WFI on MSSSO 40" Notes
====================
June 2001
As a first-time observer at the 40 inch I found there was virtually no useful
documentation, ie troubleshooting notes. However, with just a few simple
fixes, and some minor changes to the way I used CICADA, I could have saved
about three lost nights. I hope this saves you similar trouble.
Although I did have first-night training and excellent help from technical
support that doesn't cover a problem that crops up after midnight later
in the run!
These should be properly html'ised but I don't have time.
At present they are at
http://www.physics.usyd.edu.au/~ande/astronomy/WFI_40inch.html
These notes have grown out of notes written by Terry Bridges
(his file was /data/murky3/tjb/transit/WFI_notes but I can't find it now as
murky3 has gone) and the collective experience of myself, Bob & John
Shobbrook, John Goodyear, Winston Campbell, Gary Da Costa, Malcolm Harris
and others.
Things change: new problems appear, old ones get fixed, new fixes are
discovered. So feel free to copy, edit, improve, correct, add to, embelish,
simplify or whatever these notes.
Clear Skies!
Andrew Jacob
*****************************************************************************
CONTENTS
========
1/ BEFORE YOUR RUN...
2/ SOURCES OF INFORMATION
3/ GENERAL INFORMATION
BASIC DATA
CICADA, GIT & Ximtool
General
GIT (Graphic & Imaging Tool)
Ximtool
Screen Preferences for GIT and Ximtool
Other
COMPUTERS
FOCUSING
OTHER
CCD GUIDER SYSTEM
GSFIND
Procedure
Restarting Guider PC
TELESCOPE & DOME
Telescope
SDSU Cooling
CAMERA SHUTTER
Priming the shutter
Shutter Priming Delay
IRAF
3/ TROUBLESHOOTING
In the control Room
In the Dome
***************************************************************************
1/ BEFORE YOUR RUN...
Before your run I recommend you read:-
a) The 40-Inch & WFI DIY Manual.
(webpage version of a copy written in MS Word)
b) A short user's guide to the 40-inch and its CCD-camera system
c) Cicada User's Manual
d) These notes
Make sure you bring with you enough tapes or memory to back up your
data. DAT tapes are available at Siding Spring, but non-RSAA users
will be charged for their use.
Also, plan to arrive early for your run. TAKE BIAS FRAMES AND DOME
FLATS IN THE AFTERNOON to check that the system is working, so that if
you have any trouble, the technical staff will be on hand to help. You
should plan to start setting up for your nights observing in the
afternoon before observing at a bare minimum. This will allow you to
contact the technical staff to help you before they leave for the day.
***************************************************************************
2/ SOURCES OF INFORMATION
A) The 40-Inch & WFI DIY Manual.
Contents self-explanatory. Essential reading for first-time observers
using WFI at the 40inch.
Email John at sbrook@mso.anu.edu.au for a copy.
B) "40 Inch Telescope Observer's Technical Guide"
A new (June 2001) compendium of useful information and fixes.
More essential reading. You are encouraged to add to this guide to asssist
future observers.
Located on the control room bookshelf.
C) A short user's guide to the 40-inch and its CCD-camera system
This contains some basic 40" observing info. In particular, a list of what
to check when you arrive at the telescope.
D) The online Cicada User's Manual, also in hardcopy (white folder on 40" control-room bookshelf).
E) And the WFI web pages:
MSSSO
AAO (The AAO page contains QE curves or the CCDs).
F) Other Hardcopy Manuals in the Control Room
"40 inch Telescope ETS - A Short Users Manual" - located as above.
A useful guide to operating the ETS.
Only exists in hardcopy.
"A Short User's Guide to the 40-Inch and its CCD Camera System"
An old manual. Not entirely relevant to WFI but still useful in parts.
Now contained in the "40 Inch Telescope Observer's Technical Guide" folder
G) Telescope Command Reference Manual
More technical than the average user needs. Use the "40 inch Telescope
ETS - A Short Users Manual" instead (see F).
H) Additional Notes for Shobbrooks manual
Some additional or alternative notes to add to the Shobbrook manual.
Running CICADA:
-See additioinal notes below
Filling the dewars:
-Press the big red STOP button on the dome console to prevent
errant telescope movement.
-Remove (or replace) the lens-cap BEFORE filling the dewar to avoid (minor)
LN2 spillage from the camera dewar.
-Need a step by step procedure for pressurising the small LN2 tank??
-Need a step by step procedure for refilling small LN2 tank from large
LN2 tank??
Flat fields:
Dome flats
-Do these BEFORE filling the dewar to prevent major LN2 spillage.
-Dome at azimuth=~70deg. Telescope at ZD~=-70deg gets flat screen in place.
-Use dome lights (tungstens only!) or preferably sunlight by slightly
opening the upper shutter.
-Dome flats are not good. Use twilight and dark sky flats. [Terry Bridges]
Focusing the Telescope
-See additional notes below
Autoguiding
-If the autoguider PC has to be restarted:
-When asked for a password, just click "cancel".
-After windows starts open "vncviewer for WFI guider" icon.
The "VNC server" is m40slave:0 . Password is in the back of the
40" CCD Temp Log Book. You will need to scroll down the "Maxim Dl"
window to access the "Guide CCDs - WFI" chooser.
Data archiving
-the DDS3 DAT drive is now on monsoon (since Nov 2000).
-mt rewoffl rewinds and ejects the tape
-How to skip over a tape file if you want to read or extract the
second, etc file??
Shutdown
-Press the big red STOP button when not using the telescope to prevent
errant telescope movement.
Initialising CCD Hardware
-this should be done as a last resort, though there is some disagreement
on this!
Shutter priming
-See additional notes below
****************************************************************************
3/ GENERAL INFORMATION
Some useful information, most of which doesn't seem to appear anywhere else.
BASIC DATA
==========
WFI chips:
Eight 2k x 4k (2048x4096) CCDs in a 2x4 mosaic.
Each sees 13'x26' (arcmin) making a 52'x52' field.
Pixels are 15micron giving 0.375"/pix (on the 40in at f/8)
Gaps between CCDs are ~20-30arcsec.
Gains are ~1.5-2.1 electron/ADU
RONs are ~3.5-5.5 electrons
(To get Gain & RON follow procedure in "40 Inch Telescope Observer's
Technical Guide")
The awesome SDSU controllers read out the entire array in 55 seconds!!
A full array image is 140Mb in size.
One CCD is 17.5Mb
The /data/monsoon1 disc on Monsoon holds almost 100Gb - About 4-8 days data.
A DDS3 125m tape holds fractionally over 18Gb compressed
(ie using allocate mt1)
Writing to tape takes ~11.5min/Gb (~3.5hrs per full tape). So don't get
caught out on your last night, save early!
See also the AAO and MSSSO WFI pages.
CICADA, GIT & Ximtool
=====================
General
-------
Strategies to deal with CICADA.
Cicada has a tendency to hang, usually during readout, and sometimes
for no obvious reason (even if you are doing none of the bad things listed
below).
The only remedy is to restart it. To do this:
- In the Observers Window menu select actions|quit_observing_window
- In the CICADA menu select Start_Observing|Exit_CICADA
- From the command line do:
- psef |grep cicada
- Look for cicada processes. Ignore *_svc. Kill others with
cicada_cleanup -p 0
cicada_cleanup -p 1
cicada_cleanup -p 2
cicada_cleanup -p 3
- Restart using "cicada &"
In general restarting solves Cicada problems, but it's a pain so...
Follow these next practices for relatively trouble-free operation:-
Pause - DON'T use 'pause'. This crashes cicada so badly you need to logout
to restart! (apparently, but I never dared try it)
NumLock - Keep NumLock untoggled. It interferes with ximtool (and GIT?)
operation. In particular, it prevents you using right-mouse
scrolling to change brightness/contrast, and inactivates the
ximtool menus.
Initialise Hardware - advice is to use this as a last resort. (Doesn't
appear to cause any trouble tho, despite comments to the contrary)
Avoid:
-Using ABORT. You'll often then need to exit cicada and restart.
-doing too much (processor or memory intensive stuff) on monsoon during
readout. In particular avoid:
-Beginning or ending a tar command during readout (during exposing is OK)
-Doing much in Netscape (on-line banking is bad!)
-Copying images (and other operations) in iraf
Could do these things from murky but its way slower.
(For 6 CCDs (104Mb image) the limit might be ~400Mb (as shown on the
Observers window at "Memory:???MB"))
-leaving the little "graphics keystrokes" window open during readout
-changing regions (in regions window) during readout [This may not be true?]
During readout you CAN do these:
-change filter, change exposures, change duration
GIT (Graphic & Imaging Tool)
----------------------------
GIT (presently) has a problem with displaying mosaiced images. It displays
them OK on readout but it cannot set sensible scaling parameters to allow
you to calculate star profiles, get single line plots, etc. The ximtool
window just goes all black or white and displays a single value.
To overcome this limitation do the following:
-untoggle NumLock
-select options|start_private_ximtool
-select options|preferences :
-select category "General"
-untoggle "Mosaic MEF files in Cicada mode"
-enter the "MEF extension list", ie a list of which CCDs you want available
for loading into the private ximtool.
Note that if you are only using CCD numbers 1,2,7,8 (say) their MEF image
extension numbers will be 1,2,3,4. ie CCD number and extension number are
different things.
-set "Maximum Cicada images" if necessary
-The list of extensions you have requested will appear in the "image list"
window. Click one and it loads into the private ximtool.
Now you can get star profiles, cross sections, stats, etc. But when the next
image is read out it will automatically be loaded into the private ximtool.
But avoid leaving the little "graphics keystrokes" window open during readout
or GIT will crash out.
Ximtool
-------
North is right, East down.
1 pixel = 0.375arcsec
To move a star N dec offset on ETS is -ve
To " S dec " +ve
To " W RA " +ve
To " E RA " -ve
(This is better seen with a diagram)
Astigmatism leads to elliptical star images when defocused (see Focusing
below)
Screen Preferences for GIT and Ximtool
--------------------------------------
Select the monitor for these in options|preferences|general
Common practice puts GIT and Ximtool on screen 1.
Other
-----
Binning
If you change the binning, it may not change back again when you reset
it to 1 x 1, even though the regions window shows it as 1 x 1.
Watch the file sizes: should be 271440 blocks (full 8xCCDs?)
Bleeding
If readout shows bleeding on brighter stars on some CCDs and bias changes,
restart cicada.
Wrong Frame Size
If there is an error message about ximtool having the wrong frame size,
check that there is not an old cicada ximtool process running:
ps -ef | grep cicada will show it
Many Red Error Windows
If lots of red error windows appear from cicada, and restarting impossible
or doesn't work, try logging out (and back in). That can fix it. If the
error messages involve SDSU, then probably need hardware init (from action
menu): if such error message reported when cicada is restarted, may need
two hwinits. Do the hwinit right after restarting cicada.
COMPUTERS
=========
mosaic (in rack) runs controllers
monsoon (on desk) does rest - cicada
murky (in office) is a bit slow.
Data is stored on monsoon in /data/monsoon1/cicada/username
FOCUSING
========
This is difficult to maintain. There is backlash in the mechanism. And it's
pretty temperamental. But then just wait till you're 38 years old!
I found the best method was:
1/ At the beginning of a run (or if focus is lost completely)
Use Cicada's semi-automatic routine under actions|do_focus_sequence...
Follow the prompts and/or read the Cicada manual.
Move the focus by 0.03-0.05 units between each image, starting at
about 31.70. When do_focus_sequence is done set the focus position
with the bottons on the round control paddle
2/ From now on use the telescope's astigmatism. Star images are elongated:
NW-SE => focus too high: Press out [display decreases]
NE-SW => " too low : " in [ " increases]
3/ Check star profiles too as you go. Out of focus => lots of scatter
in the profile, of course.
...but then I was only taking short exposures (1-3 mins). Maybe you could
use the autoguider image to maintain focus with long exposures? See also the
Shobbrook manual.
It takes a couple of nights to get the hang of focusing.
Approach a focus position from the same direction to remove backlash. Otherwise
get double and/or smeared images.
Typical focus values range from 31.65 - 31.88
OTHER
=====
Readout speed
Use fast readout for broadband imaging when there is a choice.
CCD temperature
This appears on the Cicada observers window
(from 158-165K for TB). Was ~183.0 for my run.
Filter control
Select instrument_control|Filter:F_40 to open the filter control
window. To initialise, hit "get filter position"
Dewar vacuum
Watch the vacuum gauge - rises about 1E-07 per day.
Probably OK up to about 1E-04. LN2 hold time is about 12 hours.
If much less, dewar needs pumping (can be done cold in daytime).[TB]
CCD GUIDER SYSTEM
=================
I had no trouble with autoguiding, just used all the settings that were there.
The following are Terry's notes. [AJ]
To disable completely, turn off the switch near the mains plug boards
on the north side of guider box, low (on telescope)
GSFIND
------
This shows outline of the main and guiding CCDs relative to GSC
stars. Can start from cicada|tools menu or directly from command line.
Then drag to enlarge the window.
Choose monitor from cicada|options|preferences|general. Display on
monitor 0 to see a green outline against a black background. On
monitor 1 the outline and the sky are black [OK for me on monitor 1, AJ].
Put in the field coordinates and magnitude limits for stars (< 12 for 5 sec
exposures).
Set position angle = 90.
Move CCD outline with the LMB - gives (RA, dec) coordinates of the detector
center.
Put cursor on star and hit spacebar to see the star highlighted in table.
Also gives (x,y) position of selected star in the focal plane.
Procedure
---------
[needs editing? go ahead...]
-Select a guide star and the guide CCD.
-Set telescope, *open shutter*.
-Now go to the autoguider console.
VNCviewer is the guider program to choose from desktop. Maxim is the
autoguider program itself, similar to the one on the DBS.
-Choose "Guide CCDs - WFI" from the menu bar at the bottom of the
window. Choose which guide CCD you are using: same orientation as in GSFIND.
-In the "Maxim CCD" indow click the guide tab, enter the exposure time and
declination.
-Take an exposure with guide CCD: Click "expose" then "start".
-Wait for exposure to read out in big window.
-Move the star where you want it in the guide CCD field:
Use telescope offsets on round "box" to offset telescope;
North pulls the star *down* on guide window,
East pulls the star *right*.
-When star is centered (or where you want it), left-click on it, then
click on Track, then click Start.
-After it reads out it will create a little box in the upper left-hand
corner where you should see the star centered (roughly).
-Wait a bit to make sure telescope is tracking OK (no huge jumps in
star in guide window).
-When telescope tracking ok, close shutter and start the real
exposure.
**If you don't see any star in guide window, make sure shutter open!!**
Restarting Guider PC
--------------------
The Autoguider PC will ask for a password, just click "cancel".
-After windows starts open "vncviewer for WFI guider" icon.
The "VNC server" is m40slave:0 . Password is in the back of the
40" CCD Temp Log Book. You will need to scroll down the "m40slave"
window to access the "Guide CCDs - WFI" chooser.
Hit VNCviewer - this downloads display from m40slave (= m40spare).
Should see icons for WFI GS selector and Maxim. Run Maxim. Get the main
gui up by hitting the CCD control toggle button on the upper
menu bar. Go to setup/restart (don't worry about the "controller
off" message). Need to scroll window up and down to access the
bars at the top and bottom. Don't forget to put in the dec in the
main gui.
Settings for guider: make sure x,y axes enabled. Aggressiveness 4
is OK. X speed 10, Y speed 4.83, x axis cal time 1, y axis cal time 3.
On the guider display: all RA, dec buttons push the image.
TELESCOPE & DOME
================
**Red Emergency STOP buttons**
One is on dome console
One is on electronics rack in control room
Typically, you will run the telescope from the control room while observing
and from the dome when starting-up/closing-down, setting up for dome-flats,
filling dewar, showing-off to visitors, etc.
To transfer control TO dome:
1/ On ETS PC select Configuration|configure. Toggle Dome Control to manual.
2/ Go to dome.
3/ On dome console: switch to slew/set
4/ Move telescope, dome and floor as necessary
5/ Hit big red STOP button if filling dewar or stopping observing.
To transfer control TO control-room:
1/ Hit "emergency-stop reset" button (illuminated beside STOP button)
2/ Ensure floor is fully down
3/ Switch to auto
4/ Go to control room
5/ On ETS PC select Configuration|configure. Toggle Dome Control to automatic.
Telescope
---------
Telescope RA tracking sometimes drops out. Just redo the tracking command
to restart.
SDSU Cooling
------------
The SDSU controllers are cooled by the NESLAB cooler (downstairs in corridor
through library on the left)
If controllers get hot, probably means NESLAB has failed.
Check temperature of controller heat sinks occasionally - touch them.
CAMERA SHUTTER
==============
The shutter is badly designed. At large easterly hour-angles (>~4hrs) it
fails to reset. Recognise this by the star trails on an
image (trails on CCDs 1-4 trail up, on CCDs 5-8 trail down). There will be
a ramp in the signal across the CCDs too. [need an example image here]
To fix, you will need to prime the shutter manually.
You can listen to the shutter operation (loud CL-LICK on open and close)
with control room door open. A failed shutter sounds different, a woosy,
fainter click. You'll just have to listen! [need an example sound file here]
Priming the shutter
-------------------
Transfer control to the dome.
Move telescope to ~-45deg dec, HA=2.5hrs East. Move floor up to the level
of step 5 on the concrete stairs. This gives access to the big electronics
rack on the side of the telescope tube.
Follow the procedure in the "Observers Technical Guide" and/or Shobbrooks
procedure.
But then test the shutter is working again:
-return prime switch to NOR
-toggle mode switch to manual
-toggle display switch on
-toggle shutter switch open and closed to test.
(Don't forget to wait 8 seconds for the shutter to reset after closing)
-Return all four switches UP before resuming observing.
Shutter Priming Delay
---------------------
The shutter takes a few seconds to prime, so need to **wait** 8-10 seconds
after closing the shutter (eg after acquiring guide star) before exposing.
IRAF
====
WFI produces mosaiced images. In iraf need to specify the extension number in
the filename, as below.
Hopefully you've used IRAF before!
If not see http://msowww.anu.edu.au/computing/lun/lun35/lun35.html
Set the terminal type as xgterm (at the prompt after mkiraf)
and edit login.cl (or loginuser.cl) to set
set stdimage = imt4096
set imtype = "fits"
Then:
ximtool & :brings up an ximtool to display images in.
cl :starts iraf
And:
display filename[#CCD] 1
imcopy filename[#CCD][xmin:xmax,ymin:ymax] outfile.fits
imexamine filename[#CCD][xmin:xmax,ymin:ymax]
image operations: v does plot from two cursor input positions
r gives radial profile of star, fwhm etc
a gives mag and fwhm etc for star, no plot
Even for a single CCD readout or a small region readout, need
the extension filename[1]
Use the mscred package in iraf to reduce mosaiced images.
ds9, saotng are nice alternative display tools. ds9 can handle
mosaics.
3/ TROUBLESHOOTING
==================
If stuff goes wrong...
Here are some possible causes. Solutions are in these notes or somewhere
in the sources listed above. Sorry, I didn't have time to finish this quite
the way I would have liked. But I'm sure you'll manage :)
First, follow the suggestions above to keep Cicada as stable as possible.
In control Room
---------------
My object doesn't appear on Ximtool after readout?
-Cloud
-Dome is blocking view
-Dome shutter blocking view
-Camera shutter failure
I see streaks on my image?
-Camera shutter failure
-Initialise hardware (a last resort)
I can't analyse images (star profiles, stats, etc) in ximtool.
The image just goes to a single value!
-start a private ximtool in GIT.
Cicada has hung?
-Exit Cicada, cleanup, restart Cicada.
-Stop processor/memory intensive processes
-Avoid using abort
-Don't use pause
Lots of Red Error Windows, Can't Stop Them!
-Kill Cicada from the command line, cleanup and restart
I can't point the telescope more accurately than 20arcsec?
-Tough! The telescope only points to within about 20arcsec
-But, use buttons on round focus/guide/set paddle. Use set or guide speeds.
With practice this becomes fairly accurate (to within a few pixels).
Ximtool menus are inactive
-untoggle NumLock
I can't change brightness/contrast using left mouse scrolling?
-untoggle NumLock
New images keep loading into my private Ximtool?
-turn off automatic loading of images by toggling the "cicada mode"
button under the GIT options|preferences|general menu.
(This is suposed to work, but doesn't)
My star images are elliptical
-focus telescope
Focussing is difficult
-too bad!
-try focusing from same direction to take up backlash
There is bleeding on stars & bias has changed
-restart cicada.
I can't see my guide star (while setting up guiding)
-Open the shutter!
I can't find the "Guide CCDs - WFI" chooser window
-scroll down in the m40slave window
Autoguiding doesn't seem to be working accurately
-set the declination correctly on the "guide" tab in the Maxim CCD window
In Dome
-------
Dome doesn't follow telescope?
-Check dome position encoder (high on the western wall) is flashing its
green LEDs at you. If not restart ETS (select file|shutdown, then startup
at the prompt).
-Put dome in auto mode on ETS (select configuration|configure)
Telescope pointing is bad
-Calibrate the pointing
-Reset the time (Phone 1-1194)
Telescope doesn't move
-ensure floor is fully down
-Fault in ETS (Call technical support?)
Telescope moves unexpectedly by itself
-Press big red STOP button on dome console, or on electronics rack in
control room
Shutters don't move
-press shutter|stop button and try again (Listen for the relays clicking
downstairs)
-If upper shutter doesn't close it needs to be driven fully open (onto its
springs) first.
LN2 leaking from filling tube?
-Tighten the various nuts on the tube
LN2 leaking from camera dewar at high ZD angles?
-Remove lens-cap before filling camera dewar
-do dome-flats before filling camera dewar
LN2 small tank pressure low?
-small tank needs filling from large tank (call technical support?)
-pressurise small tank with N2(or air?) line (How?)
SDSU controllers feel hot
-Water cooler has failed. Check it and restart it, if necessary.
Dewar vacuum rising or focal plane temperature not holding
-pump dewar (call technical support?)