Images

Need to Abort an Exposure

Symptom

You need to abort an exposure.

Solution

Wait. NIRSPEC does not handle aborting exposures properly and crashes. Your best bet is to wait for the exposure to finish.

Really Need to Abort an Exposure

Symptom

You REALLY need to abort an exposure.

Solution

Ok, but don't say we didn't warn you! The server will likely crash. If it does, you will need to reinitialize all motors and retake all your calibrations.

1. Please see the entry regarding aborting exposures.
2. From an xterm on nirspecserver, type "m abort=1"
3. Wait about 15 seconds
4. Check NIRSPEC, can you show keywords? If so, you may have gotten lucky. Move along.
5. If not, please see our entry on recovering from server crashes.
6. Do not complain, you brought this on yourself.

Charge Persistence on Detectors

Symptom

Remnant images appear in subsequent images.

Problem

Charge persistence in detectors (both SPEC and SCAM can get persistence, but this is more of a problem with SPEC because it is the main science detector).

Solution

Avoid it by not exposing above 5000 counts (DN aka ADU) in a single SPEC coadd. Remove by flushing and/or waiting

1. If exposures are kept below 5000 counts per coadd, there is little, if any, charge persistence.
2. Remove by flushing: type flush from a waimea prompt.
3. There is a component of the persistence that only goes away with time, even after repeated flushes.
4. Remove by waiting or taking long darks: insert the blank filter and take several long (300s) darks frames. Examine each dark before taking the next.

Example

The image below shows a particularly bad example of the persistence effect. In this case, the original lamp flat was taken using the 42 x 0.76 arcsecond low-resolution slit and the Nirspec-7 filter. The integration time was the minimum 0.25 seconds, and 10 coadds were taken. CDS sampling mode was used. Peak counts on the original spectrum were about 25,000 per coadd, which is beyond the nominal 1% nonlinearity limit of 18,000 per coadd.

Most importantly, a signficant attempt was made to flush the persistent charge. After the original spectrum was taken, the flush command was used with the default 50 coadds, three times; flush was also run a fourth time with 200 coadds. After all that flushing, the image here was taken as a dark exposure, with itime=200 seconds and 1 coadd. The region originally exposed to the flat lamp has about 100 DN/pixel more "dark current" than the unexposed regions.

This image illustrates the importance of not overexposing the detector.

Charge Persistence Without Exposure

Question

Can charge persistence occur merely by landing a bright source on the slit with no exposure?

Answer

Yes. Beware of very bright stars on the slit

Details

To test the question, we took images with NIRSPEC in this configuration:
low resolution, 42 x 0.76 slit, NIRSPEC-1 + Thin blocker, cover closed, mirror in.

The images have itime x coadd of 60 x 3, and were taken after the detector was illuminated by the flat field lamp for approximately 180 seconds. With the flat lamp on, 60 x 3 in this configuration provides about 11,000 ADU/coadd. Both images have a 60 x 3 background image subtracted from them. The background image has the same setup as the "preflashed" images except with one or more flushes before integrating. We chose this method over darks (filter blank) as during the day the dome lights are on and there could be stray light problems.

"no lamp" image after exposure

cut across the long slit

"no lamp" image after illumination

cut across the long slit

One SPEC Quadrant appears Dead

Symptom

One quadrant of the SPEC detector looks dead

Problem

NIRSPEC server started in a strange state

Solution

Restart the server:

1. Right-click in background and run: NIRSPEC Control Menu -> End of Night Shutdown
2. Right-click in background and run: NIRSPEC Control Menu -> Start NIRSPEC Control Software

One SCAM Quadrant appears Dead

Symptom

One quadrant of the SCAM detector looks dead

Problem

Electronics issue inside the dewar

Solution

Many possible, none guaranteed to work:

1. Restart the server:
   1. Right-click in background and run: NIRSPEC Control Menu -> End of Night Shutdown
   2. Right-click in background and run: NIRSPEC Control Menu -> Start NIRSPEC Control Software
2. Power cycle the instrument:
3. Work-around: User defined nods:
   1. Define a nod pattern in /home/waimea/nirspec/setups. Currently, only moves along the slit are supported and the EFS does not check input. All moves are relative and in units of arcseconds. An example of an ABBA pattern that only uses one half of the slit is: "-2.0 -4.0 0.0 4.0 2.0".
   2. Select "User" within the EFS "Nod Pattern" pull-down
   3. Select your nod pattern
   4. Observe

SCAM Pickup Noise

Symptom

Horizontal banding in SCAM images

Problem

SCAM readout frequency is near 60 Hz harmonic

Solution

No solution. Characterization as follows:

1. Mostly affects sky-subtracted frames (nearly invisible in single frames)
2. Likely analog in origin as a vertical cut across the bands is more sinusoidal that square.
3. Peak-to-valley of 10-20 DN

No light in calibration images #1

Symptom

Calibration spectra have no light.

Problem

Calibration mirror is out of the beam.

Solution

Insert the cal mirror

1. Do either:
   1. Type m calmpos=1 from a waimea prompt
   2. Within the XNIRSPEC GUI: Click Cal Unit --> Mirror Out(button will change to Mirror In) --> Dismiss
2. Wait 10 seconds.
3. Take a short (0.1 second) SCAM image to see if light is falling down the slit.

No light in calibration images #2

Symptom

Calibration spectra have no light.

SCAM image taken during arclamps look like this:

Pinhole in beam	Pinhole out of beam

Problem

Calibration pinhole is in the beam.

Solution

Remove the pinhole

Try #1
1. Type m calpinit=1 from a waimea prompt
2. Wait 15 seconds.
3. XNIRSPEC should report that the pinhole is "Out" and the keyword calppos should be "0".
4. Take a short (0.1 second) SCAM image to see if light is falling down the slit.

Try #2
1. Type pinholeout from a waimea prompt
2. Wait 15 seconds.
3. XNIRSPEC should report that the pinhole is "Out" and the keyword calppos should be "0".
4. Take a short (0.1 second) SCAM image to see if light is falling down the slit.

No light behind AO

Symptom

There is no light from the AO fiber in images

Problem

AO pupil (in filter wheel 1) may have moved

Solution

Return the AO pupil to the beam:

1. Type set_ao_pupil from a waimea prompt
2. When complete, take a short (0.1 second) SCAM image to ensure light from the AO fiber is visible.

Weird-looking spectra

Symptom

Your spectra do not look right. See examples here.

Problem

A low-res slit is in beam with the echelle in high-res mode or vice versa

Solution

Re-select the slit or echelle position.

1. If you have doubts, use the echelle format simulator (EFS).

MAGIQ Guider

MAGIQ Troubleshooting

Link

MAGIQ Troubleshooting

Computers

waimea prompt

Symptom

Troubleshooting requires access to the NIRSPEC host computer

Problem

Need to open a window on waimea

Solution

Open a waimea window:

1. Right-click in background --> Login Windows --> xgterm 'waimea' (NIRSPEC)

no more processes

Symptom

waimea becomes unresponsive, cannot take images, move motors or even type simple commands such as ls.

Problem

The process table for user "nirspec" has filled on waimea

Solution

Reboot waimea -- although that may be hard!

1. Log into waimea under a different account like teloper
2. Perform a soft OA Reboot
3. If you cannot log in as another user, perform the following until you do get a working waimea prompt:
   1. Quit NIRSPEC GUI's (e.g. XNIRSPEC, Echelle Format Simulator, Quicklook, Image Rotator GUI, Slitnod Widget)
   2. Kill the window that says "DO NOT KILL THIS WINDOW"
4. Perform a soft OA Reboot

Soft reboot of waimea

Symptom

NIRSPEC and waimea are unresponsive and other troubleshooting steps have told you to perform an OA Reboot.

Problem

waimea, the NIRSPEC instrument host, is sick inside

Solution

Reboot (NOT to be done lightly!)

1. Log into waimea as teloper
2. Execute the command /etc/oareboot
3. Wait a couple of minutes
4. ping waimea or ping waimea.keck.hawaii.edu from another machine until you see the message: waimea is alive
5. Before you can observe, there are a few things you must do.

After a reboot

Symptom

You have just rebooted waimea

Problem

NIRSPEC is not yet ready to observe

Solution

Check the communications chain and health of instrument

1. Log into waimea as nirspec
2. Execute the command iboot from a waimea prompt (this tests the communication chain)
3. If the output looks like this:

   setting nirspec_mbox_query = 1 (wait)
   nirspec_mbox_status = ON
   setting nirspec_oibb_query = 1 (wait)
   nirspec_oibb_status = ON
   setting nirspec_crbb_query = 1 (wait)
   nirspec_crbb_status = ON
   

   Then do the following:
   1. Type runserver from your waimea prompt
   2. Type kill_all from your waimea prompt
   3. Type runserver from your waimea prompt
   If the server does not start properly, execute the command iboot all. This should clear most communication problems. Retest using the iboot command as before.
4. Start up NIRSPEC control software
5. Click here if you have problems starting the server, or you see error messages in the startup scripts.
6. Once the NIRSPEC software is up and running, issue the following commands from a waimea prompt:
   1. m test2=1
   2. cp /tmp/TESTFITS2 /tmp/TESTSKY2

iBoots do not respond

Symptom

After running iboot from a waimea prompt, you do not receive an "ON" response from one or more of the iBoots

Problem

The network router connected to the non-responsive iBoot needs a power cycle. This most often happens to the iBoot connected to the computer room blackbox.

Solution

Ask a summit tech to power cycle the router for the iBoot (it is NOT necessary to power cycle the iBoot itself).

• If the tech is unfamiliar with this procedure, photos and more details are located on the iBoot webpage.

Swap Space Warning

Symptom

A recorded voice declares that there is less than 50 Mb of swap space left on the NIRSPEC host computer.

Problem

The NIRSPEC server is running out of swap space.

Solution

Completely restart the NIRSPEC software within one hour. This will allow you to finish your observing sequence and get your cals.

Software

NIRSPEC Server Crash

Symptom

1. tklogger popup indicating server crash
2. Cannot take images
3. Cannot show keywords

Problem

The NIRSPEC server (nirspec_server_bin) has crashed

Solution

Run the recover script.

1. Right-click in background --> NIRSPEC Engineering --> Recover from server crash
2. Follow prompts of recover script (examples below)
3. Run whenever you think there may be a problem with the NIRSPEC server.
   • If no problem, you will see this message:

     The  SPEC test image was successful
     
     The server appears healthy, resetting observing parameters
     
     Continue observing
     

   • If there was a crash, after a successful recovery, you will see this message:

     Please start NIRSPEC control software from pull-down menu.
     **********************************************************************
     *                     IMPORTANT                                      *
     **********************************************************************
     * If you need flats or arcs, initialize ONLY the calibration mirror  *
     * via "m calminit=1."                                                *
     *                                                                    *
     * Once you have your cals, init all motors via "z_nirspec_init"      *
     * Do NOT enable nighttime mode before you initialize the rotator.    *
     **********************************************************************
     
     Once you have finished your cals and have initialized motors, 
     remember to enable nighttime mode.
     
     Really--don't forget to enable NIGHTTIME MODE after ALL GUIs are up.
     

4. Do NOT initialize all motors if you need flats or arcs for this instrument setting.
5. Initialize the calibration mirror: m calminit=1.
6. Insert the cal mirror (m calmpos=1) and take your cals manually using XNIRSPEC.
7. When finished, init all motors: z_nirspec_init.
8. Enable night-time mode via the pull-down.

NIRSPEC will not overwrite data. Upon a restart, the filenumbers will reset to 1; however when the first file is written to disk, the filenumber will increment to the next available number.

XNIRSPEC disappears or freezes

Symptom

XNIRSPEC disappears or stops updating.

Problem

Unknown.

Solution

Don't Panic. Restart XNIRSPEC:

1. Right-click in background --> NIRSPEC Control Menu --> Restart XNIRSPEC
2. If a SPEC exposure is in progress, wait for it to read out. The progress bar will not update, but the image will complete.
3. While waiting for the SPEC exposure, try a SCAM exposure just to see that XNIRSPEC is functional
4. If this fails, the NIRSPEC server may have crashed. Follow the server recovery instructions here.

XNIRSPEC Crash, Daytime

Symptom

XNIRSPEC disappears

Problem

DCS Simulator has died

Solution

Restart and connect to the DCS simulator

1. From a waimea prompt: make_dcs_sim
2. check_dcs_mode
3. Restart XNIRSPEC

Quicklook freezes

Symptom

Quicklook does not respond, its IDL window indicates a crash

Problem

Too many popup functions open at once, unknown causes

Solution

Restart Quicklook

1. Exit both Quicklook tools
   If "File" --> "Exit" or "Quit All" does not work:
   1. Find the iconified IDL session for the "dead QL"
   2. Double-click the icon to raise it
   3. Hit "Enter" a couple of times to get the IDL prompt
   4. Type "Exit" at the IDL prompt
2. Right-click in background --> NIRSPEC Control Menu --> Restart Quicklook

Image never displays

Symptom

User takes an image but Quicklook does not display it.

Problem

Usually one of two reasons:

1. itime x coadds less than 1 second
2. Quicklook has crashed

Solution

Solution to Problem #1

1. Increase the itime or coadds so that their product exceeds 1 second.
2. N.B. that the image is not lost. If a "test" image, it will be in the /tmp directory. If a "go" image, it will be in outdir or outdir2.

Solution to Problem #2: Restart Quicklook

1. Exit both Quicklook tools
2. Right-click in background --> NIRSPEC Control Menu --> Restart Quicklook

Quicklook popups do not respond

Symptom

Cannot access Quicklook's popup features (Gaussian Fit, Diagonal Cut, etc.)

Problem

More than one popup feature open (known Quicklook bug) or Quicklook has crashed

Solution

Close all popup features, try again

1. Click on DONE buttons of all open popups
2. Try to access the desired features one at a time.
3. If this doesn't work, restart Quicklook.

Quicklook "Move Tel" not working

Symptom

"Move Tel" gives the first prompt ("Click Mouse on Start Point") but never returns with the second prompt ("Click Mouse on End Point").

Problem

Conflict with "Diagonal Cut"

Solution

Dismiss the Diagonal Cut popup

1. Click on the DONE button within the diagonal cut popup.
2. If this doesn't work, restart Quicklook

Server does not start properly

Symptom

Error messages in startup script windows such as "RPC timeout". They may look like this:

rpc_clientOpen: RPC: Program not registered
Sorry, the show command was not able to contact the
control system: rpc_clientOpen() for waimea.keck.hawaii.edu.

Problem

The server did not start correctly, communications chain may be interrupted

Solution

Restart the server

1. From a waimea prompt: kill_all (tries to kill all server processes)
2. Once the prompt is returned: iboot all (power cycles elements of the communication chain)
3. Once that prompt is returned: check (talks to the transputers)
4. Type at the prompt: runserver (starts the NIRSPEC keyword server). The output should look like this:

   Hunting caRepeat processes on waimea...
   nirspec 25012
   Kill -9 sent to pid = 25012.
   Checking for existing servers (real and simulated)...
   Running server program.
   [1] 25191 25192
   Setting Configuration file values.
   ==================================
   setting q1offsetspec = 3240 (wait)
   setting q2offsetspec = 3240 (wait)
   setting q3offsetspec = 3240 (wait)
   setting q4offsetspec = 3240 (wait)
   setting q1offsetscam = 729 (wait)
   setting q2offsetscam = 725 (wait)
   setting q3offsetscam = 724 (wait)
   setting q4offsetscam = 722 (wait)
   setting tsptrace = 0 (wait)
   setting detbias = 450 (wait)
   setting getcryotemp = 1 (wait)
   setting getdetectortemp = 30 (wait)
   setting sensor.read = 30 (wait)
   setting outdir = /sdata600/nirspec (wait)
   setting outdir2 = /sdata600/nirspec (wait)
   

5. Start the NIRSPEC control software via a right click in the background --> NIRSPEC Control Menu --> Start NIRSPEC Control Software

Mechanism Moves

Motors will not initialize or move

Symptoms

1. NIRSPEC never acknowledges that initialization is complete
2. Motors do not move
3. Some or all cryotemps are not reporting cryotemps from a waimea prompt

Problem

The Lakeshore 330 temperature controller has been set incorrectly. The correct communication rate is 300 bps. Logic in the motor control software prevents motor moves if the temperatures are unknown or outside acceptable ranges.

Solution

Reset the Lakeshore 330 to the correct communication rate

1. Kill the NIRSPEC server.
2. Ask a summit tech to power cycle the Lakeshore 330.
3. The communication rate can be set from the Lakeshore unit.
4. Verify that the temperatures are reporting (temps should cycle on the front display of the Lakeshore 330).
5. Restart the NIRSPEC server.

Motor never stops slewing

Symptom

A motor move has been commanded, and it starts, but it never finishes. There is no message that the move failed or timed out, but the XNIRSPEC GUI reports that the mechanism is "slewing" forever.

Problem

A glitch in the communications between the server and the transputers. This problem occurs very infrequently.

Solution

Abort and re-init the offending mechanism

1. Abort: type m [stage]abort=1 from a waimea prompt.
2. Re-init: type m [stage]init=1 from a waimea prompt.
3. Re-try: Use XNIRSPEC to send the move again to the offending mechanism.
4. N.B. If the above procedure fails, a full restart of the server is necessary to fix this problem.

The keywords covered by [stage]abort are:

• irotabort abort image rotator movement
• fil1abort abort filter 1 wheel movement
• fil2abort abort filter 2 wheel movement
• slitabort abort current slit wheel movement
• echlabort abort echelle mechanism movement
• dispabort abort cross disperser movement
• calmabort abort calibration mirror movement
• calpabort abort calibration pinhole movement
• calcabort abort calibration cover movement

So, for example, to halt the Echelle grating, the command would be:

m echlabort=1

The keywords covered by [stage]init are:

• irotinit init image rotator movement
• fil1init init filter 1 wheel movement
• fil2init init filter 2 wheel movement
• slitinit init current slit wheel movement
• echlinit init echelle mechanism movement
• dispinit init cross disperser movement
• calminit init calibration mirror movement
• calpinit init calibration pinhole movement
• calcinit init calibration cover movement

So, for example, to halt the Echelle grating, the command would be:

m echlinit=1

Slit mechanism fails to init

Symptoms

1. NIRSPEC gives error message during init process such as "Mechanism timed out" or "Move failed" and "script aborted".
2. Subsequent tries to init the stage take several minutes to time out

Problem

One of the init switches for a mechanism is not working or is flaky. As of January 2009, the slit occasionally refuses to initialize. This procedure may work for other mechanisms.

Solution

Re-init (easiest) mechanism or find the mechanism and init in place.

1. Re-init mechanism (easy)
   All commands require a waimeaprompt
   1. Init the slit: type m slitinit=1
   2. Wait 5 minutes
   3. Check the slit status: type s slitstat
   4. If the result is "UNKNOWN" proceed to the next steps
2. Find the slit yourself (more difficult)
   All commands require a waimeaprompt
   1. Insert the cal mirror: m calmpos=1
   2. Turn on a cal lamp, e.g.: m neon=1
   3. If not running, start quicklook: z_run_rql
   4. Set SCAM itime: m itime2=0.1
   5. Set SCAM coadds: m coadds=10
   6. Take a SCAM exposure: goibuf2
   7. Often, even the failed init attempt will put the slit in the init postion (position 0) which is the 12 arcsec x 1 pixel slit
   8. If a 12 arcsec slit is in place, init the wheel in place: m slitinitloc=0
   9. Check the slit indexing: m slitpos=8 and take a SCAM image (the slit should be a 24 arcsec, horizontal slit)
   10. m slitpos=9 and take a SCAM image(the slit should be a 42 arcsec, vertical slit)
   11. Use the EFS to send your setup (this will turn off the lamp and remove the cal mirror)

Filter wheel fails to init or move

Symptoms

An error message indicates that one of the filter wheels failed to home or move properly.

Problem

Likely, one of the init switches for a mechanism is not working or is flaky. This procedure was developed for a failure in 2000.

Introductory Information

1. Each wheel has 12 positions with 9000 motor steps nominal between each position. Positive motor moves rotate the wheels to higher numbered positions (see the table of filter wheel positions ).
2. The filter wheels are located in a single housing located in the collimated beam in between the image rotator and the F-converter. Filter wheel 2 is the forward wheel (first in the optical path), and filter wheel 1 is behind it (when viewed from the front). In order to observe FW1 externally, then, FW2 must be in the open position. Each filter slot has a Lyot stop with a central obscuration, except for the 2.2 mm wide AO stop in position 2 of FW1.

   One can look down the beam with a flashlight to observe wheel motion; the person visually inspecting must climb into the K2 elevation journal (careful!), and the NIRSPEC front cover must be initialized and open. The person doing the visual should talk by telephone headset or radio to the other person issuing the motor moves. When viewed from the front, the wheels move CCW positive, that is, the filters will rotate from top to bottom of the observers view into NIRSPEC's light path when positive motor moves are issued.

Solution

Find the mechanism and init in place.

1. Attempt to initialize the stage one more time only using the XNIRSPEC/Engineer/Motors/Init menu item.
2. Check the status of the switches using the XNIRSPEC/Engineer/Motors/Read Switches menu item.
3. If no switches are set for the filter wheel in question, this indicates that the stage is not centered on a filter. Follow these steps to recenter:
   • Login to waimea as user nirspec.
   • Issue the command cdkw to go to the NIRSPEC keyword software directory.
   • Issue the command

     source fil1.csh

     (if filter wheel 1 is failing, else use fil2.csh). This script will move the stage 10,125 motor steps in increments of 75 and print out the status of the switches after each move. The status will be 15 as long as the wheel is between position. Interrupt the script (using Ctrl-C) immediately if the value changes from 15 to another value. If the script completes 10125 motor steps without ever seeing a switch, then a significant number of motor steps must have been lost (or the switch is failing). In this case, repeat the script once more to attempt to find the switch. If it fails again, then it is unlikely to work at all --- the wheel is probably stuck and/or the motor is burned out.
   • Check the status of the switches using the XNIRSPEC/Engineer/Motors/Read Switches menu item. If the position switch is set, then you were able to stop the mechanism in a centered position --- proceed to the next step. If not, you must backtrack to locate the switch position by typing this command:

     source rfil1.csh

     (or rfil2.csh as appropriate). Note that this script prints out both the status of the filter wheel switches and the status of the slit switches. Again, hit Ctrl-C if and when the fil1sw value changes. Re-check the switch settings as described above and, if needed, run fil1.csh to go back in the other direction. Repeat as needed until you are able to catch the position switch 'on'.
4. On the Read Switches menu, check whether the Pri Init or Sec Init light is 'on':
   • If Pri Init is on, you are a lucky soul, because you found the primary initialization position for this wheel. Consult this table to determine which filter you are at. Read the corresponding motor step position, and enter this command to initialize the wheel manually:

     m filMinitloc=N

     where M is the filter wheel number (1 or 2) and N is the number of motor steps from the table. Example: if you are on filter wheel 1 and have hit the primary init switch, then you've found M-wide at 90,000 motor steps.
   • If Sec Init is on, you are an equally lucky soul, because you found the secondary initialization position for this wheel. Consult this table to determine which filter you are at. Read the corresponding motor step position, and enter this command to initialize the wheel manually:

     m filMinitloc=N

     where M is the filter wheel number (1 or 2) and N is the number of motor steps from the table.
   • If neither the Pri Init nor the Sec Init light is 'on', then we still can't determine our location unambiguously. In this case, continue running fil1.csh (or fil2.csh) until you hit the next position switch, then again use Read Switches to check for Pri Init or Sec Init. Continue as needed until you hit one or the other.
5. Once you've located a primary or secondary initialization switch and used the filMinitloc keyword to define position, the XNIRSPEC window should properly reflect the current position. This means that you can use the Echelle Format Simulator window to manipulate the instrument as long as you set the requested filter to whatever is currently in the flaky wheel.
6. If you are brave, you can try moving to an adjacent filter using the filter selector buttons on XNIRSPEC. But beware --- if the move fails, you'll need to start again looking for position switches and initialization switches!

Rotator

Poor slit nodding #1 (cycle guiding)

Symptom

Nods along the slit fall out of slit when guiding on annular guider with the rotator in stationary mode, no other symptoms

Problem

Guiding software thinks the center of rotation has moved

Solution

Re-center object on slit (SCAM x, y = 131, 126 +/- 1 pix), then stop and restart guiding

1. Observer: re-center object via Quicklook's "Move Tel" function or via the SlitNod widget
2. OA: stop and restart guiding

Discussion

When guiding is started on the annular guider, it assumes the object is at the slit center. It interprets any subsequent telecope move as a slitnod. If guiding was started and then the object was moved to the slit center, the object will no longer be at the center of rotation.

Poor slit nodding #2 (rotator server restart)

Symptom

Nods along the slit fall out of slit and the position angle is incorrect, the rotator will not move, or TkLogger pops up with repeated rotator warnings.

Problem

The rotator server has died.

Solution

Follow the TkLogger directions:

1. Right-click in background and run: NIRSPEC Control Menu -> Restart Image Rotator
2. Wait 5 seconds or so
3. The TKlogger popups should stop. Dismiss TKlogger.
4. (Optional) Turn rotator tracking on from XNIRSPEC: IROT -> Tracking On
5. (Optional) Reset your desired rotator mode and angle from the Rotator GUI.

The object should still be in the slit, but you will want to check it.

Poor slit nodding #3 (rotator needs initializing)

Symptom

Nods along the slit fall out of slit and you see a message in XNIRSPEC suggesting you reinitialize the rotator or restarting the rotator server did not fix the problem.

Problem

Rotator has lost motor steps or tracked into a limit

Solution

Reinitialize the rotator (take cals first!!!)

1. Tell OA you need to reinitialize the rotator and ask them to stop guiding
2. Do either:
   1. Within the XNIRSPEC GUI: Click Engineering --> Motors --> Init --> Image Rotator --> Init --> Dismiss
   2. From a waimea prompt: m irotinit=1
3. The status box beneath the rotator icon in XNIRSPEC should read INIT. Once this status box reads OK, reset your desired Physical or Position Angle from the rotator GUI

Red box on GUI never turns white

Symptom

You sent a rotator command, but the indicator box on the rotator GUI never turned from red (off target) to white (on target).

Problem

Many possible problems including:

1. IROT tracking not started
2. Rotator software crash
3. Rotator slewed into a limit

Solution

Solutions given in order:

1. (Re)Start IROT tracking
   1. Within the XNIRSPEC GUI: Click IROT --> Tracking Off
   2. Within the XNIRSPEC GUI: Click IROT --> Tracking On
   3. Within Rotator GUI: Type in the Physical angle set box: 0.0 and click SET.
   4. The red box should turn white within 15 seconds.
   5. Set your desired Physical or Position Angle.
2. If the above doesn't work, restart rotator software
   1. Right click in background --> NIRSPEC Control Menu --> Restart Image Rotator
   2. Reset your desired Physical or Position Angle from the rotator GUI.
3. If that doesn't fix it, reinitialize the rotator mechanism

Position Angle is Incorrect

Symptom

The desired position angle is incorrect after a telescope slew

Problem

The rotator does not move during telescope slews

Solution

Reset the position angle

1. If guiding, ask the OA to stop guiding.
2. Type in your desired position angle into the rotator GUI.
3. Click SET
4. If this does not fix your problem, you may need to restarting the image rotator server or reinitializing the rotator mechanism may help.

What is PRESROT?

The DCS keyword PRESROT defines what the rotator will be asked to do during a telescope slew. The possible values for this Boolean keyword are 0=sky, which means "preserve the PA on the sky during the slew," and 1=phys, which means "preserve the physical angle you're at right now during the slew." PRESROT should always have a value of 1 or phys for NIRSPEC, because of the limited range of physical motion of the NIRSPEC image rotator (-90 degrees to +90 degrees). If PRESROT = sky for NIRSPEC, the rotator will work very hard during telescope slews and probably slew into a limit, requiring a rotator intialization.

Set The PA After Every Slew!

One consequence of keeping PRESROT = phys is that the user MUST reset their desired PA after every slew. If, for example, the user is tracking a science object at SlitPA = 0 degrees, then slews to a calibration star 15 degrees away, the rotator will maintain the physical angle it had on the science object, then just begin tracking on the cal star from that same physical angle, which will of course yield a different SlitPA on the sky for the cal star. There are also reports that after long slews, the rotator is so far from the last requested PA that the telescope pointing and slit nodding become inaccurate, because the demanded PA and actual PA are so different.

To get the same SlitPA on the cal star, after the slew, the user would have to enter 0 degrees and left-click "SET" on lower middle panel of the Rotator GUI. PLEASE NOTE also that this resetting of the desired SlitPA should be done after the slew is complete. Resetting while the slew is in progress will cause the reset to be inaccurate, as it will be based on the instantaneous telescope position when you click "SET" on the Rotator GUI.

Tracking Directions in Normal Operation

In normal operations, the image rotator tracks in different directions depending on the location of the object in the sky:

• Tracks CLOCKWISE if the object is SOUTH of Mauna Kea (i.e., Declinations SOUTH of about 19 deg 50 min)
• Tracks COUNTERCLOCKWISE if the object is NORTH of Mauna Kea (i.e., Declinations NORTH of about 19 deg 50 min)

Pay attention to the current physical position of the rotator when selecting a new PA. As you type numbers into the SlitPA or SCAMPA text entry boxes, the lower left Physical panel will preview the physical angle of the rotator at the current telescope pointing and your new PA. If there is a chance that the rotator might track into a limit, then choose instead the 180-degree complement of your desired angle and check the preview to see that this gives you a safe amount of tracking range.

There is an arrow in the upper left "Physical - Actual" part of the Rotator GUI that attempts to represent the tracking direction of the rotator.

Mira

No light in focus (Mira) images

Symptom

There is no light in focus (Mira) images

Problem

Either the cal unit is incorrectly configured or the wrong filter is in the beam.

Solution

Mira was designed for K-band or K-like filters

1. Ensure the Cal Cover is open, the Cal Pinhole is out, and the Cal Mirror is out. If necessary, ask the observer to re-send an instrument "Setup Only" script from the EFS.
2. Ensure the selected filter is one of the following: NIRSPEC-7, K, K-PRIME, NIRSPEC-6, NIRSPEC-5 (H); each with the Thin Blocker also in place.