Summary
- Symptom
- LRIS widgets such as XLRIS are slow to
respond to keyboard and mouse input on
hanauma, but the response time from local windows on
hanauma is normal.
- Problem
- Either lrisserver is overloaded or the network
connection to the summit is overloaded.
- Solution
-
- If you are running a Netscape window on
lrisserver, then kill it.
- Check the CPU load on lrisserver by typing the
following command in a lrisserver window:
ps -aux | more
- Check listings under the %CPU column to
identify any jobs consuming large fractions of CPU
time. Contact computer support to halt such
processes.
- If no high-CPU processes are found, then another
user may be transferring large amounts of data to or
from the summit, or the network may be slow, or
perhaps lrisserver needs rebooting. Contact
computer support for assistance.
- Symptom
- Attempts to use the configure command to
change the grating, slitmask or filter list give error messages.
- Problem
- The grating/slitmask/filter list is corrupted in memory
and on disk.
- Solution
-
- Log into lrisserver as user kics.
- Go to directory /usr/common/music/info.
- Inspect appropriate file (graname.str,
redfilt.str, or slitname.str) for
obvious errors.
- Create a backup copy of the offending file; e.g.,
cp slitname.str backups/slitname.str.11oct1999
- Edit the file (emacs, vi, etc.)
to remove the offending entry.
- Determine the process number (PID) of
infoman using ct.
- Use the following command to halt and restart
infoman:
kill -9 154 ; /usr/local/music/bin/infoman l_info v=lris &
- Re-try the configure command.
- Symptom
- When XLRIS is started, the window appears and then
vanishes within a few seconds without any warning messages.
- Problem
- The grating/slitmask/filter list is empty or corrupted.
- Solution
-
- Type the following command to check the
configuration lists:
consort
This will print out sorted lists of the configurations
for all configurable stages. Check for a list which has
either
- no entries
- invalid entries
- duplicate entries (two positions with the same string)
- If one (or more) of these lists is empty or corrupted, fix it
as follows:
- Restart XLRIS.
- Symptom
- The blue lickserv process dies. Attempting to
restart it appears to work briefly, but it dies within a
minute of being restarted.
- Problem
- The observer attempted to window the blue CCD, which is
not permitted, and the blue crate is in a bad state.
- Solution
-
- reboot the blue CCD crate (connect to the crate from
the desktop menu using Engineering >
CCD Crates... > TIP line to BLUE CCD crate) and type
the command reboot
- Restart the blue side lickserv process after the
reboot has completed using the pulldown menu item
LRIS Control Menu > Subcomponents... >
Re-start LICKSERVER... > Re-start Lickserver (BLUE)
It is not necessary to cycle power on the blue saddle bag.
- Symptom
- Show keywords give solicite the following output
startproc: connect failed
startproc: connect failed
startproc: connect failed
startproc: connect failed
startproc: giving up on connect; exiting ...
Sorry, the show command was not able to contact the control system: Can't
connect to traffic:
After starting a crate session with start_crate [red | blue], the
following output is printed to the screen:
Got network configuration
Broadcast message sent on ln0
Broadcast not sent on lo0
tropen: our host = ted.
read_remote: pathname=/local/kroot/data/music/services
read_remote: service file open
tropen: remote host = lrisserverp
startproc: parameter hostname = lrisserverp
startproc: gethostname returned ted
read_remote: pathname=/local/kroot/data/music/services
read_remote: service file open
startproc: connect failed
read_remote: pathname=/local/kroot/data/music/services
read_remote: service file open
startproc: connect failed
read_remote: pathname=/local/kroot/data/music/services
read_remote: service file open
startproc: connect failed
read_remote: pathname=/local/kroot/data/music/services
read_remote: service file open
startproc: connect failed
read_remote: pathname=/local/kroot/data/music/services
read_remote: service file open
startproc: connect failed
startproc: giving up on connect; exiting ...
"cserv.c", line 106: cserv, da8398, "cserv ()",
Error, #41: Can't make connection to traffic controller. Unable to make
connection to traffic controller, retrying...
Timeout waiting for ISERV, retrying
- Problem
- The low_level_software is not running or has died.
- Solution
- Restart the low level software
- Symptom
- The Xpose window will not allow you to start an exposure.
- Problem
- Either the GUI is in a confused state or the relevant
watch_ccd process isn't running.
- Solution
- Try each of the following to see if that re-enables control in Xpose.
- Click ABORT in the Xpose window, which
should re-enable the START button.
- Restart Xpose from the background menu.
- From the background menu select LRIS Engineering -> Reset blue
CCD flags or LRIS Engineering -> Reset blue CCD flags as
appropriate.
- Check that the
watch_ccd processes are running by typing
ctx . There should be a watch_ccd_red and a
watch_ccd_blue process. If either is missing, restart
it by logging in as lris@manuka and running lris
start watch_ccd .
- Symptom
- After images acquisition completes, no image appears on the
ds9 image display.
- Problem
- The ds9relay widget is not transferring data to ds9 becuase the ds9relay is running under some other account.
Keep in mind that the ds9relay runs on the linux server (lris).
- Solution
-
- Both the ds9relay_red/blue and the ds9_red/blue should
run only on the linux server. If they are running on the Sun
system they will not display images on the vm-lris.
- So please run lris stop ds9relay_red/blue
and lris stop ds9_red/blue on lrisserver.
- restart the ds9 from the background menu
- Symptom
- After images acquisition completes, no image appears on the
ds9 image display.
- Problem
- The ds9relay widget is not transferring data to ds9 becuase there is a problem with the X server.
Keep in mind that the ds9relay runs on the linux server (lris).
- Solution
-
- X server may be insecure. Type xhost to
check.
- If that's not the problem, then shut down and restart LRIS
software.
- If that doesn't fix it, then log out and log back in.
- Symptom
- Software cannot be started complaining that another user is running it. Running ctx shows a missing caRepeater.
- Problem
- The EPICS channel repeater is not running.
- Solution
-
- Preferred: from a lris terminal on manuka lris start repeater
- Backup: Showing a dcs keyword should restart caRepeater i.e., show -s dcs AIRMASS
- Symptom
- pressing the little G on gratings, or trying to adjust the
focus fails to open a popup menu
- Problem
- The XLRIS is not working properly on the new vm system
- Solution
-
- Check the grating configuration by running
consort on an lrisserver terminal.
- Change the focus by using the focusr value
or focusb value commands on a lrisserver terminal
- Symptom
- The popup menu under CMD does not start.
- Problem
- The blue XPOSE GUI is not working well on the linux VNC system
- Solution
-
- The only solution is to restart the blue XPOSE from the
background menu.
- Symptom
- When changing gratings or adjusting grating tilt, the
move fails soon after being initiated.
- Problem
- The grating brake drive has failed to reach the unclamp limit.
- Solution
- The solution is to restart the move by selecting 'set'
in either the grating selection subwindow, or the wavelength
selection subwindow. This will normally drive the brake
unclamp to limit, prior to completing the rest of the
compound move sequence.
- Symptom
- When attempting to change gratings, the move fails.
Running check_turret on the motor crate reveals
that the grating is ``not at home.''
- Problem
- The homing operation fails because the puka (hole) in the
rotor is plugged and the optical homing switch cannot detect
light through it.
- Solution
- The puka must be cleaned by technical staff. If this
does not fix problem, then see below.
- Symptom
- When attempting to change gratings, the move fails.
In the Motor Log tail the following output is seen:
Jul 10 11:31:43 [5981] lrs_mcs.c,345,mcs_com:
Error message from API unit received, stage grating_tilt
Jul 10 11:31:43 [5981] lrs_mcs.c,346,mcs_com:
Error message: 0 "#23 HOME NOT FOUND"
- Problem
- The homing operation fails due to some failure of the
hardware. The likely hardware failure is that the home
switch has failed. The software will not permit you to move
a grating if the home switches for the grating tilt are not
set.
- Solution
- An instrument technician will have to examine the home
switches. In the mean time, avoid the toublesome port. If the
grating is needed for observing, the grating will need to be
swapped with a second grating in a port that is
opperating. Follow these instructions for manually moving
the grating turret away from the failing port.
- Have an instrument technician confirm that the
grating is tilted to a safe location. An instrument
technician has access to the grating port ...
- Safe (homed) grating position: INSERT PICTURE
- Unsafe (tilted) grating position: INSERT PICTURE
- If the grating is unsafe, change the tilt at the API
level.
-
telnet tsred 3016
-
:0
-
stat - and check the current position
-
mov 1000 - move 1000 steps (you may need to
move more or less than this value to reach an
acceptable tilt).
- Once the grating is tilted to a safe position, in a
lrisserver window release the grating detent by:
m gdetent=0
- At the grating service port, an instrument technician
will be able to manually rotate the turret away from the
troublesome port.
- Swap gratings if necessary.
- Update SIAS and the instrument software with any
grating changes.
- Symptom
- When attempting to change gratings, slitmasks, or
filters (from either GUI or keyword level), the computer
responds as if the move completed properly, but the
instrument is unchanged.
- Problem
- The software is confused. It believes that the stage you
selected is already in place.
- Solution
-
- Issue the following command from a lrisserver
command line to force the software to
re-read the instrument settings:
m init=3
Try the move again.
- If it still fails
try selecting a different filters, gratings, or slitmasks
until one works. Then, return to the slitmask, grating, or
filter you desired.
- Symptom
- When attempting to change a mechanism (from either GUI or
keyword level), the move fails immediately. The following
message appears in either the motor log window or the window
from which the keyword command was issued:
move failed:
mread select timeout
- Problem
- The motor control software is hung.
- Solution
-
- Issue the following command from a lrisserver
window to restart the motor control software and all
other low-level software:
restart_low_level_software
- Symptom
- A requested mechanism move fails with the following
message printed in the motorlog:
DRIVE FAULT
- Problem
- One of the API motor controller is confused and needs to
be reset.
- Solution
-
- Consult the following webpage to identify the APIs associated with the failing mechanism API ports
- open the power control GUI and power cycle the suggested APIs.
- Try to move again the mechanism.
If this procedure fails, then a full power cycle might be needed. This is the procedure:
- Issue the following command from a lrisserver
window to cycle power to the entire instrument:
allpower cycle
- When it comes back up, issue the following command to
turn the guiders back on:
guiderpower on
- Use the tdl command to check communication
with the crates:
tdl red
tdl blue
See below for information on dealing with crate problems.
- Symptom
- When completing a mirror move using either XLRIS or a script such as
R-band, the mirror gets set to grating angle 0
rather than the nominal angle for imaging.
- Problem
- Either the mirror move is failing due to a bad limit
switch, or the MIRGRANG keyword has been set to
zero.
- Solution
-
- Check the value of the MIRGRANG keyword
using the show command. If it is zero, reset
it to the appropriate value for imaging (currently 13.75
deg) using the modify command.
- If MIRGRANG is okay, then check the motor
log output to verify that the move completed
successfully:
cdlog
more motorlog
- Symptom
- Although the mirror (grating position 1 in optical port)
can be commanded to any GRANGLE, other gratings
cannot be tilted more than about 5 degrees.
- Problem
- The turret switch is not reporting the proper grating
position; hence, the electronics are always trying to drive
the mirror stage (corresponding to turret switch output
0000) regardless of which stage is actually in the
service or optical port. The likely problem is no signal from
the turret switch, which multiplexes the signals to the proper
grating station.
To verify the diagnosis, checking the bits on the API unit
as follows:
- Log into lrisserver
- Open a connection to the correct port:
telnet tsred 3016
- Select API 2:
:2
- Show bits on input word 1:
ver i1
ver i1
Bits 3-6 of word I1 are supposed to encode the position of
the grating turret. If these bits are all set to 1 and do
not change when gratings are changed, then no signal from
the switches is the likely problem.
- Solution
- Check for a broken wire connecting the turret switch to
the electronics.
- Symptom
- Attempts to close the trapdoor from within XLRIS fail.
- Problem
- Due to an overflexible design the
trapdoor will not often make it to its closed limit in
certain orientations; it is
essentially closed for optical purposes, however.
- Solution
- Try again to close the trapdoor using XLRIS. If it
repeatedly fails, then assume that the door is closed and
proceed.
- Symptom
- Attempts to move the slitmasks result in an error: #25 LIMIT SWITCH ENCOUNTERED. Attempt to move using the keyord (modify -s lris SLITNAME) results in Move encountered an unexpected limit.
- Problem
- The door of the slitmask cassette is not properly closed.
- Solution
- Move the telescope to horizon, and have a summit staff
close the door properly. When properly closed, the LEDs door
indicators will be dark. If the door is ajar, the LEDs will
be lit.
- Symptom
- When trying to move a blue filter, dichroic, or grism,
the move fails with the following error:
Error reading crate: Carousel cell retainer latch is in an i
ntermediate position. Check air pressure!
See log files for further details.
- Problem
- The cell retainer latch for the appropriate stage has
moved from the latched position to an undefined position in
the middle of the move. The carousel cannot be rotated until
the latch is latched, but the latch cannot be moved until the
carousel is in a valid position.
- Solution
-
- Determine which carousel (blue filter, dichroic, or
grism) is stuck.
- Have summit personnel remove the dust cover on the
appropriate carousel and push the latch into the
latched position.
- Rotate the carousel to a defined position by issuing
the command
m dichchar=1 for example,
assuming the dichroic carousel is stuck. Use
grischar=1 for the grism and
blufchar=1 for the blue filter.
- Attempt to deploy the stage.
- Symptom
- Following installation or removal of the polarimeter
module, attempts to move the red focus stage or the moveable
guider stage fail with a serial port timeout error. Motor log
may show an entry similar to this:
Nov 13 16:43:01 [9881] sta_init: Getting stage position for red_camera_focus
Nov 13 16:43:05 [9881] lrs_mcs.c,1456,mcs_timeout(): Serial port (select) timeo
ut, stage red_camera_focus , retrying select/read
Nov 13 16:43:09 [9881] lrs_mcs.c,1456,mcs_timeout(): Serial port (select) timeo
ut, stage red_camera_focus , retrying select/read
Nov 13 16:43:13 [9881] lrs_mcs.c,1456,mcs_timeout(): Serial port (select) timeo
ut, stage red_camera_focus , retrying select/read
Nov 13 16:43:13 [9881] lrs_mcs.c,1468,mcs_timeout(): Aborting because of serial
port timeout, stage red_camera_focus
Nov 13 16:43:13 [9881] lrs_encoder.c,168,read_abs_encoder: Serial port timeout,
trying to read AR encoder, stage red_camera_focus
- Problem
- The polarimeter install/remove script was not completed
properly, leaving the encoder is powered off. The system
cannot communicate with the red focus and movable guider
encoders.
- Solution
- Turn on power to the polarimeter encoder using the LRIS
Power Control GUI.
- Symptoms
-
Attempt to make motor move fails with the following
error message:
Error setting grating: Error reading crate:
Controller unit number != unit number expected.
- Problem
- The API unit used for this motor is not echoing command
input back to the motor control software.
- Solution
- Enable command echoing on the appropriate API unit by
issuing the following command by connecting to the server:
- Log into lrisserver
- Open a connection to the correct port:
telnet tsred 3016
- Select the correct API n
:n
where n is the appropriate API unit number as
determined from consulting the Motor
Stage Data file.
- Set the echo mode:
ECHO
- Now, if one types a command in the server, the server
would echo that command. If this is not the case then issue
the echo command again
- Disconnect from the server
- Symptoms
- When attempting to move the red filter into the beam
(deploy), the move fails, and in the motor log, there is the
following error:
Dec 1 16:15:08 [734] BEGIN: red_filter_jukebox complete move
Dec 1 16:15:09 [734] red_filter_jukebox : Moving grabber into box
Dec 1 16:15:09 [734] red_filter_grabber : Applying power to motor controller
Dec 1 16:15:09 [734] lrs_mcs.c,345,mcs_com: Error message from API unit received, stage red_filter_grabber
Dec 1 16:15:09 [734] lrs_mcs.c,346,mcs_com: Error message: S=0,MO 1
3 "#25 LIMIT SWITCH ENCOUNTERED"
3>
Dec 1 16:15:09 [734] red_filter_grabber : Releasing magnetic brake
Dec 1 16:15:09 [734] red_filter_grabber : Configuring motion-control parameters
Dec 1 16:15:10 [734] red_filter_grabber : Current encoder position = 1277800
Dec 1 16:15:10 [734] red_filter_grabber : starting servo loop w/ encoder destination = 1440000.
Dec 1 16:15:10 [734] lrs_mcs.c,345,mcs_com: Error message from API unit received, stage red_filter_grabber
Dec 1 16:15:10 [734] lrs_mcs.c,346,mcs_com: Error message: PAUSE,MOV 207616
3 "#25 LIMIT SWITCH ENCOUNTERED"
3>
Dec 1 16:15:10 [734] lrs_encoder.c,543,move_abs_encoder: Bad mcs_com ret=-3079 sending API servo move command, stage red_filter_grabber
Dec 1 16:15:10 [734] lrs_encoder.c,548,move_abs_encoder: dx=162200, db=50, count=0, max_tries=5, stage red_filter_grabber
Dec 1 16:15:11 [734] red_filter_grabber : Applying magnetic brake
Dec 1 16:15:11 [734] lrs_changers.c,194,move_sc: Bad move_float_changer ret=-3079
Dec 1 16:15:11 [734] lrs_changers.c,196,move_sc: Error moving red_filter_grabber into jukebox
Dec 1 16:15:11 [734] s_redfnum.c,172: Bad move_sc ret=-3079
Dec 1 16:15:11 [734] END: red_filter_jukebox complete move
- Problem
- Either the pin securing the filters is not sufficiently
tight, or the door to the jukebox is not securely latched.
- Solution
- Have summit staff check for a red LED indication on the red
jukebox mechanism. If lit, then confirm that the pin and door
on the red filter jukebox mechanism are both secure. Note that
if the pin is even a fraction of a turn loose, it will activate
the LED.
- Symptoms
- When attempting to move the red filter into the beam
(deploy), the move fails, and in the motor log, there is the
following error
Aug 14 09:44:51 [2650] red_filter_grabber :
Current encoder position = -9520000
Aug 14 09:44:51 [2650] lrs_mcs.c,345,mcs_com:
Error message from API unit received, stage red_filter_grabber
Aug 14 09:44:51 [2650] lrs_mcs.c,346,mcs_com:
Error message: PAUSE,MOV 140288003 "#25 LIMIT SWITCH ENCOUNTERED"
In addition:
- Homing the filter grabber and selector is
successful.
- Keyword moves fail.
- The encoder is reading negative numbers.
- Possible rollover errors reported (none observed this
time).
- Problem
- It may be that the encoder has rolled over and is now
reporting negative encoder values as seen above. This will
prevent all moves initiated via keywords and guis.
- Solution
- OPTION 1: This
problem will require manual intervention to address. An
instrument technician will need to back off the encoder more
than one full turn until it is reading positive values. For
the red filter grabber encoder, you can read the encoder
value by running on lrisserver absread tsred 13 0 1 0
.
OPTION2: If you want to put in the clear filter for spectroscopy,
you can manually do this by:
- home the filter grabber and selector
- telnet tsred 3015
- at the telnet prompt type
- :3 to select the grabber
mech.
- reset 1 2 3 4 5 6 7 8
- set 6 8
- ver o1 and make sure only bits 6 and 8 are set
- mov -208000
- set 2
On XLRIS the red filter will say "unknown", but it is in
the clear position.
The steps presented above to control the hardware at the
API level are found in the Motor Control Hardware
Troubleshooting guide.
Relevant K1 Nightlog tickets are: K1-18283
- Symptoms
- The wave plate may move to an angle of 0 degrees and
some other angles, but not all the normal angles.
As an example, it may move from -90 to +30 but not to the
normal angle of +90. Check the plangle and plangler
keywords. For a plangle=0, the plangler should be 1946876.
- Problem
- The parameter file that holds the default encoder value
for the home or 0 degree angle is corrupt.
This has happened at least twice in the
past when a new software version is released. Two
known corruptions were that 1) the file dropped the least
significant digit for the encoder value and 2) the encoder
value was deleted. It is unknown why this happens during a
new software release.
- Solution
-
Using the lris development account
- ll /local/kroot/data/lris/plangle.dat to determine
which file to edit
- The file should contain 1946876. This should be the
only number in the file.
- Save the file
- Try moving the rotator, you should not have to
restart the software.
- Recheck the plange and plangler keywords.
- Symptoms
- we notice that the calibration unit is in an UNKNOWN
position. The calnum is -9999999.
- Problem
- For some reason, the calibration unit has lost is
position. The position is defined by a set of three magnet
switches and it is secured in a position by using a detent
system. If the position is undefined that measn that the magent
switches are not defined and that the detent is not fully
engaged. There is no hoiming sequence so we would need to drive
it to a known position.
- Solution
-
We will use the slitguider to verify the position of the
calibration wheel.This picture shows the set up and the cal
unit in the open position.
- start taking images with the slitguider (300 milliseconds)
- configure lris in the following way:
- trapdoor in the closed position
- slitmask in open
- turn on the Kr lamp
- move the wheel until we see the open position on the slitguider
- telnet tsred 3016
- at the telnet prompt type
- :1 to select the calibration wheel.
- reset 1 2 3 4 5 6 7 8
- set 1 6 7 8
- ver o1 and make sure that these bits are selected
- mov 1000
- continue moving until the open position is
reached. Verify by applying small moved (10 steps) to make
sure the open position is reached and the detent is secured.
- log out from the telnet session
- mov the calibration unit to open: calname open
- If the detect is fully engaged and the magnetic swiches
are set then the new calname position will be open. If not, then the calibration wheel was not in
the correct position and the user would need to apply small
motions hoping that the magnet switches will be set to
Open. Some iteration might be neede. Please note that the
telnet session has to be closed before using keywords.
- Symptoms
-
Attempts to move the Grism results in the following error
as seen in the motor log.
UNLATCH failed, stage grism_carousel
Both carousel positions switches are not set. Latch will
not move., #1=1, #2=0, stage grism_carousel
- Problem
- The flags that are used to indicate that the
carousel is in position are not "made" suggesting
that the carousel is missaligned. Because the flag
are not made, the carousel will not deploy.
The grism will not unlatch if one or both of the flags
are not made ( #1=1 and #2=0,#1=0 and #2=1, or #1=0 and #2=0).
- Solution
- Try to nudge the carousel into position at the API level.
Please use caution.
- determine which flags are not made by issuing the command in a
lrisserver window
cari1 tsblue 16
If in position the output for bits 6 and 7 will read
I1 Bit 6 set. Cell is in position according to Switch #1.
I1 Bit 7 set. Cell is in position according to Switch #2.
- telnet tsblue 3016 log in to the terminal server
on the blue side
- check the flags by stat and looking at
I1 - STATES, INPUT SET 1 = 11110111
Bits 6 and 7 should be set to one if the carousel is in position.
If either is 0 then procede to the next step
- set 4 - this releases the carousel brake.
- mov -500 move 500 steps in the negative direction.
- check the flags by stat and looking at
I1 - STATES, INPUT SET 1 = 11110111
Try the other direction if the negative direction fails.
- If you are unable to set the flag, this position will
have to be avoided until someone is able to look inside the
instrument and verify the position of the element and the
flags.
- If the flags are set, try to deploy the grism at the
keyword level m gtran=1
- Symptom
- Grism fails to change when commanded. The motor log will contain the
error
transport at positive limit (DEPLOYED), stage grism_transport
- Problem
- The transport stage has difficulty lifting the heaviest grisms at
certain rotation angles. The 600/4000 and 1200/3400 grisms are the worst
offenders.
- Solution
- Rotate the instrument to a new drive angle and retry. In the past drive
angles around +90 were problematic and angles around -90 worked.
- Symptom
- Attempts to move the red focus mechanism yield the
following error:
Error setting redfocus: lris dispatcher: Wrong control mode for request
See log files for further details.
- Problem
- The red focus mechanism is now in position (POS) mode.
- Solution
- Follow these steps to reset red focus mechanism to POS mode:
See the red-side focus documentation for further information.
- Symptom
- Attempts to move the red focus mechanism yield the
following error:
Error setting redfocus: lris dispatcher: No communications with device.
See log files for further details.
- Problem
- The communication cable for the focus Galil is probably connected to the wrong jack or left umplugged.
The cable is a coax cable that plugs in to J15 on the telescope CASS interface panel.
- Solution
- Follow these steps to reset red focus mechanism to POS mode:
- Power off the red focus stage
- Plug the cable to the correct jack (J15)
- Power on the red focus stage and home it
- Symptom
- Guider images are saturated (signal > 16000ADUs).
- Problem
-
- LRIS lamps are on, OR
- Sky is bright.
- Solutions
-
- Symptom
- MIRA images appear to be flipped or have a randon orientation. Analysis is impossible.
- Problem
-
- The ROTPPOSN keyword is missing in the MIRA image headers.
- Solutions
-
- This is usually a temporary problem, due to a race condition with DCS. Repeat MIRA and see if the orientation is correct. In fact, the problem might only affect one of the two MIRA images, and the MIRA gui can work with just one image. Remove the wrong image from the list and push Analyze.
- If the problem persists and MIRA is required, you might have to restart the DCS computer or call software support.
- Symptom
- Starting the IQM procedure, the stars remain single or the
procedure does not start at all.
- Problem
-
- The IQ prism is stuck and does not enter the light path.
- Solutions
-
- Login to k1server as k1obs
- go to: cd magiq/lris
- execute the command: ./rununstickOffsetGuiderIQL
- this command resets the galil (ignore the error
message), and then moves the IQL in and out
- It will then ask you to confirm that the IQL is moving
by providing you with the expected and measured encoder
position
- if the measured encoder positions are within 50 counts
of the expected ones, answer yes, otherwise answer no
- if you answer no, the procedure will be repeated
- Note that if you do NOT reset the galil with this
procedure, all the other offset guider motors might be
disabled (filter for example)
- Symptoms
-
LED-contaminated image
|
Images taken with LRIS show severe light contamination
with an illumination pattern strongest at the top,
resembling the image at right. Spectra of the
contamination shows it to be predominantly red in color.
- Problem
- The light-emitting diodes (LEDs) used in the grating
tilt optical homing switches have been left on, probably
due to a failure in homing the grating tilt.
- Solution
- Please consult your instrument specialist for
assistance.
- Start a motor crate session.
- At the API level on the motor crate, type:
:0
reset 1
quit
- Symptom
- CCD images read out in dual-amp mode show strange
patterns, with the rows from the left-side of the chip
randomly appearing on the right and vice versa.
- Problem
- Fiber connections between the CCD and the crate are
flaky, causing random bits to be dropped and the sense of
certain rows to appear reversed.
- Solution
- Ask Instrument Technician to test all fiber connections
to the CCD for light loss.
- Symptom
- Images taken in single-amp mode appear to have no signal,
and images taken in dual-amp moce have single in only one
amp but not the other.
- Problem
- One of the CCD amplifiers is not working, perhaps because
the saddlebag is overheating or a card in the saddlebag is
poorly mounted.
- Solution
-
- Check the saddlebag temperature using the command
show -s lris utbtemp
If it is above 25°C, then the saddlebag should be
inspected for failed glycol or failed fan.
- If the saddlebag temperature is okay, then the
saddlebag should be opened and the cards reseated.
- Symptom
- Images are saturated (signal above 65000DN) but the
trapdoor is closed and no lights are on.
- Problem
- One or both of the fibers allowing the saddlebag to
communicate with the VME crate in the control room are broken.
- Solution
-
- Change to a new (good) set of fibers between the NAS
deck interconnect panel and the control room
interconnect panel.
- If this does not fix problem, check the transmission
of the other fibers connecting the Leach board in the
saddlebag to the VME crate in the control room.
- Symptom
- Images show an elevated bias level on half of the image,
or on the entire image. Most commonly this is seen in
dual-amp mode where the right-side amplifier suddenly has a
bias level of over 40,000, as seen in the pre- and post-scan
columns. Aside from the elevated bias level, the image may
look normal (light can be detected).
- Problem
- The A/D converter for one amplifier has a stuck bit which
is raising all pixel values by some constant amount.
- Solution
-
- Try taking a dark image to verify the elevated bias
level.
- Power cycle the offending ccd crate using the
following command on lrisserver (assuming the
red side is causing the problem):
ccdpower red cycle
Remember to issue the lda 0 command on the
CCD crate to re-enable temperature control.
- If problem persists and affects only one amp, switch
to single-amp readout mode and use the good amplifier.
- If this is not possible, attempt to shorten exposure
times so that the elevated bias level and accordingly
reduced dynamic range of the image do not cause
saturated pixels for science target.
- Symptom
- After reading out part of an image successfully, the
readout pauses and the remainder of the image if filled with
zeros.
- Problem
- Communications between the CCD crate and the saddlebag
were lost during image readout.
- Solution
-
- Use the tdl command on the CCD
crate to check communications.
- It is likely you will find that the crate can't talk
to the VME interface card (i.e., tdl 0,1
fails). If so, the card must be reset. The image data
cannot be recovered.
- After resetting the card, use the tdl command to verify that
communications with the saddlebag are restored.
- If the problem persists, try switching to single amp
readout mode in order to reduce the data flow rate and
hence prevent lockups.
- Symptom
- Illuminated images show a circular patch in the image center
which has reduced flux (20% or more). Objects seen in images
may look fuzzy.
- Problem
- Water vapor has condensed onto the dewar window.
- Solution
-
- Symptom
- Bias level jumps randomly between two values during
readout, affecting both data columns and pre/post-scan
columns, as shown in this image.
- Problem
- A cable in the Leach CCD saddlebag readout electronics is
loose.
- Solution
- Inspect and reseat cables in the affected saddlebag.
- Symptom
- Bias level fluctuates wildly by dozens of DN across the
entire image, as seen in this image.
- Problem
- A connector is loose, possibly J6 MS connector at cass
interconnect panel.
- Solution
- Inspect and reseat cables at the cass panel.
- Symptom
- The exposure did not readout, but crate seems to be OK.
(NOTE: If the crate is NOT ok then go to Crate Crash instructions below). "testAll",
"tdl blue", "ctx" all indicate that all processes are running
normal. Blue exposures will finished, but the images are not
displayed in figdisp and the data is not written to disk. The
"Start" button on the XPOSE gui became inactive and needs to be
restarted to see the start button active again.
- Problem
- We don't know what the problem is at this time. Somehow
the software got tied in knots.
- Solution
- These are some things to try
- Run
LRIS Engineering --> Reset blue CCD flags from the background menu.
- take an exposure using the command line: goib
- reboot the blue crate
- Do a full software shutdown using the desktop
menu
- restart XPOSE, figdisp, and lickserve on the blue side
- Symptom
- The blue exposure reads out but looks freakishly discombobulated.
- Problem
- The blue-side windowing and binning are in an inconsistent
state. This can happen if the observer uses Xpose to change the binning without
making corresponding changes to the windowing.
- Solution
-
Use one of these options on the CMD... menu in Xpose to reset the windowing and
binning to an appropriate value:
- Window full frame mosaic (1x1)
- Binning 1x2 (spectral)
- Binning 2x2 (spatial and spectral)
- Symptom
- The blue exposure reads out the correct size, but every
pixel any given amp has the same value.
- Problem
- The +30V power supply on the LRIS blue side is powered off.
- Solution
-
Execute the command
ccdpower blue on to enable the
high voltage power.
- Symptom
- The direct image shows a dark bar running across
that looks like a hacksaw blade.
- Problem
- The slitmask move has failed and the grabber bar for the slitmask
is in beam.
- Solution
-
Try resending the slitmask move again.
- Symptoms
- Possible symptoms:
- Elapsed exposure time stops updating and timer is stuck
- Image never reads out
- The exposure did not completely readout leaving a partial image
in the ds9 window
- Problem
- The crate crashed. Note that it is possible to save
the image if the crash occured before the image readout.
- Solution
- IMPORTANT: DO NOT try to reboot the Archon without calling the support astronomer. The red side crate requires special handling.
-
- Hard Reboot the Crate: Go to LRIS Engineering -> CCD Crates -> Power cycle blue CCD crate
- start_crate blue in an lrisserver xterm
window to log into the crate. Monitor the crate
session to ensure that the crate
boots appropriately.
- If reboot is stuck at the VMXWORKS BOOT prompt,
type @ and return to initiate the rest of
the boot sequence
- At the end of the boot sequence, the prompt will
read done.
- log out of the crate: ^] followed by
quit at the telnet prompt.
- If the crash occurred before readout began,
it is possible to save the exposure. The observer
must decide whether it would be valuable to save
the image. Run the
recover_image blue script at the command line
or from the engineering portion of the pull down
menu. Script will readout the current data.
- take a 1 sec test frame to ensure everything is
okay. If the recover image script is run, it prompts you
to acquire a test image.
- If xpose gui will not take exposure run
reset_ccd_flags [red|blue]
- If still having problems, run testAll and
take appropriate actions.
- Note: The same power cycle command is available to OAs in the LRIS specific pull-down menu
- Symptom
- When crate is rebooted, the boot sequence doesn't
complete. Logging in via TIP session
(start_crate [red|blue] command) a pressing Enter yields only the following
prompt:
[VxWorks Boot]:
- Problem
- The startup script hung for some reason.
- Solution
-
- Symptom
- When crate is rebooted, the boot sequence fails before
the startup script is run. The following error message
appears on the crate console (or TIP line):
Attaching network interface ln0... done.
interrupt: ln0 lnInt: no carrier
Loading... interrupt: ln0 lnInt: no carrier
interrupt: ln0 lnInt: no carrier
interrupt: ln0 lnInt: no carrier
interrupt: ln0 lnInt: no carrier
Error loading file: errno = 0x3c.
Can't load boot file!!
[VxWorks Boot]:
- Problem
- Communications between the crate and lrisserver
are broken.
- Solution
- Check ethernet connections between the crates and
lrisserver.
- Symptom:
- CCD system suddenly cannot start an exposure after an
extended period of normal operation.
- Problem:
- It is likely that too many serv_
tasks have piled up on the crate. Verify the problem by
starting a CCD crate session (see the solution below) and
using the i command to check processes running.
Here is an example showing an excesss of suspended tasks:
-> i
NAME ENTRY TID PRI STATUS PC SP ERRNO DELAY
---------- ------------ -------- --- ---------- -------- -------- ------- -----
tExcTask _excTask 3f8ea0 0 PEND 37f60 3f8cf0 3d0001 0
tLogTask _logTask 3f6d48 0 PEND 37f60 3f6b98 d0003 0
tShell _shell 3c5818 1 READY 525cc 3c5208 30065 0
tRlogind _rlogind 3d3138 2 PEND 6f5cc 3d2d68 0 0
tTelnetd _telnetd 3d1260 2 PEND 6f5cc 3d0fc0 0 0
tNetTask _netTask 3f1410 50 PEND 6f5cc 3f1268 3d0002 0
tPortmapd _portmapd 3cfd20 100 PEND 6f5cc 3cfa30 16 0
MLOG_STDOUT_start_mlog_ 2b0da8 100 PEND 37f60 2b0878 3d0001 0
cserv _cserv 298550 100 PEND 6f5cc 298028 3d0002 0
ccdClock _ccdClock 27fcf8 100 PEND 6f5cc 27f880 3d0002 0
broad_mon _broadcast_m 2767f0 100 DELAY 70f2c 2766c0 3d0002 2595
responder _responder 24b2f8 100 PEND 37f60 24a010 3d0001 0
serv_4644 _s_show_auto 3b7ed0 100 SUSPEND 70924 3b7d40 3d0002 0
serv_4624 _s_show_numa 231a68 100 SUSPEND 70924 2318d8 3d0002 0
serv_4644 _s_show_auto 22f1a0 100 SUSPEND 70924 22f010 3d0002 0
serv_4632 _s_show_voff 22c8d8 100 SUSPEND 70924 22c748 3d0002 0
serv_4644 _s_show_auto 22a010 100 SUSPEND 70924 229e80 3d0002 0
serv_4622 _s_show_ccdg 226710 100 SUSPEND 70924 226580 3d0002 0
serv_4624 _s_show_numa 222e10 100 SUSPEND 70924 222c80 3d0002 0
serv_4624 _s_show_numa 21f510 100 SUSPEND 70924 21f380 3d0002 0
serv_4604 _s_show_binn 21cc48 100 SUSPEND 70924 21cab8 3d0002 0
serv_4630 _s_show_voff 21a380 100 SUSPEND 70924 21a1f0 3d0002 0
serv_4604 _s_show_binn 216a80 100 SUSPEND 70924 2168f0 3d0002 0
serv_4604 _s_show_binn 2141b8 100 SUSPEND 70924 214028 3d0002 0
serv_4606 _s_show_elap 2118f0 100 SUSPEND 70924 211760 3d0002 0
serv_4604 _s_show_binn 20f028 100 SUSPEND 70924 20ee98 3d0002 0
serv_4904 _s_show_raw_ 20c760 100 SUSPEND 70924 20c5d0 3d0002 0
serv_4610 _s_show_wind 209e98 100 SUSPEND 70924 209d08 3d0002 0
serv_4606 _s_show_elap 206598 100 SUSPEND 70924 206408 3d0002 0
serv_4606 _s_show_elap 203cd0 100 SUSPEND 70924 203b40 3d0002 0
serv_4502 _s_set_binni 2003d0 100 SUSPEND 70924 200240 3d0002 0
serv_4816 _s_set_utb_d 1fcad0 100 PEND 6f5cc 1fc898 3d0002 0
rccd _rccd 395390 150 SUSPEND 3fe730 395010 3d0002 0
value = 3951892 = 0x3c4d14
- Solution:
- We need to reboot the CCD crate.
- First, start a crate session using one of the two methods
below
- Reboot the crate by either
- resetting the toggle switch on the VME Sun1E board
(white reset button)
- or type "reboot":
-> reboot in the Tip
session. Note that sometimes only the reset btn will
work.
- Watch the output from the crate.
- Upon successful completion of the crate boot, the Tip
line session will print done and
pressing Enter gives you a
-> prompt.
- if boot sequence looks hung, pressing
Enter
will get you to the
[VxWorks Boot]:
. If that happens, type @ in the
tip session to cause the boot sequence to complete.
- Symptom
- vmemsgxchng (VME message exchange) errors such as the
following are seen on the CCD crate output:
vmemsgxchng: DSP reply timeout. Check power/hardware.
E1:Time=209340:"util_mem_rd.c", line 31: broad_mon, 24da70, "util_mem_rd ()",
Error, #1: Undefined error condition. Bad return 12 from vmemsgxchng() sending
RDM packet
E1:Time=209340:"get_dsp_data.c", line 108: broad_mon, 24da70, "get_dsp_data ()",
Error, #1: Undefined error condition. util_mem_rd returned -1007 reading loc
40000d of camera 0
E1:Time=209340:"broadcast_ccd_analog_inputs.c", line 72: broad_mon, 24da70,
"broadcast_ccd_analog_inputs ()",
Error, #1: Undefined error condition. get_dsp_data returned 3 reading UTB adc
chan 6 from camera 0
- Problem
- These messages indicate an inability of the CCD crate to
communicate with the utility board in the LRIS saddlebag.
Possible causes are:
- the saddlebag has no power
- fiber communications between the VME interface board
in the control room and the utility board are bad
- Solution
- Use the tdl command on the CCD
crate to determine the failure point and take actions
recommended at the bottom of that web page.
- Symptom
- After a power cycle and reboot the blue crate shows the
follwoing output:
Getting digital input word masked with 1 from utility board for camera
0
s_show_utb_dig_input: value = 0
- Problem
- These messages are produced after a reboot and power
cycle of the blue crate. It is unclear if they cause some
communication issues that in turn cause more blue crate crashes.
- Solution
- Rebooting the lrisserver manuka seems to solve the
problem.
- Symptom
- Trying to start multiexposures fails. The first exposure
ends very fast and the second exposure starts before the first
one has been read. The following error appears:
Error setting STARTEX: STARTEX: ARCHON_ERROR_STARTEX_EXPOSUID(-13058) cannot start exposure - cycle in progress
ERROR trying to start exposure
- Problem
- The lrisplus keyword RED_DETECTOR does not change and
remains on Idle.
- Solution
- restart the lrisplus daemon
- login to lrisserver
- run the following command lris stop! lrisplus
- wait a few seconds
- run lris start lrisplus
Check that the keyword changes during an exposure. If noit then
powercycle the Archon and restart lrisplus.
- Symptom
- When the START button on the
Xpose widget is pressed, the exposure does not
start. Logging into the CCD crate reveals shutter error
messages such as:
check_shutter_status: returning -1103
E1:Time=41488:"s_cshutter.c", line 491: serv_4714, 3b9310, "check_shutter_status ()",
Error, #1: Undefined error condition. Camera shutter is partially open!
E1:Time=41488:"s_cshutter.c", line 318: serv_4714, 3b9310, "do_shutter ()",
Error, #1010: Error number not found in error message table. Bad return from
check_shutter_status.
E1:Time=41488:"s_set_erase.c", line 234: serv_4714, 3b9310, "set_erase ()",
Error, #1010: Error number not found in error message table. Bad return from
close_shutter()
E1:Time=41488:"s_expose.c", line 254: serv_4714, 3b9310, "s_expose ()",
Error, #1: Undefined error condition. Bad return from set_erase()
- Problem
- The shutter is stuck partially open.
- Solution
-
- Cycle the shutter open and closed manually as directed
in this procedure. Then
try pressing the ABORT key on
the Xpose widget and attempt another exposure.
- If this fails, have technician cycle shutter open and closed
manually via the toggle switch on the LRIS shutter controller.
- If this fails, have technician cycle power on the LRIS
shutter controller.
- If this fails, check the cabling from the Leach
saddlebag to the shutter control box, and from there to
the shutter housing in front of the dewar.
- Symptom
- Attempts to start exposure or cycle the shutter fail.
Logging into the CCD crate reveals shutter error
messages such as:
safe_do_shutter: calling do_shutter.
"s_cshutter.c", line 304: serv_4740, fdb698, "do_shutter ()",
Error, #1010: Error number not found in error message table. CSH command
unsuccessful, reply was 455252
"s_cshutter.c", line 406: serv_4740, fdb698, "safe_do_shutter ()",
Error, #1010: Error number not found in error message table. Bad return
from do_shutter()
"s_cshutter.c", line 153: serv_4740, fdb698, "s_cshutter ()",
Error, #1010: Error number not found in error message table. Bad return
from safe_do_shutter()
- Problem
- The shutter is not operating.
- Solution
-
- Execute the test_shutter command from
lrisserver to verify operation of the shutter.
- If the shutter fails to move, the script will print
the following message:
------------------------------------------------------------------------
ERROR: The $side shutter is not functioning properly
------------------------------------------------------------------------
I tried to cycle the $side shutter open and closed, and it FAILED to work.
Should I power cycle the $side CCD saddlebag and the $side shutter controller
echo -n and then try again? (y/n) [y]:
- Press Enter to power cycle
the components. The script will try again to open the
shutter. Quit the script by answering n if it
does not succeed on the second try.
- If still not working, get onto the CCD crate and use
the tdl 0,3 command to verify that the crate
can communicate with the utility board, the card that
actually commands the shutter to open. If
tdl fails, try cycling power to the
saddlebag, crate, and/or VME interface card to restore
communications.
- If you can talk to the utility board, have technician
cycle shutter open and closed
manually via the toggle switch on the LRIS shutter
controller, which is located in the red-side enclosure.
The controller has three lights:
- red means closed
- green means open
- amber means unknown
An amber light is bad news and could indicate a
mechanical problem, possibly a broken linkage.
- If the shutter can opened and closed manually, the
problem is not mechanical and hence may be electrical or
electronic.
- Check the cabling from the Leach saddlebag to the
shutter control box, and from there to the shutter
housing in front of the dewar.
- If the cabling looks fine, use a voltmeter to
check the ?? cable for evidence of the +5V signal from
the utility card to the shutter controller which
triggers the open shutter operation. If this is not
found, then try to gain access to the utility card and
probe ?? for the +5V signal.
- Symptom
- Exposures on the blue side look like bias images or they have "
star trails" above a star
caused by the erase and below the star caused by the readout.
- Problem
- The shutter is not operating normally.
- Solution
-
Try the following
- Execute the cycle_shutter blue 5 command from
lrisserver to cycle the shutter open and closed. This may free
the shutter..
- Execute the test_shutter command from
lrisserver to verify operation of the shutter.
- If this fails, have technician cycle power on the LRIS
shutter controller,
- If this fails, have technician cycle shutter open and closed
manually via the toggle switch on the LRIS shutter controller.
Technician may access the shutter controller via:
- Removing the blue side electronics bay cover which is
at the bottom of the instrument when the dewars ate at
the fill position (3 O'clock when standing behind LRIS).
- locate the shutter control box at the left. It has red and
black
buttons with a toggle switch for manual and auto modes.
- switch to manual and open and close the shutter.
- Technician should hear a click as the shutter opens and closed.
- If it opens, observes can try automatic shutter mode or choose
to observe in a shutterless mode.
- If this fails, check the cabling from the Leach
saddlebag to the shutter control box, and from there to
the shutter housing in front of the dewar.
- Instructions for using the trapdoor as a shutter may be found
here
- Symptom
-
adctestAll reports the following errors:
Checking ADC daemons:
Checking galil...................ERROR! galildisp.tcl not running
Checking pulizzi.................ERROR! pulizziDispatcher not running
- Problem
- The ADC daemons are not running.
- Solution
- Log in to adcserver using
ssh kics@adcserver
and run
adc start daemons
- Symptom
- The ADC Status GUI reports that the ADC Mode is
Halt and the Status is Halted when it
should be tracking (i.e., the telescope is above an elevation of
30°).
- Problem
- The ADC mode has changed to Halted for some
reason, such as an ESTOP signal.
- Solution
- Follow these steps to put the ADC back into tracking mode:
- Log into lrisserver as lriseng
account on a local terminal.
- Execute the adctestAll command to check
ADC status. Note the ADC Status line, which
should be OK if the system is in Track
mode.
- Check the relevant ADC logfile in
/net/adcserver/usr/local/kroot/var/log/k1adclog for
clues that might indicate the reason the ADC halted (e.g.,
ESTOP signal detected).
- Verify the current mode via
show -s k1adc adcmod
- If the reported mode is not Track then
execute this command on the LRIS host to put the system into
tracking mode via
adcreset
- If this fails to return the system to tracking mode,
you can try to re-home the ADC via
adcinit
followed by another
adcreset
which should help the system return to Track mode.
- Verify that the ADC Mode displayed on the ADC GUI
changes to Track and that, after slewing to the
correct position, the ADC Status shows as
Tracking.
- If continued attempts to have the ADC track fail, then
null and stow the ADC by executing this command:
adcstow
- Symptom
- Guis report that the system is in a forward software
limit. The status message reads:
In fwd-soft limit
- Problem
- The ADC is in a forward software limit
- Solution
- To clear the limit status
- Select the "track" option to see if it will restart.
- Move to elevation that is legal and select "track" to
initiate tracking.
- Symptom
- The ADC will not change modes
- Problem
- The ADC daemons are not running appropriately
- Solution
- To fix the problem.
- login to host adcserver as user
kics using the appropriate account
- restart the ADC daemons:
adc restart daemons
- Re-initialize mechanisms and enable tracking by issuing
this command from any LRIS account on lrisserver:
adcreset
- From the desktop menus, restart the ADC gui.
- Symptom
- Obsserver ADC GUI status never says that it is tracking.
Engineering gui indicates that the separation is oscillating
between values.
- Problem
- There is a delay between information being received for
updating the position, and this delay is keeping the mechanism
motion out of phase with requested position updates.
- Solution
- To fix the problem
- Reset the ADC by executing this command from any LRIS account on
lrisserver:
adcreset
- If problem persists, halt the ADC again, place it
into position mode, and send it to the 20 mm separation:
madc adcmod=halt
madc adcmod=pos
madc adcval=20
- Re-enable tracking:
madc adcmod=track
- Symptom
- When in position mode, the ADC does not reach the intended
positon because it times out.
- Problem
- During long moves, the software may time out before the ADC
reaches the destination. This is due to a combination of a
relatively short time out and slow slew speed.
- Solution
- Re-execute the move command.
- Symptom
- ADC will not move because the primary hardware limit is set
- Problem
- The optics have moved too close to
the end of travel and have triggered the primary limit switch.
- Solution
- Remote recovery is possible as long as
the secondary limits are not triggered.
- Change to Position mode by issuing this command on the
LRIS host computer:
madc adcmod=pos
- Set the destination to 20 mm:
madc adcval=20
This should get the ADC out of the limit.
- Re-initialize via:
madc adccal=homed
- Start Track mode, to begin tracking:
madc adcmod=Track
- Symptom
- ADC will not move or initialize. The status indicates that the primary
and secondary hardware limits are set
- Problem
- The primary and secondary hardware limits are set.
- Solution
-
This requires manual intervention at the ADC to solve. If this
occurs at night, the protocol is to operate LRIS with the ADC
stuck in this state. If this occurs in the afternoon when
trained staff members are available, then they should address
the problem.
To move the ADC out of the primary and secondary limits, a
trained staff member has to manually turn the lead screw on
the back of the ADC. The ADC does not need to be removed
from the CASS position, however, if LRIS in in beam, LRIS
may need to be pulled back to gain access to the lead
screw. Someone should monitor the ADC status while the lead
screw is turned. As the lead screw is tunred, the secondary
limit will dissapear and then the primary limit will
dissapear in the status message window on the eng and
observer guis. After the limits are cleared, ensure that the
technician is safely outside the instrument and then
re-initialize the ADC. Set the mode to track to test whether
the ADC will track appropriately.
- Symptom
- The teperature of the CCD and saddle bags on either or both the red
and blue side are out of range. The LRIS dewar status monitor
task "/home/lrisserver/lris/bin/monitor_tempdet2"
running on the LRIS host computer "lrisserver" may have detected the
problem and is now sending e-mail warnings.
- Problems
-
There are several reasons for the temperature warning.
- The observers could be acquiring several short exposures
in a row.
- The saddle bag is powered down
- We have lost communication to the crate
- The CCD controller has been powered up without temperature
control running, and the temperature control loop has just been
started.
- The dewar is dry of LN2, and is starting to warm up.
- Solution
-
- Try running "testAll" on lrisserver
- Try running "tdl red" or "tdl blue" on lrisserver and follow
instructions if there are problems.
- If both tests pass, run "temps" on lrisserver. Typical output is:
[69] lris@lrisserver: temps
System Current Temp Acceptable Range Status
------------------- ------------ ---------------- ------
Blue CCD temp -120.0 -121.0 to -119.0 OK
Red CCD temp -100.0 -101.0 to -99.0 OK
Blue saddlebag temp -3.5 -10.0 to 0.0 OK
Red saddlebag temp -4.5 -8.0 to 0.0 OK
- if several temps output indicates that the temperature
is converging to the acceptable range, then continue to
monitor these messages until the CCD temperature
converges.
- If several temps output indicates that the temperature
is rising, i.e. continuing to warm up, then the dewar is
indeed dry and an immediate LN2 fill is necessary, to
prevent warm up and loss of vacuum. Contact the support
astronomer on duty if this happens.
- Symptom
- Reconfig software hangs while trying to retireve slitmask
information from the slitmask database.
- Problem
- It may be that polo is hung. Polo serves the slitmask
retrieval software because it is authenticated with the
slitmask database.
- Solution
-
- ping polo to see if it is alive
- if polo is alive, try to log in as lriseng
- if it hangs on login, contact computer
support
An example of how polo was hung from 30 July 2008. A
nirspec host computer (waikoko) disk was cross
mounted on polo. Waikoko was having problems and
this hung polo. JC unmounted the waikoko disk and
then polo was fine. Then the reconfig software ran
normally.
- Symptom
- Current status is not displaying valid values. If it is
a problem during the night, troubleshooting that would result in lost
sky time should not be completed. The Red dewar should be fine without
continually monitoring it. It is not a critical component.
- Problem
- It may be that the ion pump monitoring software is lost
following a power cycle or the ion pump electronics need to be power cycled.
- Solution
-
If power cycling is needed.
- Start the power gui: LRIS Engineering -> LRIS Power
Control
- Cycle power on the ion pump. This will cycle power and
restart the software.
If a software restart is needed.
- As lriseng@lrisserver run ionStart
Following a software restart, you may need to wait up to
five minutes before a valid measurment is recorded.
For further information see: ion pump technical doc
- Symptom
- Ion pump electronics are not responding
appropriately. Under normal opperation, lrmode = 1 and lrstat =3. An
example of common bad data from an ionshow is provided below. If it is
a problem during the night, troubleshooting that would result in lost
sky time should not be completed. The Red dewar should be fine without
continually monitoring it. It is not a critical component.
--- Current lris ion pump keyword values
lrcur = 0.000000000 amp
lrvolt = 0.000000 KV
lrmode = 0
lrstat = 0 (MAJOR: value is less than "low" limit)
- Problem
- The high voltage is not set appropriately.
- Solution
- First try restarting the software a few times. If this
fails, then staff will need to power cycle the high voltage
current on the ion pump electronics located in the Red side
electronics bay.
If it is working "lrmode = 1" and "lrstat = 3"
Following a software restart, you may need to wait up to
five minutes before a valid measurment is recorded.
For further information see: ion pump technical doc
|