing carts

  • Good metrology signals are important to the proper positioning of the carts. Monitor the signal strength by running [RUN MULTIPLE] on the Metrology Monitor. Place the windows above the TV windows for each scope you are using. They should show white sine waves that are around the height of the window. Erratic, fluctuating waves indicate self-interference or a weak signal. This may cause the carts to lose their place as the signal strength falls too low. Red waves indicate that some displayed signal has gone too low and the carts will all need to be homed. A careful adjustment of the MET2 mirror can sometimes bring the signal back. Do not adjust the MET1 mirror.
  • To home the carts, turn off the [OL] and [MAN] buttons on each cart and it will automatically return to the front switch. If the cart has no issue, it will arrive at the target position of 0m and the home switch at the same time. The X in the OPLE server under the HM will indicate it has stopped on the home switch. If a cart does not reach the home switch when it returns to position 0m, it was lost and likely the cause of the difficulty in finding fringes. The X does not guarantee the cart has retained its home position. This home position can be checked by using the CHECK button on the ople gui or by typing the command “homechk S1” into the ople server. When the error value is displayed on the ople server, hit ESC to clear the display. Click on the [OL] and [MAN] buttons and hit [TRACK] to send the cart to the desired cart position.

Beam Samplers

Beam Samplers are not moving

This happens most frequently with the E1 beam sampler. If you are doing alignments and are using E1, try to move it forward first to see if it dies. If it does, follow this procedure.

  • Turn off power on Newport Universal Motion Controller/Driver for the Beam Samplers (on top of rack in computer area, labelled “BS”)
  • Then go behind the racks and locate the serial port for BS. Look for all of the blue ethernet cables coming out. BS is labelled in prt 2. Remove the ethernet cable for BS. The step stool is available to make it easier to reach the blue ethernet cable at the back of the BS controller box.
  • Restart BS controller (press power button again). Wait for all axes to appear on screen.
  • Plug the ethernet cable back in.
  • Restart BS server on ctrscrut using commands in /etc/rc.local (see instructions below)
  • Reopen Beam Sampler GUI. You will need to home each Beam Sampler one by one. You should be able to do this from the regular Beam Sampler GUI. If that doesn't work, try opening “espgtk BS” from the command line.

Restarting the Beam Sampler Server

The most common cause for getting multiple copies of a server is using the socket manage RESTART button, which is flaky at best. Don't do this. If you need to restart a server you should do it manually.

1. Make sure you are logged into the right machine: ssh ctrscrut

2. If you are not sure see if the socket manager will tell you. If it doesn't have a look in the file: /ctrscrut/chara/etc/sock_manager/socket_manager.list

3. See if there are any ghosts running

ps aux | grep esp_server

Yes, it's hard to know which one it is with the esp servers. You can work out which ones are ghosts by typing the command

tsockman |grep ctrscrut

which will give you a list of the servers running on the machine you are interested in. Checking for non-matching PIDs will tell you which processes you need to stop.

4. Stop those processes:

kill -9 PID1 PID2 ….

5. Make sure there is no sign of it in the socket manager

tsockman rm BS

6. Restart the beam sampler servers (this starts both beam sampler servers BS1 and BS2):

bootlaunch_beamsamp

"Failed to request position of S1"

Cosmic Debris reports “Failed to request position of S1” when trying to set the beam order. This indicates that ople is no longer talking to the beam samplers. If the beam sampler server has been restarted recently, then the ople server will also need to be restarted to re-establish the connection.

The beam sample server runs on ctrscrut. The ople server tries to open a connection to the beam sampler when it starts. If the beam sampler is not there, or dies,
you need to restart the ople server as there is no command to reopen that connection.

Remote Observing

VPN connection is not working

To re-establish the VPN connection, follow these steps:

  • Open VPN status page from the firefox web broswer. To do this, click on the “AROC VPN reconnect” icon on the zoot desktop. Alternatively, you can open the firefox web browser and go to 192.168.3.1/vpn_sta.htm
  • The browser will prompt you for the admin user name and password. If these entries are not already filled in, then look for this information on the papers near the console.
  • If the browser does not open, then check if there are other instances of firefox already running. To do this, open a terminal window, type “ps auxw | grep firefox”, and kill the other instances of firefox so that you can open a new window.
  • When the VPN connection is working, it will report “Phase 1: M-ESTABLISHED / Phase 2: ESTABLISHED” under the state column.
  • If either Phase 1 or Phase 2 states say IDLE for a given VPN, then you need to re-establish the connection. To do this, click on the “DROP” button for that VPN. Then click on the “CONNECT” button and wait for the Phase 1 and Phase 2 connections to be established.
  • Make sure to close Firefox by using CTRL-Q to quit the program and not just close the window.

General Problems

GUIs are frozen

Check to see if Sockman is working. Click the [LIST] button on sockman. If the list doesn't come up, Sockman may be hung up. Follow the Sockman restart procedure in the text file on the desktop if it won't respond.

Hit [REOPEN] on the gui to see if it reconnects with its server. if that does not fix it, the server may be dead. From the LIST on Sockman, select the server suspected of being frozen. Ping it and if it reports back as being dead, see Server is frozen section below.

Telescope guis often freeze. Hit reopen on the gui to bring it back to life. Hitting [MOVE] on the telescope gui before the star has stopped moving can cause the gui to freeze. Avoid doing this. Close the dead one and open a new one to fix the problem.

Server display is blank or filled with jibberish

  • If the top of a server goes blank, try typing “sb” (start background) on the server command line.
  • If the server screen fills with jibberish, try hitting CTRL-L in the server to clear it.
  • If a server freezes, sometimes hitting CTRL-C and then N, for NO, will unfreeze it.

Server is frozen

Try quitting the server using CTRL-C and entering “Y”. If that doesn't work, try the following:

  • On Sockman, click [LIST] and select the appropriate server. It will give a PID number you can use to kill the server in a terminal. Open a terminal and follow the directions for Restarting Servers above.
  • If a server window does not close, click the X in the upper right corner to close it, but only after getting the PID from Sockman.
  • Click [REOPEN] on Cosmic Debris and relevant GUIs to re-initialize communication with the server after it is restarted.
  • A folder is on the desktop that has the restart commands for CTRSCRUT servers and the Shutters server restart command running on OPLE. Use it to restart servers that will not reopen from the menu.
  • if a telescope server does not close after using CTRL-C when restarting them after UT midnight and the Send ON button is green on the obsgtk, turn the Send ON to OFF and the server will close. Turn the Send ON button back on to resume the setup for the night.

PAVO Server - Error communicating with IFW

The following error message sometimes appears when starting the PAVO server: “Error communicating with IFW. Could not read from IFW. Is the filter wheel plugged in and on?” Most of the time, restarting the PAVO server will clear this error message and allow PAVO to communicate with the filter wheel. If restarting the server doesn't help, then check that the cables going into the small, black IFW box on the PAVO table and into the back of the PAVO computer are plugged in securely.

Adding or Finding a Star in the CHARA Database

The command “dbadd” can be used to look up the CHARA number for a star or to add a star to the CHARA database. To use dbadd, open a terminal window, and type “dbadd starname” where starname is the common name (e.g., Vega) or identifier of the star (e.g., IRC, HR, HD, SAO, FK5, HIP, GJ, or 2MASS designation). If the object is in the database, then it will return the CHARA number.

You can also look up different identifiers for the star by using the SIMBAD database: http://simbad.u-strasbg.fr/simbad/sim-fbasic

DBADD uses SIMBAD to look up the coordinates of the star and then searches to see if the star is already in the CHARA database. If the star is in the database, it will return the CHARA number. If the star is not in the CHARA database, it will ask if you want to add it and assign a CHARA number if you answer yes.

If an object is not in SIMBAD (such as a recently discovered nova), then you must enter the object manually using the instructions below.

Finding a star in the CHARA database

Here is an example of how to find the CHARA number of a star, nova, or AGN that is already in the CHARA database:

[ctrscrut:502] dbadd NGC 4151
Object “NGC 4151” found as CHARA number: 320263

Note: If the target information stored in the CHARA database does not match SIMBAD, then dbadd will provide the coordinates and magnitudes listed in SIMBAD, and it will also list the CHARA numbers of objects in the database located within 10 arcsec. Please check these nearby entries before adding a new star to the database.

Targets that are not in the CHARA database

If an object is not in the CHARA database, then you can use dbadd to query SIMBAD and add the star to the CHARA database. This might need to be done for faint stars, novae, and AGN. If the object is found in SIMBAD, then dbadd will return the target's coordinates, magnitudes, and spectral type and ask if this it the object for which you are looking. If you answer yes, then dbadd will issue a CHARA number and add the object to the database. Please check the target coordinates to make sure that they are correct before adding a new star to the database.

Here is an example of how to add an object to the CHARA database:

[ctrscrut:534] dbadd BD 41 3807

SIMBAD found this for “BD 41 3807”:

ID 2MASS_J20340850 4136592
CHARA 320414
2000.0
CAT8 2MASS_J20340850 4136592
RA 20 34 08.5148
DEC 41 36 59.408
PMRA -2.32
PMDEC -3.87
PAR 0.58
RVEL -27.50

VMAG 10.03
HMAG 6.23
KMAG 5.99
SP O5.5Ifc
END

Is the above the object you are looking for? [y/n] y
Added “BD 41 3807” to the database.

In the example above, the object “BD 41 3807” has been added to the database as CHARA number 320414.

Targets that are not in SIMBAD

Some objects like recently discovered novae are not in SIMBAD. Entering dbadd <star designation> for a star not in the SIMBAD database returns the message “Simbad is unable to find an object matching <star designation>. Try using a different catalog designation, or use the ”-m“ switch to add the object manually.”

The command dbadd -m is used by entering the star name and coordinates in this format:

Usage: dbadd -m <name> <RA> <Dec>

<name>: Object ID (no spaces)

<RA> : XXhXXmXX.X or XXhXX.X (no spaces)

<Dec> : XXdXXmXX.X or XXdXX.X (no spaces)

There are cases where an object is not in SIMBAD and doesn't return a result in dbadd, but is in fact in the CHARA database. Novae and AGN's are the likely objects that cause this result. At times, objects have multiple entries and several CHARA numbers since the names can be so unique. A system will be developed to find these entries without knowing or having a conventional database designation.

Binary stars

Some binary stars have a common HD number with an A or B after them. These can cause problems as Cosmic Debris does not accept non numerical entries when filling in star designation. These stars are likely in CHARA's database but need to be searched in dbadd or on SIMBAD to get the CHARA number or another designation.

When a recognized identifier is entered, the HD number with A or B will usually display. Confirm that the coordinates, magnitudes and spectral type match the star desired. If they do not match, you may have the wrong star or the database may have the wrong info and the baseline solution will be affected. Offsets based on incorrect coordinates or misidentifications can move the fringes many thousands of microns away from the calculated position. This can happen when observing a dim companion (B) using the brighter (A) companion's coordinates. Inform the observers that noting and using the CHARA number will save time next time the target is observed.

Editing the database: If you find a mistake in the database, please send an email to Nils to have it corrected. Identify what you believe to be the error and what is the correct information.

Adding an anchor to the wiki page

To add an anchor to the wikipage, type the following where you want the anchor inserted (without the spaces between the symbols):
[ [ #AnchorName ] ]

Highlight the text you want to be a link to the anchor, click the link symbol above and select anchor. Select a page(such as Trouble Shooting) where the anchor is located and give the same name for the anchor.

Displays for the MIRC-X wolverine2 monitors

Follow instructions in the linked page to display the MIRC-X wolverine2 monitors in the control room.


Last updated 2021-01-18 by Norm Vargas

chara/trouble_shooting.txt · Last modified: 2022/02/10 18:46 by charaobs